Most founders obsess over product, fundraising, and growth. They track revenue, CAC, conversion rates, and maybe retention. But according to Attila Tóth, co-founder and chief strategist of Cognitive Creators, one of the biggest risks in a startup is often hiding somewhere less obvious: inside its digital strategy and data infrastructure.
Attila’s path into marketing data started with a teenage side project. As a young cyclist, he built a basic webshop for his father’s business. When the first sale came in, his first reaction was not celebration. It was curiosity: why only one? That question pushed him into analytics, consumer behavior, conversion, and eventually digital due diligence.
Today, Attila helps companies, investors, and acquirers understand the hidden risks inside digital businesses—from marketing inefficiency to messy data systems to cloud infrastructure mistakes. In one M&A audit, his team uncovered roughly €2.5 million in risk inside an €80 million deal. The lesson: digital risk is not theoretical. It can directly change valuation, negotiation, and outcomes.
The Paid Marketing Treadmill
One of Attila’s sharpest arguments is that many companies are trapped in a paid marketing system they don’t fully understand.
The trap starts simply. A company spends money on Google, Meta, TikTok, or another paid channel. The campaign works. Revenue grows. So the company spends more.
But then competitors enter the same auction. Costs rise. The company has to spend more just to achieve the same results. Over time, the business becomes dependent on platforms where it does not control the rules.
Attila compares this to a bakery bidding on search terms like “fresh sourdough bread.” If one advertiser pays $0.50 per click, another may bid $0.51. Then a larger competitor enters and pushes the price to $1. Smaller players either raise their spend or disappear from that digital market.
That is the treadmill: you keep running, but the economics do not necessarily improve.
For startups, this matters because early CAC can be misleading. A company may look efficient in its first niche or first geography, but that does not mean the same economics will hold when it expands. Attila has seen startups celebrate strong CAC, only to discover that the next market—such as moving from the UK to the US—is dramatically more expensive.
First-Party Data Is the Escape Hatch
Attila does not argue that companies should abandon paid marketing. That is unrealistic. His point is that companies need more control, and the best source of control is often their own first-party data.
Most companies already have useful data sitting inside their business. The problem is that they do not use it well.
His tire example makes the point clearly. If a customer buys summer tires today, that customer probably does not need to see tire ads from the same company for the next several weeks—or maybe even years. Yet many companies keep retargeting people who already purchased, wasting budget and creating a bad customer experience.
The same problem shows up in banking. Attila described receiving loan offers from a bank despite never taking personal loans and consistently using investment products instead. The bank likely had enough data to understand his behavior, but its marketing system was still blasting irrelevant campaigns.
This is not just a marketing mistake. It is an operating problem. Data often sits in silos. It is messy, incomplete, duplicated, or missing key context like dates and behavioral signals. Without clean, centralized, usable data, personalization becomes impossible.
The Cloud Credit Trap
Attila also warns founders about another hidden startup risk: free software and cloud credits.
Startup programs from major cloud providers and software companies can be helpful. Free credits make it easier to launch, test, and scale early. But they can also create bad habits.
Founders may build on infrastructure that is oversized, poorly configured, or unnecessarily expensive because the bill is hidden by credits. When the credits expire, the company suddenly faces costs it never designed around.
This is especially dangerous because early technical decisions often compound. A stack that looks “free” at seed stage can become expensive technical debt by Series A or Series B.
The question founders should ask is not just: Can we get this tool for free?
It is: Will this still make sense when we are paying real money for it?
What VCs Miss in Digital Due Diligence
For investors, Attila argues that digital strategy deserves more scrutiny.
Traditional diligence often looks at market size, revenue growth, customer concentration, product differentiation, and team quality. Those matter. But Attila believes investors often miss the market’s digital footprint.
That means understanding how customers actually search, compare, discuss, and signal demand online. Search behavior, sentiment, category growth, geography-specific interest, and platform dynamics can all reveal whether a startup is riding a real market wave or merely selling into a narrow pocket of temporary demand.
