In 2015, waiting six weeks for user research felt normal.
In 2026, it feels like sending a fax.
Yet most companies still treat feedback like it’s a quarterly ritual. Write the test plan. Recruit participants. Run interviews. Watch the videos. Build the deck. Present the deck. Finally… maybe fix the thing. By then, your competitor has shipped three versions.
Chris Hicken has lived this cycle from the inside. As the fourth hire and President of UserTesting.com, he helped scale the company from a few hundred thousand in revenue to nearly $100M and through IPO. He understands the traditional research model better than almost anyone.
And now he’s rebuilding it from scratch.
The Problem: Nobody Wants to “Do Research”
Here’s the uncomfortable truth Chris shares:
Nobody wants to do research. They want to know what to fix.
Traditional user testing is slow because it was built for a different era:
Slower development cycles
Heavier engineering constraints
Fewer product experiments
Today?
Junior engineers ship in days using AI-assisted coding tools.
Founders test new features weekly.
Product teams iterate constantly.
But research hasn’t caught up.
So Chris founded Theysaid, an AI-native feedback platform designed around one core principle:
Push a button. Get an insight.
Not in weeks.
In hours.
Why the “5 User Test” Is Dead
For years, product teams leaned on the famous idea: test with five users and you’ll uncover most usability issues.
That made sense when:
Running tests was expensive
Watching sessions was manual
Analysis took days
But with AI reviewing sessions, summarizing patterns, and clustering feedback, you’re no longer limited by human bandwidth.
Why stop at five users when you can run 100?
More importantly:
Why separate qualitative insight from quantitative confidence when AI can combine both?
The old constraint was human time.
The new constraint is how fast you’re willing to learn.
AI Hype vs. AI Reality
There’s a lot of noise in AI right now.
Solo founders vibe-code their way to billion-dollar acquisitions.
Startups hit $100M ARR in record time.
Everyone declares SaaS dead every other Tuesday.
Chris offers a calmer take:
The fundamentals haven’t changed.
You still need to solve a painful problem.
You still need product-market fit.
You still need execution.
AI makes building faster. It does not make thinking optional.
In fact, shipping AI-native software for enterprise is harder than most people realize. Out-of-the-box models hallucinate. They miss edge cases. They misinterpret nuance.
To deliver trustworthy insights, Theysaid built layered AI systems—agents reviewing agents, QA loops, structured validation.
This isn’t “ask ChatGPT once and call it done.”
It’s iteration—hundreds of times over.
The Hardest Problem in SaaS: Product-Market Fit Never Ends
One of the most valuable insights from the conversation wasn’t about AI at all.
It was about product-market fit.
Chris built a framework called the CPV score to evaluate fit across four dimensions:
Problem
Product
Price
People
And here’s the key:
Product-market fit isn’t permanent.
It’s a snapshot in time.
Every time you:
Enter a new segment
Move upmarket
Change pricing
Add a new use case
You are crossing the chasm again.
Many companies stall at $10M because they assume early success guarantees expansion. It doesn’t. Each new vertical is a new test.
The companies that survive don’t assume.
They measure.
A Founder’s Edge: Domain Expertise vs. Problem Obsession
Interestingly, Chris doesn’t believe domain expertise is the secret to startup success—even though this is his third company and second time in the research category.
His take:
Good founders aren’t product-obsessed.
Great founders are problem-obsessed.
The advantage isn’t “I worked here before.”
It’s “I can see the pain clearly and I won’t stop until it’s solved.”
In this case, the pain is obvious:
Modern teams build at AI speed.
Feedback still runs at 2015 speed.
That gap is the opportunity.
What Changes in the Next 5 Years?
If Chris is right, user research becomes:
Instant
Embedded in workflow
Accessible to anyone, not just research specialists
Think smartphone camera vs. professional DSLR.
There will always be experts.
But most teams just need quick, reliable feedback.
And when insight becomes instant, something bigger happens:
More experiments.
More iteration.
Better products.
Learning becomes the competitive advantage.
The Bigger Pattern
Every 10–15 years, software undergoes a platform shift:
Web
Mobile
Cloud
AI
Each time, the tools change.
Each time, the pace accelerates.
But the core game remains the same:
Find a painful problem.
Solve it better.
Keep adapting as the market moves.
Chris didn’t come back to user research because the old system worked.
He came back because it didn’t.
And in an AI-native world, the winners won’t be the teams with the biggest research budgets.
They’ll be the teams who can learn fastest.
👂🎧 Watch, listen, and follow on your favorite platform: https://tr.ee/S2ayrbx_fL
🙏 Join the conversation on your favorite social network: https://linktr.ee/theignitepodcast
Chapters:
00:01 – Welcome and Chris Hicken Introduction
02:21 – Origin Story and UserTesting Journey
04:18 – Why Traditional User Research Is Broken
05:08 – Compressing Research from Weeks to Hours with AI
05:31 – Synthetic Testers and AI Personas
08:18 – Building Massive AI Testing Panels
09:12 – Mission Accomplished or Not at UserTesting
09:47 – Is the 5 User Test Dead
12:17 – Marrying Quant and Qual with AI
12:42 – Lessons Scaling to 100M ARR
13:53 – Professional Services vs SaaS Revenue
16:50 – Jobs to Be Done and Insight Delivery
19:31 – AI Workflows and Engineering Acceleration
23:18 – How Product Teams Are Changing
24:53 – Building in One Unified Workspace
25:45 – Third Time Founder Reflections
27:37 – Thriving on the Edge of Failure
28:19 – Theysaid 3.0 Launch
30:06 – SaaS in the Age of AI
32:02 – Hype vs Reality in AI Startups
36:18 – Fundamentals Still Win
39:14 – AI Limitations and Enterprise Reliability
41:28 – Competing Against UserTesting
43:04 – Domain Expertise vs Problem Obsession
45:17 – Preventing AI Hallucinations
49:28 – Leadership and Productivity Systems
52:28 – Single Source of Truth for Work
56:11 – Filtering Noise in Fast Moving AI Markets
57:52 – First Startup and Amazon Competition
59:48 – Metrics That Actually Matter
01:01:36 – Product Market Fit vs Culture
01:04:35 – Expanding into New Segments
01:06:21 – Worst Advice in SaaS
01:06:48 – Learning from Failed Founders
01:08:21 – Books and Frameworks
01:09:42 – Building Team Ignite
01:10:52 – The Future of UX and AI Research
Transcript
Brian Bell
(00:01:15): Hey, everyone, welcome back to the Ignite podcast. (00:01:17): Today, (00:01:18): we’re thrilled to have Chris Hicken, co-founder and CEO of They Said, an AI-powered user testing and research platform.
Brian Bell
(00:01:23): Chris was the fourth hire and president at usertesting.com for eight years, (00:01:27): where he took the company from a few hundred thousand in sales to just under 100 (00:01:31): million. (00:01:32): Since their IPO user testing and all of its major competitors have been acquired by (00:01:35): private equity, (00:01:36): wow, (00:01:36): which created a great opportunity for Chris to get back into the category and build (00:01:40): the next generation of feedback tools. (00:01:41): The way the world builds products has changed dramatically. (00:01:44): Vibe coding, MCP, AI code editors, agents, et cetera. (00:01:47): And Chris is staying at the cutting edge by giving builders a way to get feedback (00:01:52): rapidly while they build cool stuff. (00:01:53): Thanks for coming on, Chris.
Chris Hicken
(00:01:54): Hey, bud. (00:01:55): Thanks for having me on. (00:01:56): We’ve been talking about this for a long time, so I’m glad we’re finally making it happen.
Brian Bell
(00:01:59): Yeah, yeah. (00:01:59): So Chris and I are old friends. (00:02:01): I’ve literally stayed at this house. (00:02:02): So we spent the last 20 minutes catching up before we started recording. (00:02:05): So just a little context if we feel kind of chummy. (00:02:08): But yeah, (00:02:09): I’d love to start with kind of the core question,
Chris Hicken
(00:02:11): which is like, (00:02:12): you know, (00:02:12): I said a little bit about your background, (00:02:14): which is incredible, (00:02:15): but maybe you could just give your origin story real quick and we’ll kind of dive (00:02:18): into your founder journey. (00:02:19): Yeah, (00:02:20): well, (00:02:20): the work story is I can kind of summarize it as I’ve been building startups pretty (00:02:25): much my whole career. (00:02:26): I’ll start with usertesting.com, (00:02:29): which was it’s kind of I believe it’s the biggest product that’s used by design and (00:02:36): UX research teams in the world. (00:02:38): It went IPO in 2021. (00:02:40): I left there to start a company called Nuff Said, and we’ll talk more about that later. (00:02:44): But it was a tool that aggregated all of your apps, all of your work apps in one place. (00:02:49): So your email, your Slack, your calendar, your tasks in one. (00:02:54): And we sold that company to another business with the same exact vision, (00:02:58): which is ClickUp, which is a San Diego based company. (00:03:01): And with the advent of ChatGPT and all of its variants since, (00:03:05): there was an obvious gap in the market to come back and do another feedback (00:03:09): company. (00:03:10): So that’s what we’re doing. (00:03:10): That’s what we’re doing now.
Brian Bell
(00:03:12): Pretty interesting. (00:03:12): So basically, (00:03:14): you know, (00:03:14): you can think of user testing as sort of the 10 years ago, (00:03:17): if you’re building a user testing platform, (00:03:19): how you would build it. (00:03:21): What’s different about what you’re running now? (00:03:23): And how did you get to kind of founding this company?
Chris Hicken
(00:03:25): Well, okay, so two parts to that. (00:03:29): What’s different about what we’re doing now? (00:03:30): The process of doing research is painful and it sucks. (00:03:34): And the reality is nobody wants to do research. (00:03:37): In fact, even user researchers don’t want to do research. (00:03:40): Because the process is pretty painful. (00:03:42): So it starts with like, (00:03:44): you have to come up with a series of questions or a test plan for your survey or (00:03:48): for your user test or your interview. (00:03:50): And that’s a very time consuming process to get that right. (00:03:52): And then you have to go and recruit people to do your research project with. (00:03:58): And then you have to either do the interviews yourself or do the user test yourself (00:04:02): or you hire a company like UserTest and do it. (00:04:05): And then afterwards, you have to watch the interview or you have to watch the user tests. (00:04:09): And then you have to do the summary and the analysis. (00:04:12): And then you have to produce the reports for your executive team or for your peers (00:04:15): in order to summarize like what you learned. (00:04:17): So this whole process, (00:04:18): even if you’re using a company like UserTest, (00:04:20): still takes weeks, (00:04:22): sometimes months to accomplish. (00:04:23): And I think we use user testing back then. (00:04:26): And so you get the panel and you kind of walk them through UX experiences. (00:04:29): And that whole process ended and did take about a month. (00:04:32): And we had a full time team, like a scrum team, like four or five people working on it. (00:04:37): Yeah, (00:04:38): we’d have a hypothesis like, (00:04:39): oh, (00:04:39): we want to test this, (00:04:40): you know, (00:04:41): here’s this clickable Figma prototype, (00:04:43): or here’s how we kind of want to design this thing. (00:04:44): And it just took a long time. (00:04:46): There’s a lot of manpower to kind of get through this process. (00:04:49): Yeah, (00:04:49): I mean, (00:04:50): unless you’re a ginormous team, (00:04:51): I mean, (00:04:51): who has four to eight weeks to spend doing research? (00:04:55): I mean, especially nowadays with the way that people are building products today. (00:04:58): I mean, there’s absolutely no time to do that. (00:05:00): There’ll be a new version of Claude or ChatGPT out that designs a whole new (00:05:03): interface in eight weeks. (00:05:05): Yeah, exactly. (00:05:07): So what we’re doing now is it’s all AI-powered based research. (00:05:10): So AI is doing everything from writing your test plan to... (00:05:14): moderating your user testing sessions to analyzing the results to publishing, (00:05:18): you know, (00:05:19): report, (00:05:19): you know, (00:05:19): reports that are ready to share with your with your executive team or with your (00:05:22): peers. (00:05:23): And the goal is to get all that time that four to six weeks down to like one or two (00:05:27): hours where you ask a question, (00:05:29): like push a button, (00:05:29): get insights. (00:05:30): That’s that’s what we’re building with this company. (00:05:32): So ensure like AI singular singularity exponential times, (00:05:37): you’re compressing this timeline of getting the testing from weeks to like (00:05:42): practically a day. (00:05:43): Ours. (00:05:43): Yeah, really ours is what we’re shooting for.