This is especially important for timing. A startup can have a strong product and team, but if the market is not ready, growth will be harder and more expensive. Conversely, a startup entering a market with rising digital demand can ride a tailwind others have not yet noticed.
Brand Is a Resilience Mechanism
Attila also pushes back on the shallow definition of brand.
Brand is not just a logo, color palette, or tagline. Those things matter, but they are not the core. To Attila, brand is about connection. Real connection with an audience creates resilience.
His example: Apple could make unpopular product decisions—like removing ports from MacBooks—and still survive because the brand had deep customer trust. A no-name company making the same mistake might not survive.
For startups, this matters because performance marketing alone is fragile. If customers only know you through paid ads, you are vulnerable to rising CAC, copycat competitors, and platform shifts. But if your audience has a real relationship with the brand, you have more room to recover, adapt, and compound.
In VC terms, this connects to category creation. The best startups do not just sell into a category. They define one.
AI Will Make Marketing Worse Before It Makes It Better
One of Attila’s more provocative points is that AI may initially make marketing worse.
Why? Because many companies are using tools like ChatGPT and Claude lazily. They ask generic prompts, accept generic outputs, and publish campaigns that sound like everyone else’s campaigns.
The result is sameness.
As more companies rely on default AI-generated messaging, differentiation may collapse. Ads, emails, landing pages, and brand copy will start to converge. Customers will see more noise, not more relevance.
Attila’s view is not anti-AI. The better path is to combine AI with proprietary customer behavior data, market signals, and sharper human judgment. AI can accelerate iteration, personalization, and campaign testing—but only if companies feed it something more distinctive than a generic prompt.
The Investor Question Founders Should Be Ready For
Near the end of the conversation, Attila offered a question he thinks more investors should ask founders:
If a similar company appears in six months, how will you react?
It is a deceptively strong question.
It tests more than competitive awareness. It reveals whether the founder understands their moat, distribution edge, data advantage, brand position, and speed of execution. A weak founder answers with vague confidence. A strong founder has a specific response.
For startups, this is the real challenge. It is not enough to grow while the market is quiet. You need to know what happens when competitors notice the same opportunity.
The Bottom Line
Attila’s message is blunt: growth is not just about spending more, moving faster, or trusting platform dashboards.
Startups need to know where their data lives. Investors need to understand whether CAC is sustainable. Founders need to think beyond the first beachhead market. And everyone needs to be more skeptical of digital strategies that look good only because no one has audited the underlying system.
The companies that win will not be the ones that blindly pour money into paid channels. They will be the ones that understand their data, own their audience, define their category, and build growth systems that can survive competition.
👂🎧 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 — Intro to Attila Tóth and Cognitive Creators
00:25 — Attila’s origin story
00:29 — From teenage cyclist to accidental web builder
02:54 — The first online sale
03:10 — Discovering analytics, tracking, and consumer behavior
04:29 — Launching Sight Doctor at 18
05:00 — Early startup failure and hard lessons
05:40 — Digital business modeling for traditional industries
06:45 — The M&A audit that exposed €2.5M in risk
08:57 — Writing Hyper and the frustration behind marketing data
11:14 — The rising cost-per-click problem
12:18 — The bakery ad-spend analogy
14:52 — The paid marketing trap
16:43 — The marketing spend treadmill
18:10 — Searching for an escape from platform dependency
20:00 — Turning years of experiments into a book
22:04 — Self-publishing Hyper
22:34 — Defining the marketing data trap
24:00 — First-party data as the escape plan
24:22 — The tire purchase example
26:29 — Banks, bad segmentation, and irrelevant offers
28:26 — Data silos inside large companies
31:00 — B2B marketing stacks and startup tooling
31:40 — Why there is no perfect tool list
32:35 — The hidden cost of startup cloud credits
34:04 — Questioning the tech stack after credits expire
35:33 — What founders and VCs misread in campaign performance
36:08 — CAC sustainability beyond the first beachhead
38:35 — The UK-to-US expansion problem
40:16 — Digital brand value as startup resilience
42:29 — Brand connection beyond logos and colors
44:01 — Category-defining startups
44:43 — Slack, Teams, and category creation
46:43 — The unsolved interoperability gap
47:50 — Market digital footprint as a VC diligence lens
50:49 — AI, marketing sameness, and lazy prompting
53:41 — The coming wave of AI-generated marketing noise
55:12 — Iteration, personalization, and AI-assisted campaign testing
56:28 — Event-driven marketing and localized campaign signals
Transcript
Brian Bell (00:00:53):
Hey everyone, welcome back to the Ignite Podcast. Today we’re thrilled to have Attila Tóth on the mic. He is the chief strategist and co-founder of cognitive creators and the author of Hyper the untold story of marketing data a book that reframes how marketers, founders, and investors should think about measurement data and digital strategy in an era of rising automation and privacy constraints thanks for coming on Attila sure happy to be here. I’d love to start with your origin story what’s your background
Attila Tóth (00:01:19):
Well, it’s a story that’s not connected to the world of digital. It’s a story that’s related to cycling. I used to be a professional cyclist when I was a teenager and my father, basically my My family were my sole sponsor. My parents were paying for the bikes, for the trips, for the equipment, whatever you need for the races. And my father is an entrepreneur and back then he started a new business. I was in high school. school 11th grade focusing on IT and English and my father thought that my son is learning IT let me give him a summer assignment and the summer assignment was I need a webshop it’s like that we are learning Turbo Pascal and see it’s nothing to do with with web development or anything like that but he was like I really needed for this new business and I said okay I had this internal motivation that they helped me go enjoy my biking and go to different races in country and out of country so that’s the least Amount of thing I can do to at least try it and if I’m not able to do it okay I’m going to say I failed but I want to give it a try so the summer went by and of course with the help of friends Google, YouTube, you name it. I made the webshop happen. Nothing fancy, but it was functional. It had a checkout process and basically gave the keys to my father, the login to the admin panel and forget about the project. And I still remember the day it was raining quite heavily and my father came to my room and said, we made the first sale. First sale? Where? What? It’s like, well, on the website you built, the website I built, that scrappy little thing, I barely made it happen. Yeah, we have the first sale. So then I asked, okay, why just one? how many people visited the website soon I realized there’s no way to track on that website the visitors or what they do on the website so I started learning about data tracking data analytics then asked my parents to buy a couple of books about consumer behavior the psychology of how people think around making decisions making buying decisions so next to cycling quite soon as a hot-headed teenager I had a new hobby that was connected to data and digital behavior. So this is how my story started at 17. At 18, I launched my first business. It was called Site Doctor. The idea was that I can heal websites to convert better. It’s a stupid teenager idea. And yeah, it was an interesting try. A couple of co-founders who were my best high school friends and failed massively after roughly two and a half years. We went basically bankrupt. We needed to stop the business but yeah it was an interesting learning and one of the first failures then which put me on the path of learning more deeply digital business modeling which is a fancy one A way of saying, helping old fashioned businesses create new revenue streams with digital products, digital services across automotive, tourism, healthcare. So really, I would say the boring Big traditional businesses who have nothing to do with digital and then helping them find an edge and build these new services and that’s what I’ve been doing for quite a few years and long story For short, coming to what my focus is today. Roughly a decade ago, we had a client who we built a successful digital business model. They were making a couple of millions of added revenue from this digital services we set up for them. And they called me and said, well, we have now an MP. deal we are buying a scale up and could you check these guys because you built our digital models and this new company is digital first so it would probably make sense for you guys to check it and we did check it We had a service level agreement and said, sure, we’ll use a couple of hours from there. So we check it and we found massive risks worth back then, roughly around two, two and a half million euros at 80 million deal. which basically helped them negotiate better price and it’s really interesting that those risks were not even something that the founders of the scale wanted to hide those were coming from just growing too fast the business