Brian Bell
(00:05:45): Are each just actually kind of going through the experience and kind of testing it? (00:05:48): Or is it humans or some combination?
Chris Hicken
(00:05:51): It’s a combination right now. (00:05:52): So the platform that we’ve rolled out right now is available to test with humans primarily. (00:05:57): However, we’re also developing AI. (00:06:01): We’re calling them synthetic testers. (00:06:02): But these are AI... (00:06:05): personalities that we can deploy to do user tests for you. (00:06:09): And each one actually has a very different personality and it interacts with your (00:06:12): test differently. (00:06:13): And, (00:06:14): you know, (00:06:14): I’ve kind of, (00:06:14): I’ve simplified the process, (00:06:16): but the, (00:06:16): you know, (00:06:17): getting AI to do this is actually a real pain in the ass. (00:06:21): It does not do this well out of the box. (00:06:23): And so we’ve had to do a ton of development to make it
Brian Bell
(00:06:27): work the way that a researcher or a product designer would want this to work to get (00:06:31): feedback while they’re building stuff. (00:06:33): This is one of the really cool things about AI is that you can prompt it in such a (00:06:36): way to get it to assume a persona. (00:06:40): I’m a 55 year old female that still uses Hotmail and I’m not very familiar with the UI. (00:06:46): I prefer not to use it. (00:06:48): and go, right? (00:06:49): Is that a little bit how it works? (00:06:50): Yeah, (00:06:51): I mean, (00:06:51): that’s a very basic version of it, (00:06:53): but there’s so much more that you can build into the persona, (00:06:56): like the roles and responsibilities, (00:06:58): how big of a company they work for, (00:07:00): the pace of their development, (00:07:02): how they develop, (00:07:03): how they do their job, (00:07:05): what stage they’re at in terms of their research of buying a product, (00:07:09): for example. (00:07:10): So there’s 100 different factors that you can use to give AI a persona, (00:07:17): And different models perform differently. (00:07:19): So if you put Claude up against ChatGPT against Gemini, (00:07:22): they’re all going to have different responses. (00:07:24): Different insights and responses and feedback. (00:07:28): So you kind of do this like almost like this collection of models. (00:07:30): You can put like Kimmy in there, you can put Manus in there, Llama in there. (00:07:35): Exactly. (00:07:35): Yeah, you’re orchestrating all these agents to sort of go test the UI. (00:07:39): Right. (00:07:40): So each, you know, think about all these different factors. (00:07:42): So you can, (00:07:43): for each model, (00:07:44): you can have thousands of personas with an avatar and a voice and you can, (00:07:50): and then each model. (00:07:51): So chat GPT 5.2 would get a model and chat GPT 5.1 will get a model and 4.0 mini (00:07:57): would get a model or would get different personalities. (00:07:59): So you actually can build this humongous panel of synthetic testers to do your tests. (00:08:04): That’s incredible. (00:08:06): What do you think it is about customer feedback that keeps drawing you back to that problem? (00:08:09): You’ve been working on this now for 10 or 15 years. (00:08:12): Yeah. (00:08:13): Well, (00:08:13): you know how George Bush kind of after the Iraq operations, (00:08:18): he was sitting on that aircraft carrier. (00:08:20): He kind of lands in the jet. (00:08:22): He kind of lands in the jet. (00:08:22): He comes out and he says, mission accomplished. (00:08:25): And after user testing, I definitely did not feel like it was mission accomplished. (00:08:30): There was a lot that we wanted to do that we never were able to. (00:08:33): able to develop. (00:08:34): And now with the advent of AI and our ability to build so quickly, (00:08:38): we can roll out, (00:08:39): I can finally build the product that I wish we could have brought to market 10 (00:08:43): years ago. (00:08:44): So that’s what really brings me back because I want to finish what I started at user testing. (00:08:49): Don’t you always think though that it’s going to be this like treadmill of it’s (00:08:52): never done, (00:08:53): right?
Chris Hicken
(00:08:53): Because every 10 years, we’re going to have a platform shift. (00:08:56): You know, 10 years ago, it was mobile and cloud. (00:08:57): 10 years from now, it’ll probably be something like quantum. (00:09:00): It’ll be like, (00:09:00): oh, (00:09:00): now that quantum is out, (00:09:02): we can go simulate like user testing across billions of realities. (00:09:06): Yeah. (00:09:07): I mean, I think I’m generally okay with that.
Brian Bell
(00:09:10): Every 10 years, I’ll disrupt what I did before and keep making me better. (00:09:13): That’s okay. (00:09:14): That’s a good founder. (00:09:15): So you’ve talked about the five-person user test being dead. (00:09:18): What does that mean and what’s changed? (00:09:20): Right. (00:09:21): First, some context for people who are listening in. (00:09:24): There’s a guy named Steve Krug who wrote a book, Don’t Make Me Think. (00:09:27): Love that book. (00:09:28): Essentially, (00:09:28): it really became... (00:09:29): Everybody who works in software should read that book before they do anything with (00:09:33): software.
Chris Hicken
(00:09:33): It’s it’s such a good book. (00:09:35): And Steve Krug is a really smart guy. (00:09:38): Essentially, (00:09:38): when he’s talking about user testing, (00:09:40): he says, (00:09:40): if you, (00:09:41): you know, (00:09:42): user test, (00:09:42): he admits user testing takes a lot of time. (00:09:45): And each one, you know, each test can be half an hour to an hour. (00:09:48): So it’s a huge investment of time. (00:09:50): But if you’re being thoughtful and asking probing questions, (00:09:53): you can uncover 80% of the usability issues across five user tests. (00:10:00): So that kind of, (00:10:02): it gives people like a small, (00:10:03): even though it would be much better if you could do 20 or 30 or 40 of these, (00:10:07): if you only have time to do a couple, (00:10:10): just do five and you’ll get 80% of the insights. (00:10:13): So it kind of asymptotes after five and you’ve gotten 80 and maybe it’s 90 at 10 (00:10:17): and it’s like 95 to 120, (00:10:17): something like that. (00:10:20): Right, exactly. (00:10:21): And basically what he says is at that point, (00:10:24): you at least know what 80% of the issues are, (00:10:26): but you don’t have enough data to know whether or not those are severe issues or (00:10:30): not. (00:10:30): So now you have to use a bit of your own context and understanding of the business. (00:10:35): to prioritize which things actually matter. (00:10:38): The reason why I think that is dead now is with AI, (00:10:43): you can actually do 100 user tests in a couple of hours.
Brian Bell
(00:10:48): And you don’t have to sit down and watch 100 videos because AI is watching the videos for you. (00:10:53): AI is summarizing what it’s learned so far. (00:10:56): and ai is making recommendations of what to do to fix the problems that come up so (00:11:00): you can say you know of a hundred out of these hundred user tests 95 of them had (00:11:05): the same problem but it wasn’t very severe but these seven people had it and it (00:11:09): completely stopped their ability to move forward so now you’ve got in a lot of ways (00:11:13): you’re marrying the the what people do a survey because they want statistical (00:11:18): significance they want enough data to be able to make decisions so you get to marry (00:11:22): the quant (00:11:23): which is the decision-making power of a large sample size with the qualitative, (00:11:26): which is the understanding of why people are doing what they’re doing. (00:11:29): So you can marry the two with AI-powered user testing or AI-powered surveys or (00:11:35): AI-powered interviews. (00:11:36): Yeah, maybe you re-named the company the AI said or something like that, but... (00:11:40): Actually, that would be a good name in a couple of years. (00:11:42): What the AI said? (00:11:43): Yeah, AI said. (00:11:44): Love that. (00:11:45): So back at user testing, I mean, you were COO during a period of rapid scale. (00:11:51): What’s something that you’re doing differently as a CEO now from what you learned there?
Chris Hicken
(00:11:55): Well, besides AI, I mean, AI is... (00:11:58): is the big thing. (00:11:59): But I’d say, (00:12:00): you know, (00:12:01): outside of AI, (00:12:02): one of the big things that user testing that we did was we were always nervous to (00:12:06): implement professional services. (00:12:08): We were nervous that if we introduced professional services at the company, (00:12:11): we would be seen as a services company and our revenue would be treated as services (00:12:17): revenue, (00:12:17): would be valued as services revenue rather than as software revenue.
Brian Bell
(00:12:20): So for the audience, explain what that is. (00:12:22): What is professional services in that context and why do people discount it? (00:12:26): Yeah. (00:12:26): Yeah. (00:12:27): That’s okay. (00:12:27): So that’s a good, that’s a good clarification. (00:12:29): So every business, most businesses have revenue that comes from the software that they sell.
Chris Hicken
(00:12:35): That’s like the annual, I’m paying 10 grand a year to use your platform kind of thing. (00:12:41): Exactly. (00:12:41): A seat license or, (00:12:42): you know, (00:12:42): actually most SAS is you’re going to buy a seat license and that’s the software. (00:12:46): That’s, (00:12:46): the thing that you log into, the web application, the software. (00:12:50): Also, you’re oftentimes buying, you’re buying professional services. (00:12:54): This could be things like helping a team will help you onboard. (00:12:58): A team will help you get started. (00:13:01): You can get access to a customer success person who you can call anytime. (00:13:07): You can pay for, (00:13:08): you know, (00:13:08): an advanced customer support plan where you get 24-7, (00:13:12): you know, (00:13:14): customer support and people will always be on call to... (00:13:17): answer your team’s questions. (00:13:20): So all of that is considered services revenue. (00:13:23): Now, (00:13:23): investors generally, (00:13:25): they will reward software revenue at a much higher multiple because software (00:13:31): revenue has a much higher gross profit margin. (00:13:34): So software revenue... (00:13:36): typically has 80 to 90% gross profit margins, (00:13:38): where oftentimes services revenue might only have a 20 to 40%. (00:13:42): Yeah, it’s like 20% gross margin. (00:13:45): Right. (00:13:45): So as a result, (00:13:46): you know, (00:13:47): most companies try to optimize their revenue to be as much as they can services. (00:13:51): Now, (00:13:52): Salesforce kind of changed the world of the mix between software and services (00:13:55): because Salesforce, (00:13:56): I mean, (00:13:57): a huge chunk of their revenue of services, (00:13:59): it’s, (00:13:59): you know, (00:13:59): can (00:14:00): It’s climbed upwards of 20% in some years. (00:14:02): And basically what Wall Street said is like, (00:14:04): look, (00:14:04): as long as you keep, (00:14:05): what we’ve learned is that public markets, (00:14:07): as long as you keep your services revenue in the 15% or less range, (00:14:12): it will reward all of the revenue as software revenue. (00:14:15): So you can actually have 85% software, (00:14:18): 15% services, (00:14:19): but that whole 100% will be valued as software. (00:14:22): Some equity analyst on Wall Street just decided that and that’s what everybody does. (00:14:25): Honestly, it was probably more companies like Salesforce who were early adopters. (00:14:30): They say, no, this is the mix and this is how you should treat it. (00:14:33): Yeah. (00:14:33): And they prove that you can add services to your business without hurting gross profit margins. (00:14:40): And it actually can be an accelerant of growth. (00:14:44): And renewals, upsells, cross-sells, all that stuff. (00:14:46): Exactly. (00:14:47): Net revenue retention, they could probably study that. (00:14:50): Where we implemented service revenue or professional pro-serve, (00:14:54): we got better retention and better net revenue, (00:14:57): net dollar retention, (00:14:59): net logo retention, (00:14:59): stuff like that. (00:15:01): Exactly, yeah. (00:15:01): So I think that, (00:15:02): so, (00:15:03): but when we were building user testing, (00:15:05): those guardrails weren’t as well established. (00:15:08): So we were really nervous about adding services to the company. (00:15:11): Even though looking back, that would have been an incredible accelerator for the business. (00:15:15): So this time around, (00:15:16): one of the things that I’m doing differently is I’m not scared at all about adding (00:15:20): professional services into the business. (00:15:21): In fact, (00:15:22): I think it can be one of our key unlocks to helping us get to this vision of push a
Brian Bell
(00:15:27): button, (00:15:27): get an insight. (00:15:28): Yeah. (00:15:28): Well, yeah, because people are hiring you to do this job. (00:15:31): It’s like the jobs to be done framework, right? (00:15:32): They’re not hiring you to give me a tool to use. (00:15:37): I want the insight so I can build a better product. (00:15:39): Yeah. (00:15:39): I don’t really care. (00:15:40): I don’t want to do the research. (00:15:41): I don’t want to do the research. (00:15:42): Just tell me what to fix. (00:15:43): Yeah, I can hire some agency for whatever five-figure cost to go do this study for me. (00:15:49): Or I can use they said and just like the AI does it and it’s as good or nearly as (00:15:55): good for a fraction of the cost. (00:15:56): Is that kind of the value prop here? (00:15:58): Exactly. (00:15:59): And it may be even more likely at a big company. (00:16:01): It’s like, I’ve got this really good research team, but they’re backlogged for months. (00:16:05): Yeah. (00:16:05): Oh, I’ll get to that in a Q3. (00:16:06): And you’re like, wait, it’s Q1. (00:16:08): Right. (00:16:09): And it’s like professional photographers still use a digital SLR because it takes (00:16:13): the best pictures. (00:16:14): But most pictures taken today, (00:16:16): the vast majority of pictures taken today are from a smartphone, (00:16:19): which are still really good. (00:16:21): Not as good as an SLR, but convenient. (00:16:23): It’s with you all the time. (00:16:25): And so... (00:16:25): Yeah, sometimes you don’t have to fly a 747 if you can take a Gulfstream, right? (00:16:29): Right, right. (00:16:31): But what I’m thinking about, I’m using this analogy to talk about ubiquity. (00:16:35): User testing has traditionally been something that’s done by professionals. (00:16:39): These are the SLR people. (00:16:41): They really know what they’re doing. (00:16:42): Super high quality output. (00:16:43): Yeah. (00:16:44): They even have a darkroom at their house. (00:16:46): Right. (00:16:46): But the future is the smartphone future where now anyone can do anyone can push a (00:16:51): button and get an insight. (00:16:52): And now there will always be a place for the professional SLR person always. (00:16:56): But a lot of people just need feedback quick. (00:16:59): And you can use a smartphone. (00:17:00): And especially the kind of the long tail of non-Fortune 1000 companies, let’s call it. (00:17:04): Exactly. (00:17:05): That’s exactly it, which is most businesses, right? (00:17:07): So yeah, (00:17:07): so I think if we deliver on our vision, (00:17:09): we will have many, (00:17:11): many, (00:17:11): many more research tests, (00:17:13): many more insights will be generated for companies in the next 15 years than in the (00:17:18): previous 15 years. (00:17:19): How does this dovetail with kind of the movement for forward deployed engineers? (00:17:23): Because that seems to be a movement right now in software as well. (00:17:25): What do you mean by forward deployed engineers?
Chris Hicken
(00:17:27): Yeah, (00:17:27): so this is more like they’re basically professional services engineers that go on (00:17:33): site and help you implement the product and things like that. (00:17:36): Seems to be like something that a lot of software companies are doing now. (00:17:39): Interesting. (00:17:39): Well, I don’t think that will be a trend for long. (00:17:43): And let me explain why. (00:17:44): AI workflows, they really are remarkable. (00:17:47): When you really get them implemented in your system, (00:17:49): you can do unbelievable software development. (00:17:53): Everything from having your designer, (00:17:55): assuming that they’ve built their design system, (00:17:57): crank out polished, (00:17:59): beautiful designs in hours using Figma Make. (00:18:03): It’s kind of like vibe coding everything. (00:18:04): We’re vibing every process in business. (00:18:08): true we’re just yeah just basically telling the ai like hey man i need a user study (00:18:12): done cool like like and it just built this whole platform yeah well okay so so you (00:18:18): go from the design in figma to now you’re using figma mcp where cursor can grab all (00:18:25): of that and convert it into front-end code basically we we’ve been able to make (00:18:30): relatively junior and it (00:18:32): I say relatively junior. (00:18:33): I mean, these are engineers with less than three years of coding experience. (00:18:37): Some of them are straight out of college or in college. (00:18:39): And they’re producing more throughput than previous engineers that I worked with (00:18:44): that didn’t have the benefit of AI by a factor of three or four. (00:18:47): They are incredible engineers. (00:18:49): This must be really hard for product managers because I came from product for years and years. (00:18:53): I was never an engineer. (00:18:55): And so I’d write a PRD and then I’d just go meet people for the next two weeks (00:18:58): while they kind of built it. (00:19:00): Yeah. (00:19:00): You know, (00:19:01): like this must be really hard because now it’s almost like this, (00:19:03): like almost a daily process. (00:19:05): Like, (00:19:05): are you guys experiencing it where like you almost have to review the output every (00:19:09): single day and user test it every single day. (00:19:11): The job is actually easier. (00:19:12): And the reason why is you can give more ownership. (00:19:16): Like before the product manager really had to think everything through. (00:19:20): because it was so expensive to use engineering time that you just couldn’t afford (00:19:24): to screw it up. (00:19:25): You don’t want to waste your time. (00:19:26): Like, make sure I’m building the right thing in the right way with all the right functionality. (00:19:29): And you better be sure, because it’s very costly. (00:19:32): Do not waste. (00:19:33): And do not roll out something that doesn’t get used. (00:19:36): Otherwise, shame on you. (00:19:37): Now, (00:19:38): you know, (00:19:38): the engineer is building so quickly that you can actually say like, (00:19:41): look, (00:19:42): this is like the general direction I want to head with this, (00:19:44): but I want it to work within the, (00:19:46): within your system. (00:19:47): So however you can build this fastest, go do that. (00:19:49): Show me, show me what you came up with and then let’s iterate. (00:19:52): So it’s like, (00:19:53): it’s way, (00:19:54): way more rapid iteration with the product manager and the engineer, (00:19:57): way less time on the, (00:19:58): on the original PRD, (00:20:00): easy fixes when engineering has to make changes. (00:20:03): So yeah. (00:20:04): And engineers are less upset when you get feedback because I remember like just grinding. (00:20:07): At one point I had five scrum teams under me and it was like, (00:20:11): you know, (00:20:12): just pulling teeth sometimes to get them to iterate on things. (00:20:16): Yeah. (00:20:17): You had this backlog of like 500 tickets that you’re prioritizing and now they’re (00:20:21): just cranking through those. (00:20:23): Yeah, I agree with all that. (00:20:24): And the engineers feel more ownership for what they’re building because they’re (00:20:29): part of the process. (00:20:30): They get to see how the product is made, the product decisions. (00:20:33): They were part of the product decisions. (00:20:34): They get to weigh in on the experience. (00:20:36): We have QA involved really early in the process. (00:20:39): So QA gets to also watch it get built so that they can test the right things. (00:20:43): So it’s just a much more collaborative team-based approach to building stuff, (00:20:47): which is rapid, (00:20:48): iterative, (00:20:49): more ownership across the board. (00:20:50): And I think it’s more fun.
Brian Bell
(00:20:52): That’s amazing. (00:20:53): And that was kind of like the next question I wanted to ask is how are product teams changing? (00:20:58): And I think we kind of nailed it, but anything else you want to add there?
Chris Hicken
(00:21:00): Well, (00:21:01): yeah, (00:21:01): let me just think of like there’s any other details that’d be worth sharing about (00:21:04): how the product teams work. (00:21:06): Well, I would say there have been tools. (00:21:09): I’m going to talk about ClickUp. (00:21:11): I sold my company to ClickUp, (00:21:12): but it’s really because I’m a believer in the vision of Nuff Said, (00:21:16): which I sold to ClickUp, (00:21:17): which is get all of your work in one place. (00:21:20): It makes a huge difference for product teams when they can have their ticketing, (00:21:24): their Slack, (00:21:25): communication, (00:21:27): their tasks. (00:21:28): their figma designs their customer research it’s all in one place so it’s not like (00:21:33): you’re not going to five different places you don’t have data silos just like all (00:21:36): the context that you need to get work done is all happening in one place so that (00:21:41): ticket is the single source of truth for everything that you need to know about the (00:21:45): project all the slack slack communication i mean click up has its own internal chat
Brian Bell
(00:21:51): it has its own commenting system it has its own (00:21:54): direct messaging, it has its own file system. (00:21:56): So all that comes in one place. (00:21:58): So I think really effective product teams are doing a good job of bringing all (00:22:02): their workflow into a single workspace, (00:22:05): which makes life so much easier for keeping everybody on the same page and working (00:22:09): quickly. (00:22:10): Yeah. (00:22:10): Yeah. (00:22:10): I love that. (00:22:12): I’d love to talk on that thread a little bit too, (00:22:15): because you went through this huge scale up, (00:22:17): huge IPO, (00:22:18): C-level exec from zero to a hundred million revenue. (00:22:22): That’s pretty amazing. (00:22:23): And then you did another company and sold it. (00:22:24): And now you’re kind of really on your third startup in a way and twice as a (00:22:28): founder, (00:22:29): once as an exec. (00:22:30): What are you taking in this like third iteration? (00:22:32): I think there’s a magic... (00:22:35): magical quality to three to the third time. (00:22:38): Like what are you kind of getting right now that you’ve kind of taken away? (00:22:41): I asked that question about user testing, (00:22:43): but I’d love to kind of know like as a, (00:22:45): you know, (00:22:45): kind of this third rodeo, (00:22:47): like what are you doing now that that’s different?
Chris Hicken
(00:22:49): From an execution standpoint, I’m not sure that there’s that much difference. (00:22:54): And what I mean by that is, (00:22:55): you know, (00:22:55): early stage company building is about just rapid nonstop testing of everything (00:23:03): because (00:23:04): You have a hypothesis and your first hypothesis is going to be wrong. (00:23:08): You already know it’s going to be wrong. (00:23:10): So just get something out to market, (00:23:12): test it, (00:23:13): iterate on it, (00:23:14): move quickly, (00:23:15): kill the stuff that’s not working. (00:23:16): Try to sell it. (00:23:17): Yeah, sell it before it exists. (00:23:19): Right. (00:23:19): That’s all part of the startup process. (00:23:20): So I think a lot of that is the same. (00:23:22): I think probably the thing that’s different is I have more patience to know that
Brian Bell
(00:23:26): like stuff isn’t it’s OK that stuff isn’t going to work initially. (00:23:30): Like in the early, like the first couple of times at it, something doesn’t work. (00:23:33): It’s like, you know, have I failed? (00:23:35): You know, am I not as good as I think I am? (00:23:38): You know, did I make the wrong choice? (00:23:41): Right. (00:23:42): Did I make the wrong choice with this business? (00:23:44): Piece of shit. (00:23:45): Maybe I should shut this down. (00:23:46): Why did I give up this high paying job to go do this startup where I’m not making any money? (00:23:50): Team Ignite every day. (00:23:51): Like I could just be at Microsoft just collecting the equity, man.
Chris Hicken
(00:23:55): And yeah, it’s like grinding. (00:23:57): Right. (00:23:57): Oh, (00:23:58): I think now, (00:23:59): now it’s like, (00:23:59): I know that that’s, (00:24:00): you know, (00:24:01): when I’m signing up for this journey,
Brian Bell
(00:24:02): I know that’s what I’m signing up for. (00:24:04): And there’s something like, (00:24:05): you know, (00:24:05): people ask Elon Musk, (00:24:06): like, (00:24:06): why do you keep doing startups? (00:24:08): Like you’ve made a ton of money. (00:24:09): You don’t need to do this anymore. (00:24:11): Like why, why keep $150, $200 billion, $300 billion, whatever it is now. (00:24:15): Yeah. (00:24:15): He’s like, it’s like six or 700 billion now. (00:24:18): And so,
Chris Hicken
(00:24:19): but I think it’s the same, (00:24:20): it’s the same thing, (00:24:20): which is like constantly on the edge of failure keeps you really sharp. (00:24:25): It keeps you thinking, it keeps, it keeps the pressure on. (00:24:28): It keeps you moving at your fastest pace. (00:24:30): It keeps you learning. (00:24:32): It keeps you experimenting. (00:24:34): And I’m just really good in that. (00:24:36): Like I’m really good in that kind of environment. (00:24:39): So love it. (00:24:40): Yeah. (00:24:40): That’s where I thrive. (00:24:41): And that’s why we invested. (00:24:42): Fair disclosure, but we’re friends, but I also invested in. (00:24:45): Yes. (00:24:45): Yeah. (00:24:46): Thank you. (00:24:46): Thank you. (00:24:46): Of course. (00:24:47): Yeah, of course.