and not understanding some basic principles in cloud architecture and technology and creating some data redundancies which were not necessary, which then created massive cloud build service consumption on Microsoft Azure. Again, something we immediately spotted because when you’re a builder, you have to build first with the minimum necessary and then scale it from there. These guys who basically built a really good scale up. It was their second company and didn’t have the experience to spot those issues. But at the end of the day, our client was super happy. We helped them save 2.5 million euros, which even back then was a serious amount of money and this is my focus today that was the light bulb moment that not only building digital business models but actually auditing them on different transactions whether that’s a series a investment or an a or a carve out or a larger merger. It’s something that worked doing and we call this service digital due diligence, which basically looks at the business from a digital perspective. That’s data, technology, marketing, digital brand value, everything related. So that’s where I’m today. That’s my focus. And yeah, in the meantime, I wrote a book about marketing data because it was a topic. It was frustrating many of our clients, why they are spending more and more back then mainly on Google. Of course, ever since the power shifted to social media, the meta ecosystem, TikTok, but even TikTok. And I think that today there’s a big issue on the market like if you go on these platforms, you basically play by their games, and no one is telling you the rules of that game. So if you don’t understand that game, You likely are overspending, whether you’re a corporate or whether you’re just scaling up your startup. So the book is something that started from a frustration. I wanted to understand why this happening. We did a lot of research, a lot of tests and experiments and understood, okay, what’s happening? How can you better work with your own data? As you said, in the era of privacy, GDPR, it’s an interesting topic.
Brian Bell (00:08:50):
Yeah, fascinating. It’s how I kind of got my start in tech as well. You know, about 15 years ago, I started in marketing demand gen.
Attila Tóth (00:08:57):
Okay.
Brian Bell (00:08:58):
Yeah, I ran marketing for a few VC backed startups. So I was probably part of the problem that you’re describing, you know, just throwing lots of cash at trying to grow startups. And I eventually moved to product for a variety of reasons. One was I realized that after running so many multivariate tests, I had to actually go in and change how users got to the aha moment inside of a product, which was kind of what you might call like growth product management today. where that term did not exist. And we might have called it growth hacking back then, like 2013. Let’s just dive into the book. It’s hyper and the current state of marketing data. What year was that book written? Who was the intended reader?
Attila Tóth (00:09:37):
So let me share a bit of of the backstory of the book was 2018. So this is pre COVID times. And we had a client who was doing really well, the digital business model we set up was generating revenue nicely. The only issue we had, and that was a major issue was cost per click marketing spend. So in order to create more revenue each year we have to ask for more money and a conversion rate was not improving meaning if we spend last year 500,000 next year we might need to spend and 550 just to reach the same level of revenue and the reason behind that is because even in 2018 more and more businesses started spending money on the platforms and if we just stay with to simplify the example one platform let’s pick Google search and let’s say you and me have a bakery right in the same town and we advertise freshly baked I don’t know sourdough bread New York okay if you’re the first advertiser then you basically set the benchmark of how much you will spend for each impression and each click I’m the second advertiser and I will try to outbid you so let’s say you spend 50 cents on a click I will outbid you with 51 cents or 52 cents We are two small players, but let’s say a big corporate bakery comes into New York. This is a cheap market. There are only two guys. Let’s just increase ad spend from 50 cents to one dollar because we have the budgets, right? So we, the small guys, we have basically two options. Either we increase our ad spend to pay from 50 cents or 51 cents to a dollar or maybe a dollar and 10 cents. or we are left out from that digital market. And of course, small players sometimes are left out from the market, but the market is huge and there are a lot of corporates who are just pushing up and up the prices. And this is working kind of like the stock market. So the more players you have for a certain audience and a certain set of keywords, if we stick with Google, then those costs are just skyrocketing.