Brian Bell
(00:24:48): So you just launched 3.0. (00:24:49): Tell us about that. (00:24:50): What’s exciting about this new launch? (00:24:54): 3.0 is us delivering a lot of features that the market was asking for. (00:24:57): So in 2.0, we made this big push to get a bunch of stuff out there. (00:25:01): We got a ton of feedback about what people liked and what they didn’t like. (00:25:04): One of the things that people really wanted was really good user testing capabilities. (00:25:08): So 3.0 included, we’d had AI,
Chris Hicken
(00:25:12): user testing. (00:25:13): We had AI surveys, AI forums, which is like a type forum, but it’s AI assisted, super cool tech. (00:25:19): Nice. (00:25:20): We added the ability to add voice anywhere in the platform. (00:25:23): So you can have users, (00:25:24): you know, (00:25:24): testers or survey takers can speak their responses rather than typing. (00:25:30): AI can talk back to them. (00:25:31): We added things like conditional logic, which, you know, gives the (00:25:35): the survey creator, the ability to branch the questions depending on how people answer. (00:25:40): We added the ability to recruit out of your own panel or recruit out of our panel. (00:25:46): We made it an international product where you can launch your test to 70 different languages. (00:25:55): So anyway, (00:25:56): And remember, we’re an AI native company. (00:25:58): So this is like three or four months worth of work. (00:26:01): So we’re delivering like these massive features. (00:26:03): Yeah, (00:26:03): I mean, (00:26:03): you just described like probably almost two or three years back in my software (00:26:07): brain. (00:26:08): Right, right, yeah. (00:26:09): Yeah, and this is now months, months of work. (00:26:12): So, (00:26:13): and we’re already working on, (00:26:14): like we’re pretty far into 4.0,
Brian Bell
(00:26:17): which is the next round of features. (00:26:19): So yeah, that’s, you know, yeah, being AI native is fun, man. (00:26:23): It’s a different world. (00:26:24): It’s a different world than we grew up in. (00:26:25): How do you, like, that was going to be my next question too. (00:26:28): Like, how do you think this whole industry changes?
Chris Hicken
(00:26:31): You know, we’ve been kind of working in SaaS. (00:26:33): I’ve been investing in SaaS now and working in it for 15 years plus. (00:26:36): And, (00:26:37): you know, (00:26:37): a lot of, (00:26:38): you know, (00:26:38): you saw the SaaS apocalypse recently where, (00:26:40): you know, (00:26:41): there’s, (00:26:41): I don’t know how many tens or hundreds of billions of the stock market value was (00:26:45): crushed. (00:26:46): I’m sure user testing got crushed in that too. (00:26:48): What’s your crystal ball over the next five or 10 years with SaaS and software and (00:26:53): AI, everything. (00:26:54): Yeah, I’d love to get your thoughts on this. (00:26:55): I have some too that I’ll share. (00:26:57): Sure. (00:26:57): Yeah, actually, I would love to hear your thoughts. (00:26:59): I don’t think anyone can predict. (00:27:00): I don’t know if you watch the All In podcast. (00:27:02): I watch to see what those guys are kind of predicting too. (00:27:04): And even they’re like... You know the one I’ve liked recently is Moonshots. (00:27:07): I’ve been listening to Moonshots a lot. (00:27:09): Yeah, really good. (00:27:09): Yeah. (00:27:11): Those guys are living in the future and talking about it in the present day. (00:27:15): Whereas All In’s become a very political and every once in a while you’ll get some (00:27:20): tech stuff, (00:27:21): but... (00:27:21): Yeah, (00:27:21): I mean,
Brian Bell
(00:27:22): they’ve tried to bring in, (00:27:23): they’ve tried to bring in more kind of AI discussion recently. (00:27:26): You know, (00:27:27): especially the beginning of the year episodes, (00:27:29): they try to make predictions for the, (00:27:30): you know, (00:27:31): for what’s going to happen in the year. (00:27:32): So there’s, there are some interesting. (00:27:35): And Sachs just want to talk about politics. (00:27:37): Yeah, they do. (00:27:38): Freeberg and Jason are like, no, let’s talk about startups and tech. (00:27:41): Yeah, that’s kind of the schism there. (00:27:43): Well, I mean, yeah, obviously Sachs has to talk politics now, right? (00:27:47): It’s like his whole book now, yeah. (00:27:48): Right, he’s part of the administration. (00:27:50): But yeah, so impossible to see how it’s going.
Chris Hicken
(00:27:53): And like every week you get like these new huge things. (00:27:55): I mean, you know, you had OpenClaw and Maltbook come out last week or two weeks ago. (00:28:01): And already the OpenClaw got acquired by OpenAI as we’re recording this. (00:28:05): Yeah, I don’t know if he acquired OpenClaw, but they certainly hired the guy who made it. (00:28:09): Well, I hear like, and I have to, I’ve just glanced at the news. (00:28:14): I have to like do your own research, but I heard that Meta offered $2 billion to bring them in. (00:28:21): And so OpenAI had to at least match that. (00:28:24): Wow. (00:28:24): So I think what we’re witnessing is probably the first unicorn exit in history.
Brian Bell
(00:28:29): A solo founded AI unicorn exit. (00:28:33): Yeah, in like seven days, in seven business days. (00:28:36): I mean, he launched it like, what, a month ago?
Chris Hicken
(00:28:40): Yeah, less than a month ago. (00:28:41): I mean, that’s like literally weeks ago. (00:28:43): He started working on it back in, he said he vibe coded over a weekend in November of 2025. (00:28:47): Yeah. (00:28:47): We’re talking about from like zero to a billion dollar exit. (00:28:52): And this is the acceleration. (00:28:54): This is where that question’s coming from is we’re just seeing an acceleration (00:28:57): across our economy, (00:28:59): across startup creation, (00:29:00): value creation, (00:29:01): everything. (00:29:02): I mean, for an investor like me, like looking at him, like he didn’t raise any money.
Brian Bell
(00:29:06): He just vibe coded something for a billion dollars. (00:29:08): Right. (00:29:09): Yeah, exactly. (00:29:09): So what even happens to me? (00:29:11): Like, okay, so here’s my take. (00:29:14): What about me? (00:29:15): Yeah. (00:29:15): Who moved my cheese? (00:29:17): So it’s, it’s really hard.
Chris Hicken
(00:29:19): It’s really easy. (00:29:20): Like there’s a lot of noise. (00:29:22): I think you even said this before. (00:29:23): There’s a lot of noise in the AI market. (00:29:25): And for me, (00:29:25): the noise is like the one guy who sold his company for a billion dollars or the (00:29:31): cursor who went from zero to a hundred million in six months or whatever. (00:29:35): Right. (00:29:35): Or the, you know, that one AI company that hit product market fit. (00:29:40): And, you know, like for me, that’s noise because that’s not most of us. (00:29:44): Like, of course, there’s going to be the outliers.
Brian Bell
(00:29:46): Let them do their thing. (00:29:47): The 0.001% of companies that are just... (00:29:51): And there’s not, there’s not much we can learn from them. (00:29:54): I mean, what do you learn from a guy who vibe coded?
Chris Hicken
(00:29:57): I mean, it’s like everybody’s vibe coding right now. (00:29:59): So it’s not like it was, (00:30:00): you know, (00:30:01): maybe what you can, (00:30:02): maybe the takeaway from that story is there’s a huge appetite in the market for AI (00:30:07): that can deeply integrate into your workflows and just do stuff for you.
Brian Bell
(00:30:11): I don’t know. (00:30:12): Was that that big of an insight?
Chris Hicken
(00:30:13): I think he was just the first person to actually do it. (00:30:15): So I kind of just took all these things that were sitting around and said, (00:30:18): Hey, (00:30:18): I can wire this up to like do stuff. (00:30:20): Right. (00:30:20): And I could do useful things. (00:30:21): And actually, (00:30:22): if you look at the things that the AI could do, (00:30:24): they’re actually not that useful, (00:30:25): actually, (00:30:26): in the grand scheme of things in terms of saving you time. (00:30:28): But it was the realization that like, (00:30:31): oh, (00:30:31): AI can do a lot of my rote, (00:30:34): you know, (00:30:34): administrative tasks that I don’t want to do. (00:30:36): So it’s kind of like the promise of where this technology could go that I think (00:30:40): everyone got really excited about. (00:30:41): Because if you look at the actual tasks that OpenBook performs for you, (00:30:45): they’re actually not that. (00:30:46): I mean, yes, of course, they’re helpful. (00:30:48): I mean, you take some stuff off your plate, but it’s not like it’s not going to go. (00:30:51): I spent five or six hours setting it up, got it to work. (00:30:54): Went in a hostinger, got my VPS set up, doing a lot of bash commands on the terminal. (00:31:00): And I’m pretty technical. (00:31:01): And I spent five or six hours trying to get it to work.
Brian Bell
(00:31:03): And I was like, this isn’t that special. (00:31:05): Yeah. (00:31:06): It’s not that helpful. (00:31:07): It’s not that helpful. (00:31:10): It’s almost like the promise of it, (00:31:11): you know, (00:31:11): because you can see like if this becomes very useful in the next year or two, (00:31:17): like all knowledge work just completely shifts, (00:31:19): right? (00:31:19): And one thing I hear Jason talk about a lot (00:31:21): is, oh, this is changing the entire equation on scaling startups. (00:31:26): I mean, are you feeling that a little bit as a founder who’s done it before? (00:31:30): I mean, I see it as an investor. (00:31:33): People are more capital efficient. (00:31:34): They could ship more code. (00:31:36): I still think there’s like a Jevons paradox or Jevons. (00:31:39): I hear people say it both ways, (00:31:40): but that will happen with software, (00:31:42): which is when the cost of producing software asymptotes is zero, (00:31:46): there’s more software than ever created and more demand for it than ever. (00:31:49): Yes, that’s generally hit. (00:31:50): Yeah. (00:31:51): So the way that I’m feeling about this is the outliers are noise. (00:31:55): I mean, there’s nothing we can learn from them. (00:31:58): There’s a lot of speculation in the market still, (00:32:00): a lot of like FOMO, (00:32:02): you know, (00:32:03): and that drives crazy valuations. (00:32:05): And so, of course, as an investor, you feel frothy, like, oh, my gosh, how do I get it? (00:32:10): How do I be a part of this? (00:32:11): Right. (00:32:11): Yeah. (00:32:11): Now I got to go to everybody with a GitHub project and start throwing money at them. (00:32:15): Yeah, (00:32:15): but for the other 999,999 companies that are building right now, (00:32:19): they’re not experiencing that wave right now. (00:32:23): And so I think the fundamentals of building business still exist, (00:32:28): which is you got to go find people with a problem. (00:32:31): You got to find people with a severe problem where they’re willing to spend money (00:32:34): to solve that problem. (00:32:35): And AI needs to be a tool to help you solve that problem more efficiently than or (00:32:40): quickly or at a lower cost than you could before. (00:32:42): Right. (00:32:43): I mean, you just described startups perfectly. (00:32:46): It’s like you’re paying time or money to solve a problem in a very painful way, (00:32:51): and we could do it better with this new product. (00:32:53): That’s it. (00:32:54): That’s it. (00:32:54): That’s basically it. (00:32:55): That’s basically it. (00:32:56): And Cursor, (00:32:57): is it worth billions of dollars because it helps engineers code 50% faster than (00:33:03): they could have before, (00:33:04): or maybe even 100% faster than they could have before? (00:33:06): Yeah, I mean, that’s certainly a valuable product. (00:33:08): There’s going to be a million copycats of that product. (00:33:10): When there already is, you got cloud code, you got code. (00:33:12): Right, right, exactly. (00:33:13): Right, right. (00:33:14): And then you’ve got to be for them, actually, I think. (00:33:16): Yeah, (00:33:17): you got this other, (00:33:18): you know, (00:33:18): this other argument that’s being made now that, (00:33:20): you know, (00:33:20): maybe all software will be (00:33:22): you know, self prompted. (00:33:24): I want my own CRM. (00:33:25): And so I’m not going to buy Salesforce. (00:33:27): My, my agent is going to keep saying, right? (00:33:30): Right. (00:33:31): Yeah. (00:33:31): And, (00:33:32): and I’m probably more in the, (00:33:33): in the camp of David Sachs, (00:33:35): at least for the next five to 10 years, (00:33:36): which is like software, (00:33:38): like I’m in it. (00:33:39): So I’m in it with AI.
Chris Hicken
(00:33:40): Yeah. (00:33:41): And yes, the initial product building can happen very, very quickly with AI. (00:33:47): AI is really bad at debugging. (00:33:49): AI is really bad at managing existing code. (00:33:52): And as your application grows, (00:33:54): I mean, (00:33:54): you know this, (00:33:54): Brian, (00:33:55): as your application grows, (00:33:56): you go from like, (00:33:59): 75 or 80% innovation time to like 5% innovation time. (00:34:03): And the other 90% goes to like keeping the product on the cutting edge, (00:34:07): fixing bugs, (00:34:09): you know, (00:34:09): making the product more performant, (00:34:11): you know, (00:34:11): adding the death knell of every SaaS company ever. (00:34:14): Well, we’re going to spend six months refactoring it. (00:34:16): Right, right. (00:34:17): But the refactoring, you need to shard your database. (00:34:20): Okay, so now you need to shard your database. (00:34:21): You need to do international hosting. (00:34:23): Like all those things are like AI is not, at some point AI will be good at these things. (00:34:28): It’s almost like a bias at play where everybody’s like, (00:34:30): oh, (00:34:30): AI is going to come in and just destroy everything. (00:34:32): It’s like, you know, like the internet didn’t do that. (00:34:34): It took 25 years for that. (00:34:36): And we’re still kind of (00:34:37): living through that and same with mobile and cloud and everything else and so I (00:34:41): think to your point it’s going to take 10 to 20 years for this all to play out but (00:34:44): I think we’re going to see a massive acceleration of you know a thousand flowers (00:34:48): blooming of software everywhere yeah I mean I think we’ll see a massive we’ll see a (00:34:52): massive number of new companies developing a large amount of software for sure (00:34:56): we’re going to see that but at the end of the day the winner is still going to be (00:35:00): the company that solves the problem that does the best job of solving the problem (00:35:06): execute better (00:35:07): It still just comes down to executing. (00:35:09): Yeah. (00:35:10): Faster, cheaper, better. (00:35:12): I mean, that’s going to be the winner ultimately. (00:35:16): And so, (00:35:17): yeah, (00:35:17): the last point that I was going to make is that even, (00:35:19): okay, (00:35:20): so we’re building an AI that can have conversations with customers. (00:35:24): And from the time that the company started until now, I have to say there’s been (00:35:31): A tiny improvement in AI’s ability off the shelf to have conversations the way that (00:35:36): a human would want to interact. (00:35:38): Like people are protective of having conversations with customers. (00:35:41): So they want a really high quality conversation with good follow-up questions. (00:35:46): I’ve been very disappointed with the quality of AI out of the box. (00:35:49): We’ve had to do an enormous amount of agentic agents working together to try to
Brian Bell
(00:35:54): figure out the best way to have a conversation with a human and make it feel more (00:35:57): human-like. (00:35:59): And even now, (00:35:59): I still, (00:36:00): even though the product is good now, (00:36:01): I still think there’s a lot of things that we can and should do to make it feel (00:36:05): even more human-like. (00:36:06): So even though we keep saying like acceleration, (00:36:08): acceleration, (00:36:08): acceleration, (00:36:09): there are lots of things where like AI has just not made big improvements at all. (00:36:13): And this is one, this is one use case. (00:36:15): Right. (00:36:15): Yeah, I love that. (00:36:16): So, I mean, speaking of big companies, user testing is a very big company. (00:36:20): How are you finding it selling against your old company? (00:36:23): That must be really interesting experience. (00:36:25): User testing, (00:36:26): we talked about this at the beginning, (00:36:27): because user testing is slow and because user testing was acquired by private (00:36:32): equity. (00:36:32): And actually, it wasn’t user testing. (00:36:34): It was user testing and three of its biggest competitors. (00:36:37): Say no one. (00:36:38): Yeah. (00:36:38): Yeah, they stopped innovating basically. (00:36:40): Yeah. (00:36:40): Right. (00:36:41): So now it’s a $350, $400 million a year business. (00:36:42): Cash count. (00:36:46): cash cow for private equity and but as a result of that is very expensive so now (00:36:51): it’s you got this slow expensive product so very few companies can afford it and (00:36:57): very few companies are moving slowly enough to have enough time to do user testing (00:37:02): the old way right so yeah so if you ask like how i’m feeling selling to the market
Chris Hicken
(00:37:07): it feels great i’ve never all the skeletons are yeah like hey i was i was the coo (00:37:12): of that company let me tell you about them you know (00:37:14): I literally know where all of the weaknesses of the company are. (00:37:19): Ask them these three questions next time you talk. (00:37:21): Yeah. (00:37:22): There’s literally no one who can sell against me in a user testing process because (00:37:27): I just know the business so well. (00:37:29): Do you feel like, putting your VC hat on for a sec, do you feel like
Brian Bell
(00:37:33): the best founders are like you that have the domain expertise or they’re kind of (00:37:38): like from outside the domain. (00:37:39): And so they don’t, they’re ignorant enough to kind of go reinvent everything. (00:37:43): Or could you be successful with both kinds of personas?
Chris Hicken
(00:37:45): I’m thinking back on the companies where I’ve made the most money investing as an investor. (00:37:49): Yeah. (00:37:50): Right. (00:37:51): Cause yeah, (00:37:51): you’ve, (00:37:51): you’ve been an angel and invested in lots of companies and yeah, (00:37:54): we’ve had some fantastic wins too. (00:37:56): I don’t think I’ve ever invested in a founder with domain expertise. (00:38:04): that had a big win. (00:38:05): I think all of my... (00:38:07): Even in my career, (00:38:08): this is the first company I’ve started where I have domain expertise. (00:38:14): Even in my own career and founders that I’ve invested in, (00:38:17): none of them had domain or they hadn’t worked at a company in the same category (00:38:23): before. (00:38:24): So I think it’s actually...
Brian Bell
(00:38:27): A good founder can identify a painful problem to start with. (00:38:33): Really, it’s like an obsession with a customer problem. (00:38:35): That’s really what matters at the end of the day. (00:38:38): The wrong founder is product obsessed and a great founder is problem obsessed. (00:38:42): And so the great founders just have this knack for seeing a problem and then just (00:38:47): ravaging it from all angles to figure out how to solve it. (00:38:50): Yeah, I love that. (00:38:51): And I’ve invested over 300 companies now. (00:38:53): I’ve definitely seen founders be successful on both sides. (00:38:56): And that kind of is the art and science of, (00:38:59): I think, (00:38:59): being an investor is figuring that out, (00:39:01): right? (00:39:02): Kind of assessing the team and trying to figure out what type of team it is. (00:39:06): I think you can come from the domain and reinvent it and vice versa. (00:39:09): You can Jeff Bezos it and just be like, I want to start a bookstore. (00:39:12): I’m from hedge funds, right? (00:39:14): Yeah. (00:39:14): Or Bill Gates starting, you know, an operating system company. (00:39:17): I mean, (00:39:18): probably what you’re saying is domain expertise is not the winning factor, (00:39:23): the predictor of whether or not someone’s going to be successful. (00:39:26): Right. (00:39:26): So speaking of AI, (00:39:28): LLMs generally that are prone to making shit up and hallucinating, (00:39:32): how do you kind of main trust and signal in these insights?
Chris Hicken
(00:39:35): Yeah. (00:39:36): Well, (00:39:36): we’ve actually done an extraordinary amount of development behind the scenes to (00:39:39): make AI not hallucinate so that you can actually trust the results. (00:39:43): So, (00:39:43): but without spilling all the beans, (00:39:45): what I can say is you can’t use, (00:39:46): you cannot use AI out of the box to look at a survey or user test and get reliable (00:39:52): results. (00:39:52): Even though, (00:39:53): you know, (00:39:54): models like, (00:39:55): you know, (00:39:55): Gemini would have, (00:39:56): you know, (00:39:56): you can put a million tokens into the context window. (00:40:00): the LLM’s ability to find the needle in the haystack is very, very small. (00:40:05): I mean, I think the hit rate is like, I don’t know, 25 or 30%. (00:40:07): So you’re going to get- It’s getting better every generation, but yeah. (00:40:11): Yeah, I think the new Claude Opus is up to like 60%. (00:40:15): But still, (00:40:15): you’re talking about like an enterprise software, (00:40:17): you can’t get it wrong 30% of the time. (00:40:19): So you have to use a combination of traditional technology plus LLMs to keep AI (00:40:25): honest on what you’re actually learning from the insights. (00:40:28): For us, (00:40:29): in terms of how do we create conversations that are more human-like in terms of (00:40:33): quality, (00:40:34): we’ve got this massive agentic AI framework where we’ve got agents that work (00:40:38): together to build a great conversation with the customer. (00:40:42): And we use different LLMs in that agentic flow, right? (00:40:45): is powered by a single LLM model. (00:40:47): It has instructions. (00:40:49): Oftentimes two or three agents are working together. (00:40:51): And then oftentimes we have a QA agent that inspects the work that was done and (00:40:56): sends it back to the group if it doesn’t like the output. (00:40:58): So that’s how we’re able to maintain, (00:41:01): you know, (00:41:01): and it’s not like it’s, (00:41:03): you know, (00:41:03): sometimes you still get a hallucination through sometimes, (00:41:07): but we’ve reduced it to a really, (00:41:08): really small percentage of the time. (00:41:10): Yeah. (00:41:10): Yeah. (00:41:10): That makes sense. (00:41:11): Yeah, (00:41:11): what I find, (00:41:12): because I use AI in all my processes at Team Ignite, (00:41:16): it’s basically me with a bunch of AI underneath me. (00:41:18): That’s Team Ignite. (00:41:20): Plus all the LPs and operators that help us. (00:41:23): But in the day-to-day running of Team Ignite, it’s basically AI. (00:41:27): And what I find is you have to basically prompt the AI multiple times in multiple (00:41:32): ways to get a good output. (00:41:35): You just can’t rely on a single shot answer. (00:41:37): So there are tools like Langchain, (00:41:40): FlowWise, (00:41:41): N8N, (00:41:42): where you can build like hardened agentic AI flows where it’s like a flow chart
Brian Bell
(00:41:47): where you have agent, (00:41:48): right? (00:41:48): And it takes an extraordinary amount of, (00:41:51): like you’ve experienced this, (00:41:53): so many revisions to get AI to reliably produce the output that you want.
Chris Hicken
(00:41:58): Yeah, now go think about that again. (00:41:59): Now go think about that again. (00:42:00): Now go think about that again, like literally. (00:42:03): Yeah, like the agentic flow, (00:42:08): that produces our conversations i think we have gone through 13 1370 iterations on (00:42:17): it and there’s probably another 1300 that need to be done before we feel like it’s (00:42:22): perfect yeah my my ai for analyzing startups is now like on version 200 something (00:42:27): yeah i basically almost iterate every day on it there’s almost like a new iteration (00:42:31): every single day on the process or the prompt or the model or (00:42:35): Some combination of the three. (00:42:36): Yeah. (00:42:37): Yeah. (00:42:37): We’re the same way. (00:42:38): And then we’re, you know, oftentimes we’re doing multiple, multiple per day. (00:42:41): So, yeah. (00:42:42): Yeah. (00:42:42): And that’s, and that’s just what’s required because AI is not good enough out of the box. (00:42:47): I mean, I think this is part of like people like us that are in it. (00:42:49): Like there, there’s people who, who see AI from the outside. (00:42:52): They maybe use chat GPT. (00:42:54): But if you’re trying to build software for the enterprise, (00:42:58): there’s no tolerance for a crappy output. (00:43:00): No, it’s got to be at least three nines, basically. (00:43:02): Yeah, it’s got to be right every time. (00:43:04): And so the amount of work that it takes to go from the chat GPT level of quality to (00:43:10): autonomous, (00:43:11): run on your own,
Brian Bell
(00:43:12): be right all the time, (00:43:13): it’s an enormous amount of work. (00:43:14): Yeah. (00:43:15): Yeah, I love that. (00:43:16): So what are some leadership lessons that you’re applying today that you picked up (00:43:21): over your last two stints? (00:43:22): Well, (00:43:23): I’m going to go back to something I said earlier, (00:43:26): because it’s been such a valuable part of me building companies. (00:43:30): At user testing, this started at user testing. (00:43:33): There were some people at the company who were incredibly productive. (00:43:36): There are other people who are so smart, (00:43:38): like I knew them, (00:43:39): and we gave everybody Wunderlich tests when they joined the company. (00:43:43): What’s the test? (00:43:44): Okay, so Wunderlich is like a general aptitude test. (00:43:48): They use it for the NFL and a couple other places. (00:43:50): It’s kind of like, (00:43:51): it’s not so much IQ, (00:43:53): but what it does is it reveals someone’s aptitude to learn new things. (00:43:57): And over time, (00:43:58): I’ve kind of learned what each score range means for someone’s ability to learn new (00:44:03): things. (00:44:03): Is there a bar that you won’t hire somebody if they take this test? (00:44:06): Do you got to get a certain score? (00:44:08): It depends. (00:44:10): Some jobs, it actually doesn’t matter as much. (00:44:13): But some jobs, it really matters. (00:44:15): But yes, for each job, there’s a threshold below. (00:44:19): If you hire this person, it’s not like they couldn’t do the job. (00:44:21): It’s going to take them a long time to learn what needs to be learned. (00:44:24): If you love them, they’ve got a great personality, great attitude. (00:44:28): Maybe they have industry expertise. (00:44:30): Maybe you still make the hire anyway. (00:44:32): But generally, it’s another factor you consider. (00:44:34): Yeah, it’s another factor. (00:44:35): But so there are some people at user testing who were so smart and holy crap, (00:44:40): they couldn’t get anything done. (00:44:41): They were so ineffective. (00:44:43): And oftentimes what I learned was, (00:44:45): you know, (00:44:45): smart people always, (00:44:47): you know, (00:44:47): they kind of grew up and they were always able to handle everything in their head (00:44:50): because, (00:44:50): you know, (00:44:51): great. (00:44:51): They’ve got great memory. (00:44:52): They can always remember. (00:44:53): But, (00:44:54): you know, (00:44:54): when you start working at a high paced startup, (00:44:56): you’ve got like 30 things you need to contact switching. (00:44:59): You got 30 different things going on. (00:45:00): And yeah. (00:45:01): So the thing that was different between the high productivity and the low (00:45:05): productivity people, (00:45:06): regardless of their Wunderlich score, (00:45:09): was ability to organize all of their work into a single list. (00:45:13): So they weren’t looking at email, (00:45:16): like some people like on their Gmail, (00:45:18): they would leave the messages unread that they had to get back to, (00:45:21): or they put a star next to it. (00:45:22): Now, (00:45:23): the high productivity people took any tasks that they needed, (00:45:26): put it into a task list, (00:45:27): archived their inbox so they never have to look at it. (00:45:30): Their tasks were cleaned and organized. (00:45:33): They didn’t leave to-dos on their calendar. (00:45:35): They didn’t leave to-dos in Slack with reminders and all that. (00:45:39): They just had one place where they could say, (00:45:41): okay, (00:45:42): this is the universe of everything I need to do. (00:45:44): So for me, that is an email. (00:45:45): Like I put my entire email. (00:45:47): Yeah. (00:45:48): Right. (00:45:48): So actually some people do use email as their primary. (00:45:51): My task is how I prioritize is how I folder everything into like do now, (00:45:56): do later, (00:45:57): review later, (00:45:58): review now. (00:45:58): Like everything’s like in my email. (00:46:00): Yes. (00:46:01): People like you end up being the most productive because they have intentionally (00:46:06): said, (00:46:07): I’m going to make one place my single source of truth. (00:46:09): Mm-hmm. (00:46:10): And I’m going to get all of my shit into that one place. (00:46:12): And that will let me see every day when I come into the office, (00:46:14): I can see the universe of everything I need to pay attention to. (00:46:17): That makes it way easier to spot the stuff that’s urgent. (00:46:20): Way easier to set aside the stuff that doesn’t need attention right now. (00:46:23): And it lets you kind of focus and prioritize and work quickly on those high impact activities. (00:46:28): Plus, like for me personally, I have to get to inbox zero because I’m a VC. (00:46:32): So there’s like, (00:46:33): oh, (00:46:33): I’m waiting on this doc for the series A. (00:46:35): And like, (00:46:36): hey, (00:46:37): man, (00:46:37): I need this doc signed. (00:46:39): Right, exactly. (00:46:39): I have to get to inbox zero. (00:46:40): I have to get through everything. (00:46:42): And I only do it once a week now. (00:46:43): I used to do it every day. (00:46:44): But like every week I get to inbox zero. (00:46:46): I used to be able to do it every day. (00:46:47): I can’t anymore, but. (00:46:49): In a lot of ways, (00:46:50): what I found is consistently the very smartest people never had to put in place an (00:46:57): organization system because it never mattered for them because they were smart (00:47:00): enough to keep it all in their brain. (00:47:02): Once they got into a high-paced environment, everything that they had learned didn’t work. (00:47:05): How do you filter what kind of interview questions? (00:47:08): Is there a test? (00:47:09): How do you filter for people like that? (00:47:11): Well, I don’t. (00:47:11): I still want to hire the smartest people. (00:47:13): I just teach them how to manage their work better. (00:47:16): So one of the things that, (00:47:17): you know, (00:47:18): they said one of the things we do as a forcing function is we just get all the work (00:47:22): in one place. (00:47:23): So you don’t even have a choice to be going off into these other places because if (00:47:27): you want to work at the company, (00:47:29): you have to be in ClickUp where all the work and all the context is one place. (00:47:31): So what’s your system for that now? (00:47:33): Like, is there a software package you use or? (00:47:35): Yeah, it’s just ClickUp. (00:47:37): So ClickUp comes bundled with tasks, project management, calendar, chat, like Slack-based chat.
Chris Hicken
(00:47:44): You can do emails. (00:47:45): You can do many emails from your ClickUp. (00:47:49): Loom videos are all built into ClickUp. (00:47:51): Reporting and dashboards. (00:47:53): You can have run AI agents within it. (00:47:55): You can have automations. (00:47:56): So we do all of our... (00:47:57): product and engineering ticketing. (00:47:58): We do all of our CRM in ClickUp. (00:48:00): We do all of our HR. (00:48:02): So all those systems, rather than having like separate systems for all that stuff. (00:48:05): Nope. (00:48:06): It’s a piece of like business operation software, basically. (00:48:09): It does everything that you, as a knowledge worker, you could possibly need to do is ClickUp. (00:48:15): Yeah. (00:48:15): So as a forcing function, (00:48:16): it gets everyone working together in the same place, (00:48:19): context in one place. (00:48:21): It’s almost like it’s hard to work out of another system because so like all the work is there. (00:48:26): So if you want access to the context, you go to where the work is. (00:48:31): I got to check out ClickUp, I guess. (00:48:32): Because right now we’re like... You definitely got to check out ClickUp. (00:48:35): Yeah. (00:48:35): Yeah. (00:48:35): I got to check it out. (00:48:36): Because right now I’m in so many different systems like Airtable for one thing, (00:48:40): Google Docs for another thing, (00:48:41): email for everything else. (00:48:42): And obviously got Calendar and got a CRM. (00:48:46): Yeah. (00:48:46): It just makes you less efficient because now you have to scan every day when you come in. (00:48:50): And during the course of the day,
Brian Bell
(00:48:51): you have to be scanning five different systems to figure out what you should be (00:48:55): working on. (00:48:55): Right. (00:48:56): Yeah. (00:48:56): So how do you how do you coach your team to kind of stay focused in a noisy, (00:49:01): fast moving AI startup and market? (00:49:04): Yeah, well, I think this goes back to like, how do you filter the noise from the signal?
Chris Hicken
(00:49:08): And what I’m often telling the team is that, (00:49:11): you know, (00:49:12): there’s a lot of debates online, (00:49:13): you know, (00:49:13): optimist versus pessimist debate of where AI is going. (00:49:17): Don’t even engage with that. (00:49:18): It’s all noise. (00:49:19): No one can predict what’s happening. (00:49:20): It’s just for clicks. (00:49:21): You know, the (00:49:22): The 1% of 1% companies that have breakout success, (00:49:26): don’t pay attention because there’s nothing we can learn from that story. (00:49:29): But what I do do is I watch, (00:49:31): I actively encourage everyone to be experimenting with new technologies as they (00:49:35): come out. (00:49:36): So as a company, we’re (00:49:37): when something new comes out and if it looks really promising, (00:49:40): we jump on it and try to see if it adds value. (00:49:44): And I gotta be honest, a lot of times, a lot of these AI tools add almost no value. (00:49:47): So, but we’re always experimenting. (00:49:49): So what I’d say is like, (00:49:50): it’s actually critical as a CEO that you encourage the team to spend some portion (00:49:54): of every week (00:49:55): experimenting with new tools because one out of 10 is really good and it’s worth (00:50:00): adopting into the into the way that you build stuff so yeah so so how do we how do (00:50:04): we keep the team up to date i largely me and the cto are focused we’re scanning the (00:50:09): world for the stuff that’s actually valuable we filter it down to the team and we (00:50:13): encourage regular testing of new technologies as they become available but we kill
Brian Bell
(00:50:18): stuff really quick that isn’t working (00:50:19): Yeah, that makes sense. (00:50:20): Well, let’s wrap up with some rapid fire questions. (00:50:23): Okay. (00:50:23): First startup you ever worked on. (00:50:25): All right. (00:50:26): The short version is I worked for a company called headsets.com. (00:50:30): This was in the early 2000s, right? (00:50:31): So there was no like, there’s no killer app yet for the internet, right? (00:50:35): Amazon had some, I don’t even think Google had gone public yet. (00:50:38): Amazon has some good early success. (00:50:40): So in my mind, (00:50:41): I was thinking, (00:50:41): okay, (00:50:42): like maybe internet retail is the killer app for the internet. (00:50:46): So I joined headsets.com. (00:50:48): We sold telephone headsets online and it grew from a million to 40 million in sales
Chris Hicken
(00:50:55): in like four years. (00:50:57): For online retailer, that’s like unbelievable growth. (00:51:00): That’s a lot, yeah. (00:51:01): And then we got killed by Amazon. (00:51:03): So that was that story. (00:51:05): It was a really good business for four or five years. (00:51:07): And then it became really hard to scale once Amazon realized that they could sell (00:51:12): our products and go directly to our manufacturers and sell our products instead of
Brian Bell
(00:51:17): us selling them. (00:51:18): Why do you think you had that outcome of that company versus like a Zappos or (00:51:21): diapers.com or name your favorite category killer that Amazon had to acquire? (00:51:25): Did Zappos just get so big that Amazon couldn’t compete and it was just easier to buy? (00:51:30): Totally. (00:51:30): Yeah. (00:51:31): Because if you think about it, Amazon did not acquire that many, that many retailers. (00:51:34): Zappos was one of the new standouts. (00:51:36): And Tony Hsieh, I knew Tony Hsieh. (00:51:38): And he, (00:51:39): one of the things that was really special about him was his like obsessive focus on (00:51:43): creating a great customer experience. (00:51:45): And that aligned really well with Jeff Bezos’s model. (00:51:48): So like the fact that they had scaled so quickly, (00:51:50): you know, (00:51:51): they’re at a hundred plus, (00:51:52): maybe 150 million by the time they were acquired. (00:51:55): And that aligns so well with Amazon’s model for selling online that I think that was a good. (00:52:02): What’s your favorite metric that actually matters to a CEO?
Chris Hicken
(00:52:06): I created a score called CPV, which is a product market fit score. (00:52:11): And it has four components, problem, product, price, and people. (00:52:17): And actually, (00:52:18): if you look around, (00:52:19): it actually aligns pretty well with some of the other kind of Sean Ellis and some (00:52:23): of the other guys that I’ve (00:52:24): done a lot of thinking about product market fit. (00:52:25): But essentially, (00:52:26): it’s the score that we, (00:52:28): like every one of our sales deals is scored on this product market fit chart.
Brian Bell
(00:52:32): What that allows us to do is assess over time whether the business is getting (00:52:37): closer to or farther from product market fit. (00:52:39): And if we’re moving farther, why? (00:52:41): What part of our calculus is wrong? (00:52:44): Are we solving the wrong problem for somebody? (00:52:46): And if so, who was that person so we know what was wrong about it? (00:52:50): What, you know, product, does the product solve the problem that they wanted to solve? (00:52:55): And if not, what was missing pricing? (00:52:57): Can they justify the price? (00:52:58): Like even if the product is good and the problem was there, (00:53:02): maybe the price is too high to justify the solution. (00:53:06): And then people, (00:53:06): you know, (00:53:07): so are we, (00:53:08): are we speaking to the right people and are we offering the right people to help (00:53:11): them buy? (00:53:12): So I call it the CPV score and it’s, (00:53:14): it’s our number one tool that we use to iterate on product market fit. (00:53:17): Love that. (00:53:18): Never heard that. (00:53:18): So yeah, you made it up and maybe write a blog on it and I’ll share it with some of my founders. (00:53:23): It’s really cool. (00:53:23): What’s harder to scale, product market fit or culture? (00:53:26): Both are really hard. (00:53:27): I would say product market fit is harder. (00:53:29): Most businesses fail. (00:53:31): You know, (00:53:32): something like, (00:53:32): I don’t remember the exact stats, (00:53:34): but something like 80% of businesses fail by the time they get to year five. (00:53:38): And of the remaining ones, almost nobody makes it over a million in sales. (00:53:41): And the reason is we typically invest in companies that are post-launch, post-revenue, right? (00:53:46): I find that the mortality rate going from zero to even a quarter million in (00:53:50): revenue, (00:53:51): like you drop the mortality rate by like 80%. (00:53:54): Totally. (00:53:55): Because at least at that point, they’ve gotten some product market fit. (00:53:57): Yeah. (00:53:58): You can kind of like, you almost have infinite runway, I think, at that point. (00:54:01): Because you can kind of just go back to the founders and maybe one employee and (00:54:05): kind of grind it out at that point. (00:54:06): Keep grinding it. (00:54:07): Yeah. (00:54:08): So product market fit is just incredibly difficult to get right. (00:54:13): And even once you get it right, you know, how do you take it from 250 to a million? (00:54:17): Yeah. (00:54:18): Like for those companies that quickly get to 10 and then can’t scale beyond 10, right? (00:54:22): You hear a lot of those stories. (00:54:23): A lot of these like hot startups that like never make it. (00:54:25): Well, they run into a wall. (00:54:26): They think like, oh, I can do no wrong. (00:54:28): I’m going to turn burn up to three, (00:54:30): four, (00:54:31): 500,000 a month because I just raised a 10 million A, (00:54:34): right? (00:54:34): Right. (00:54:35): Or 10 million B, whatever it is. (00:54:36): And then they just go. (00:54:37): Right. (00:54:37): And the thing that they didn’t have right yet is they didn’t have product market (00:54:40): fit in a big enough category yet. (00:54:43): And that process, like you never actually have product market fit forever. (00:54:48): You have product market fit for, it’s like a slice in time. (00:54:52): Yeah, the initial subcategory, subvertical, submarket. (00:54:57): And then you have to keep expanding that pie and you’re always kind of on the edge (00:55:01): of no product market fit. (00:55:02): That’s exactly it. (00:55:03): And it’s so hard to get that right over time. (00:55:06): Yes. (00:55:06): By the way, (00:55:07): I don’t want to diminish the value of culture because that’s also super,
Chris Hicken
(00:55:10): super critical and very hard to maintain over time. (00:55:13): But the product market fit is like it’s existential. (00:55:15): Like you have to figure out a path to continue to iterate on product market fit. (00:55:19): And if you don’t have a way, (00:55:21): like a structured way in your mind to do that, (00:55:23): you as a CEO, (00:55:24): you significantly increase the chance of the business failing or plateauing. (00:55:29): And so I think for that reason, I’d have to say product market fits harder.
Brian Bell
(00:55:32): Yeah, (00:55:32): you know what’s interesting about this is I’m thinking about like kind of like (00:55:34): restaurants that go from one to two or three locations and they end up failing. (00:55:40): Yeah. (00:55:40): It’s kind of the same thing with startups. (00:55:42): Like you went from this like category and you got some product market fit there. (00:55:46): Now you’re in the second category. (00:55:48): or the second vertical, trying to grow revenue. (00:55:50): And that’s where a lot of people and founders and founding teams kind of trip up. (00:55:56): How do you prevent that? (00:55:57): Is it just putting the right structure in place where, (00:55:59): okay, (00:56:00): now you have a new team that has a product marketing manager and an account team (00:56:05): and a sales team. (00:56:06): And you literally have to set up almost like a new sub-business every time you’re (00:56:10): going into a new market. (00:56:11): How do you tackle that? (00:56:12): I mean, that’s probably a pretty healthy way to think about it. (00:56:15): I think from a corporate governance and fiscal responsibility perspective, (00:56:21): you don’t pump money into the new market until you’ve proven product market fit. (00:56:27): And that’s where a lot of companies go wrong. (00:56:28): It’s like, well, it worked in this category, so it’s going to work in this category. (00:56:32): Boom, run full speed ahead. (00:56:33): Yeah. (00:56:34): Yeah, (00:56:34): let’s let’s let’s pump tons of money into sales and marketing and go, (00:56:36): you know, (00:56:37): conquer that category before the company is really understood. (00:56:40): What’s the problem? (00:56:42): How severe is the problem for this new buyer score? (00:56:45): Right. (00:56:45): Yeah, right. (00:56:46): Exactly. (00:56:47): In your framework. (00:56:48): Yeah, exactly. (00:56:49): How good is the product at solving their problem and what’s missing? (00:56:52): How do people feel about the price? (00:56:54): Can they justify the price? (00:56:55): Can they get access to the budget needed to pay for the solution? (00:56:59): And then are the people that we hired before to sell before, (00:57:02): are they the right people to sell to the new buyer? (00:57:05): Because it could be that the people that you had before are actually not great (00:57:08): people to market to and sell to and service the new buyer. (00:57:12): So yeah, long story short, we... (00:57:16): We as investors, (00:57:17): and I’ve done this too, (00:57:18): we get excited that an early sign of fit, (00:57:22): we believe that it can expand into other categories without actually testing it (00:57:25): first. (00:57:26): And we invest too early into it before proving that the product market fit is there. (00:57:30): Yeah, I love that. (00:57:31): Really, really good insights. (00:57:32): What’s a founder you like to study or people should study?
Chris Hicken
(00:57:35): I mean, (00:57:35): just as a general rule, (00:57:36): I don’t study people who are successful because it’s, (00:57:40): you know, (00:57:40): I’ve even done this myself in some points. (00:57:42): Like once you’re successful, (00:57:44): you kind of retell history to help make yourself sound really smart. (00:57:49): I think we’re all guilty of that.
Brian Bell
(00:57:50): Like once you have some success, (00:57:51): you can kind of look back and say like, (00:57:52): wow, (00:57:52): man, (00:57:53): I made some great decisions. (00:57:54): Let me walk you through those decisions I made. (00:57:55): Even though at the time you were just like testing a lot of shit and hope something worked. (00:57:59): So, (00:58:00): and you know, (00:58:00): a lot of it was just like brute force effort rather than anything smart that you (00:58:05): did. (00:58:05): So I’m much more interested in talking to founders who experimented with a lot of (00:58:10): different things and asking them what they needed. (00:58:13): If they could go back in time, (00:58:15): what they would have done differently, (00:58:16): what they think they did wrong and what they learned. (00:58:20): I think that’s a much more interesting, (00:58:21): like from a founder perspective, (00:58:23): it’s much more interesting to learn from those folks. (00:58:25): There’s nothing I can learn from Jeff Bezos at this point, right? (00:58:27): There’s nothing I can learn from Bill Gates. (00:58:29): They’ve rewritten their story a hundred times and who knows what made them (00:58:34): successful in the early days. (00:58:35): They don’t even know what made them successful. (00:58:37): Yeah. (00:58:37): And I think a lot of it’s just luck and random, like right place, right time, right idea. (00:58:42): And just everything kind of pinballs down into like the Monte Carlo simulation that is reality. (00:58:48): Yeah. (00:58:48): Yeah. (00:58:49): I don’t want to take anything away from a successful founder. (00:58:51): I mean, people work really hard as founders. (00:58:53): You know what it’s like. (00:58:54): It’s brutal. (00:58:54): I mean, it’s just like nonstop work and you make all these sacrifices. (00:58:58): So like, yeah, good for them for their success. (00:59:00): But in terms of learning, I’d rather learn from people who failed. (00:59:02): Love that. (00:59:03): What’s the worst advice you hear in SaaS today? (00:59:06): The worst advice, you know, as a general, okay, so again, I’m speaking in generalities. (00:59:11): Whatever the market or my investors are telling me to do is generally the opposite (00:59:17): of what I want to do. (00:59:18): It’s like the same thing as Warren Buffett’s, (00:59:20): like investors tend to, (00:59:22): they bunch, (00:59:24): they all kind of swarm around an idea. (00:59:27): And I always stop when it happens. (00:59:30): I stop and think, (00:59:30): okay, (00:59:31): what’s the opposite of that idea that I could do right now that would be give me an (00:59:35): unnatural advantage compared to everyone else who’s listening to their VCs right (00:59:38): now. (00:59:38): But the opposite is true too. (00:59:40): Like when everyone’s saying pour on the gasoline, (00:59:41): I’m like, (00:59:42): okay, (00:59:43): what’s the frothy thing that I want to avoid? (00:59:46): It’s going to be, you know, overcrowded. (00:59:47): That’s why I love working in pre-seed and being a generalist investor. (00:59:50): Because I could have done like an AI fund. (00:59:52): I think we talked about that. (00:59:53): Because I led AI, built a bunch of AI, I led AI at Amazon. (00:59:57): And I definitely have all the chops and pedigree to do that. (01:00:00): But I also knew that if I did that, I would be pigeonholing myself into suboptimal returns. (01:00:06): Because I’m not completely confident I can get into all the best and hottest AI companies. (01:00:10): And... (01:00:11): Being a generalist, I can find value. (01:00:13): It’s almost like being Warren Buffett, right?
Chris Hicken
(01:00:14): Warren Buffett is sitting right now as we speak like on $330 billion in cash (01:00:19): because everything’s just like in his mind at Berkshire, (01:00:23): like everything’s just too frothy. (01:00:24): There’s nothing valuable to buy. (01:00:26): I mean, (01:00:26): I want to add something to this, (01:00:27): which is that you are remarkable at getting in front of deals though. (01:00:31): Like I’ve never seen anybody get access to so many deals compared to you. (01:00:35): I mean, you’re, you’re like, I don’t know how, honestly, I don’t know how you do it. (01:00:38): It’s like one of your secret weapons,
Brian Bell
(01:00:39): but you, (01:00:40): you’ve gotten access to some amazing deals that I haven’t been able to find anyone (01:00:45): else. (01:00:45): Like I’ve asked, like, how do I get into these early stage deals? (01:00:48): Like who, like, how do you do this? (01:00:49): And you’ve cracked the code. (01:00:50): So kudos to you for figuring that out. (01:00:52): And I add a lot of value, right?
Chris Hicken
(01:00:53): Yeah. (01:00:54): Team Ignite, I think, is a competitive advantage. (01:00:57): People always say, oh, I have a great network and my network’s my value. (01:01:01): It really is though. (01:01:02): The way we build Team Ignite is with 300 plus portfolio companies doing the (01:01:06): podcast, (01:01:07): adding value, (01:01:08): making intros. (01:01:10): If you have somebody that you know is raising, you’ll introduce them to me. (01:01:12): Because you’re like, oh, Brian’s great. (01:01:14): He made a really quick decision. (01:01:15): He has a lot of value. (01:01:16): And that word kind of spreads. (01:01:18): And then I just get into lots of deals because of it. (01:01:21): And I just make very quick decisions. (01:01:23): Yeah. (01:01:24): Well, it’s working. (01:01:25): All I’m saying is it’s working. (01:01:27): I’ve raised money from, you know, some of the top VCs in the world. (01:01:31): And I’ve never seen anyone like you. (01:01:34): I mean, (01:01:34): the kind of network and access that you have at your disposal is like unbelievable, (01:01:39): like truly remarkable. (01:01:41): You know, my wife, Lee Hong, we work together here. (01:01:42): And sometimes I’m like, how is Brian getting access to these deals? (01:01:45): It’s crazy. (01:01:47): So anyway, good for you, man. (01:01:49): It’s a solid network. (01:01:49): And then it’s like, everything’s AI powered. (01:01:51): Yeah. (01:01:52): I can move very quickly. (01:01:53): Yeah. (01:01:53): I can, I can process 10,000 pitch decks a year. (01:01:57): Right. (01:01:57): With a thousand of them and invest in a hundred of them. (01:01:59): That’s kind of like the formula. (01:02:01): Great. (01:02:01): And then everything else that we do is like, it’s AI powered with, with a, (01:02:06): you know, like a part-time contractor person. (01:02:08): I can get these people with master’s degrees in the Philippines. (01:02:12): Yeah. (01:02:12): And then I give them the AI system, the prompt to follow. (01:02:14): Yeah. (01:02:15): And so we’re processing, (01:02:16): you know, (01:02:16): we have, (01:02:17): you know, (01:02:17): two or 300 updates coming in from all our founders every month. (01:02:20): And it’s kind of like, it’s a lot to process. (01:02:22): And so, (01:02:23): you know, (01:02:23): the contractor will look, (01:02:24): our portfolio manager looks through everything,
Brian Bell
(01:02:26): processes everything, (01:02:27): and then basically offers help. (01:02:29): Like, (01:02:29): well, (01:02:29): I see like X, (01:02:30): Y, (01:02:30): and Z. (01:02:31): And you put everything we can do into the context window of AI at midnight. (01:02:34): Yeah, yeah. (01:02:35): Yeah. (01:02:36): And it’s just like, what do you really need from me? (01:02:38): You really need like intros. (01:02:39): Yeah. (01:02:40): Intros. (01:02:41): Fundraising intros, customer intros, and talent intros, right?
Chris Hicken
(01:02:45): Me advising you on how to like run your company, (01:02:48): like you know how to do that way better than I do. (01:02:50): In fact, (01:02:51): if I had a founder struggling with some of the stuff we’ve talked about in the
Brian Bell
(01:02:53): podcast, (01:02:53): I’d be like, (01:02:54): go talk to Chris. (01:02:55): Yeah. (01:02:55): Like he’s been there and done that. (01:02:56): And like, (01:02:57): oh, (01:02:57): go watch, (01:02:57): by the way, (01:02:58): I interviewed him on my podcast and you can go listen to the podcast and learn a (01:03:02): ton just by like hearing his experiences. (01:03:05): That’s another way to add leverage in Team Ignite. (01:03:07): Yeah. (01:03:08): And you do. (01:03:08): You add a lot of value to us as a team and as a founder, (01:03:12): but also as an investor too,
Chris Hicken
(01:03:13): like as an LP. (01:03:14): I mean, just like, it’s a great, your gift of networking. (01:03:17): I mean, it’s unbelievable.
Brian Bell
(01:03:19): So yeah, appreciate you for that. (01:03:20): Gosh, I’ll pay you later for all the kind words, but... (01:03:23): I take payment in the form of whiskey. (01:03:25): Exactly. (01:03:26): What’s a book or framework you revisit regularly? (01:03:29): Okay, let’s go back to what I said earlier. (01:03:33): Steve Krug’s Don’t Make Me Think. (01:03:34): That’s a must read. (01:03:35): I three or four times a year go back and reread Jeff Bezos’s. (01:03:39): He has a lot of really great letters to investors, but his 2016 one I thought was his best ever. (01:03:44): Really? (01:03:46): just such a good like and this was back when he was like hardcore operating at the (01:03:50): top of his game he hadn’t quite started retiring out in 2019 when he started kind (01:03:56): of pulling back that was still this is like peak bezos yeah this peak bezos yeah (01:04:00): like really really good stuff but i mean his 2015 letters 2017 letter 2014 letter (01:04:04): they were all really good but i need to go back and read all his letters yeah yeah (01:04:09): now they’re just really exceptional like good insight into like a you know (01:04:13): someone who’s thinking big picture about the world, (01:04:15): you know, (01:04:15): and thinking in terms of structure and value-based thinking. (01:04:20): Why did he retire? (01:04:21): You talk about Elon not retiring because he just like loves the thrill of first (01:04:25): principle value creation and... (01:04:28): Yeah, but I mean, did he really retire?
Chris Hicken
(01:04:30): I mean, he’s got like four or five new... Yeah, he’s in a lot of stuff right now. (01:04:35): So I don’t think he’s actually retired. (01:04:36): I think he’s actually involved in a lot of stuff right now.
Brian Bell
(01:04:39): And maybe Amazon got too big to kind of do day one kind of big picture kind of stuff. (01:04:44): Yeah, too hard to do the stuff that he probably loves doing, which is building something new. (01:04:49): Because he’s in AI now. (01:04:51): He’s in... And Amazon has a big AGI push right now. (01:04:55): He’s in... (01:04:57): Blue Origin, he’s invested in a ton of startups. (01:05:00): So yeah, I think he’s still in the game. (01:05:02): Nice. (01:05:03): And then is there anything else? (01:05:06): Last one is Crossing the Chasm. (01:05:08): It’s a really good product market fit book. (01:05:10): I’m sure most people who are listening have read it, (01:05:12): but definitely the, (01:05:13): you know, (01:05:14): understanding the difference between a technology enthusiast, (01:05:17): an early adopter, (01:05:19): and then the chasm between the early adopter and the early majority is such a (01:05:23): valuable concept when you’re thinking about product market fit. (01:05:25): I just, I come back to that a lot. (01:05:27): Yeah. (01:05:27): And, (01:05:27): you know, (01:05:27): back to our kind of portfolio thinking of expanding to new verticals and (01:05:31): categories, (01:05:31): you’re basically crossing the chasm over and over again. (01:05:34): Exactly. (01:05:35): That’s it. (01:05:35): Yeah. (01:05:35): Each time you go into a new segment, it’s crossing the chasm. (01:05:38): Yeah. (01:05:38): You kind of have to figure out where, (01:05:39): you know, (01:05:40): are we in the early majority phase in this new segment or category? (01:05:43): Or are we kind of like in the early adopter segment, kind of knowing where the market is? (01:05:48): Yeah. (01:05:48): And a lot of times you’re like, yeah, you haven’t crossed the chasm. (01:05:51): Most of the time you haven’t crossed the chasm. (01:05:52): Even if it’s something like at user testing, we sold user testing to SMBs in mid-market. (01:05:57): And now we want to sell them to enterprise. (01:05:59): And the way that enterprises buy and the way that they do user testing and the way (01:06:03): that they buy software products, (01:06:05): completely different. (01:06:06): I mean, it was actually brutal for us to cross the chasm at user testing. (01:06:10): Yeah. (01:06:11): Which B2B company do you admire most right now?
Chris Hicken
(01:06:13): Okay, so my wife and I have been, our guilty pleasure is we play a game called Clash of Clans. (01:06:18): It’s run by a company called Supercell. (01:06:20): Yeah, I know the head of VC there. (01:06:22): He was at Club Ignite in November. (01:06:25): You guys probably got the invite to that. (01:06:26): That was really fun. (01:06:27): Awesome. (01:06:29): So they were acquired by a Chinese company, (01:06:33): but the operations are still largely run out of Scandinavia. (01:06:35): And they’re just really good. (01:06:37): I really admire their community building. (01:06:40): They just do such a good job of gathering feedback from the community, (01:06:45): looking at metrics and coming up with like (01:06:47): super clever product solutions to the problems that have come up from the community (01:06:52): it’s just like very very all my kids play that yeah all three of them and i haven’t (01:06:56): jumped on it yet but maybe i’ll jump on and play with you well as an adult as an (01:07:00): adult it’s great because you can play for like three minutes at a time and yeah (01:07:03): it’s like a quick match and right quick match and you’re done right and then you (01:07:06): don’t have to then you don’t have to think about it again i like that so yeah it’s (01:07:09): actually it’s great (01:07:10): My main game is Escape from Tarkov, (01:07:11): which I think we’ve talked about, (01:07:12): which is like, (01:07:13): you know, (01:07:14): my friend and I will jump on at 10 PM at night and it’s like, (01:07:16): oh crap, (01:07:17): we’ve played one match and it’s 11. (01:07:20): Like it’s such a time commit. (01:07:22): Yeah, yeah, exactly. (01:07:23): And we love games like Rocket League where you can kind of jump in and play a match (01:07:26): and it’s five to seven minute and you can always jump off. (01:07:31): Yeah, well, in college, I played a game called Unreal Tournament.
Brian Bell
(01:07:35): I was ranked number three globally. (01:07:38): So I was a really good player. (01:07:39): So I love games. (01:07:40): I just, you know. (01:07:41): You don’t have time. (01:07:42): With kids, there’s no time. (01:07:43): There’s no time, yeah. (01:07:44): No, I get like, sometimes like, you know, because I work seven days a week. (01:07:48): I basically work a seven, seven. (01:07:50): seven or something like that yeah whatever it’s not a nine what is it nine nine six (01:07:55): is it nine nine six that’s the Chinese one yeah I’m like I’m like a nine five seven (01:07:59): yeah so I basically work every day and so like I I’m able to kind of like get into (01:08:04): this rhythm of every day I try to live like I’m retired if that makes sense (01:08:09): Yeah. (01:08:09): Great. (01:08:10): I try to always live like I’m a billionaire already. (01:08:12): Like, what would I do today if I was a billionaire? (01:08:14): Yeah, I’d meet some founders. (01:08:15): I’d do the podcast and like, you know, talk to some LPs and like, oh, I play video games. (01:08:19): Sometimes my kids, I kind of try to live almost every day is the same for me.
Chris Hicken
(01:08:24): That’s good, man. (01:08:25): I mean, you kind of live in the life. (01:08:26): Yeah. (01:08:27): You’ve got this great house and a great place in California. (01:08:29): I mean, you got the dream life. (01:08:31): No, (01:08:31): it’s like I’m living like I’m just, (01:08:33): I would actually pull back on Team Ignite a little bit if I, (01:08:35): you know, (01:08:36): was already successful at it. (01:08:38): Like, and I had, you know, 100 million in the bank. (01:08:40): I’d still do it though. (01:08:41): Yeah. (01:08:41): Oh, you absolutely would. (01:08:42): Humans and AI doing everything else for me. (01:08:45): And I’m in this journey of automating everything about Team Ignite. (01:08:48): So like, I just show up and all I have to do is like three things in a day. (01:08:52): meet a founder or two or three or four or five, (01:08:55): whatever, (01:08:55): record one podcast like I’m doing now, (01:08:57): and then meet an LP or two. (01:08:58): Yeah. (01:08:59): And everything else is like just taken care of. (01:09:02): Yeah. (01:09:02): But we’re builders. (01:09:03): I mean, you know, we, a lot of our purpose comes from adding value to the world.
Brian Bell
(01:09:07): So if we weren’t doing that, what would we spend our time on? (01:09:10): Right. (01:09:11): I mean, reading books. (01:09:12): I mean, who’s, who wants to do that?
Chris Hicken
(01:09:13): Let’s, let’s, let’s build stuff. (01:09:15): You know, I think we’re both from that. (01:09:16): It’s fun. (01:09:17): It’s actually, (01:09:17): it’s, (01:09:18): it’s this vision I have for Team Ignite, (01:09:19): which is like, (01:09:20): everything is like an AI and a process. (01:09:23): And there’s like this system where it just runs. (01:09:26): And I could, (01:09:26): I could literally tell my LPs, (01:09:27): I could step away and this thing would just run and just kick off cash. (01:09:30): Yes. (01:09:31): Yeah. (01:09:31): You know, right. (01:09:32): And you can just put money into this machine that is Team Ignite. (01:09:35): And because of all the AI and software that we’ve built and all this process (01:09:39): automation I’ve done,
Brian Bell
(01:09:39): I literally don’t even have to, (01:09:41): like, (01:09:42): I’m just a figurehead. (01:09:43): Yeah. (01:09:43): And just everything just kind of runs without me. (01:09:45): Yeah. (01:09:46): Keep us on the LP list. (01:09:48): I’m fully sold on that vision. (01:09:49): Nice. (01:09:49): Love it. (01:09:50): Yeah. (01:09:50): Well, yeah, definitely. (01:09:52): Yeah. (01:09:52): That’s another cool thing about being a VC is like now that I back 300 founders, (01:09:56): there’s going to be some percentage of those that’ll be investors in the future. (01:09:59): For sure. (01:10:01): Yeah, absolutely. (01:10:01): There will be. (01:10:02): Yeah, for sure. (01:10:03): So last question,
Chris Hicken
(01:10:05): what will UX design and research look like five years from now and maybe even 10 (01:10:09): years from now, (01:10:10): if you kind of realize your vision for they said?
Brian Bell
(01:10:12): Yeah, I mean, you know, there’s rapid fire. (01:10:14): It’s push a button, get an insight. (01:10:16): That’s it. (01:10:17): Push the button, get it. (01:10:17): No one wants to do research. (01:10:19): We’ll build, (01:10:19): software will exist that lets anybody, (01:10:23): like a camera, (01:10:24): open the camera, (01:10:25): get a picture, (01:10:26): click a button, (01:10:26): get an insight. (01:10:27): Yeah, I love that.
Chris Hicken
(01:10:27): I think that’s where we’re headed.
Brian Bell
(01:10:28): Well, Chris, thanks so much. (01:10:30): This was definitely a longer episode than usual, (01:10:33): which is a good indication of how much fun I had. (01:10:35): So thanks for coming on.
Chris Hicken
(01:10:36): Yeah, likewise. (01:10:36): Really appreciate you inviting me. (01:10:38): And thank you for being a great supporter, investor. (01:10:42): And yeah, you’ve been a great partner to the company. (01:10:44): So I really appreciate that. (01:10:45): Thank you. (01:10:45): Appreciate that. (01:10:46): All right.