Brian Bell (00:12:09):
it’s actually a business theory actually you know they teach in school that you want to increase your you want to set your marginal revenue to marginal cost right yeah the marketers will come in and they’ll basically just turn up that marketing spend in every channel until every dollar that they get in contribution margin equals the ad spend exactly and so like marketing ruins everything eventually this is the trap right because you come in and you’re just like like everybody else is spending a dollar so I got to spend a dollar
Attila Tóth (00:12:35):
Exactly, exactly. And I took Google as an example, but it’s the same whether we go for Instagram, TikTok, it’s the same engine. The only difference is that new channels are more for giving. So TikTok used to be cheaper until they have reached a certain level of audience and a certain level of paid clients. But if you just think a big back Let’s say 2007, 2008, Facebook used to be like super simple, almost no ads. I don’t exactly remember which year was exactly when they started introducing ads, but even those channels were mainly organic. Now, if you post something on Instagram or Facebook, it’s basically lost. Nobody is going to see it until you spend some money on it. So you pay for it.
Brian Bell (00:13:29):
And then an insidious thing can happen on those platforms now too where once you turn on the paid spend, you actually hurt your organic traffic as well. Something in the algorithm is like, oh, well, This must not be as popular anymore because it’s just not getting as many of impressions, clicks, views, comments, shares. And so the algorithm will now kind of de-weight you in its prediction algorithm because you were paying for that engagement before. And as soon as you turn off paying for engagement, all the engagement that you had before you started paying for engagement, it goes down.
Attila Tóth (00:14:03):
Absolutely. Absolutely. So that was the pain.
Brian Bell (00:14:07):
That’s the treadmill of marketing spend that you get on.
Attila Tóth (00:14:11):
exactly and basically I had a tough question coming from the CMOs one of our clients said why we’re spending more and not increasing accordingly our our income and I had to explain this whole theory how it works but I felt like Google doesn’t pay me so why I’m why I’m explaining this with like kind of like positivity and enthusiasm and telling the client yeah this is how the market works and this is normal and blah blah it was on one end I knew why we are doing that but I didn’t have a better answer to say how to avoid this situation and that was probably a tough conversation but it started I have a question in my head, like, is this really the game we need to play? Is this the only game we can play? And I wanted to find an answer to that question. And I was not sure what that answer will be. was the underlying question that started the book. So the idea started around late 2018. So we started doing a lot of tests, trying different things, as you call it, growth hacking, different ways to still use these platforms, but do also something differently. Most of these failed because we ended up in the same, playing the same game. So whenever we wanted to scale, We were back to square one and needed to increase ad spend. So we started thinking, okay, how can we get outside the system? What’s holding companies back to get outside of the system? And by, I think it was mid-2000... 21 where I had enough evidence how to basically see the trap recognize it see your capabilities and and basically understand what can you do as a company. You cannot be independent from these platforms, probably a myth, and anybody who promises that probably is a snake or a salesman, you cannot be completely independent however there are ways and mechanics that can help you to have a bit of control on what you are doing and that was the moment when I said okay I have enough proof based on our tests and experiences with clients, what are the things that are working and by then I saw this pattern that can be applied from an SME to a big corporate and it’s functional. I was just finishing I think it was late December in 2021 when I decided okay from next year I have to collect everything what we learned since 2018 and try to put it in a format that can be easily digested and can be understood also by people who are not CMOs, simple business decision makers. So the process started and from 2022 until the beginning of 2020, basically I wrote the main copy of the book and then in early 2024 I sent the draft to a couple of people who I think can challenge my My thinking in the right way. These were people who were marketing professors at different universities who were CMOs at different companies or who were data people and wanted to get like a brutal feedback. What am I missing or what am I not covering? Is it consistent enough? Does it make sense? So I collected quite a lot of feedback, incorporated that feedback, and I think it was summer of 2024 when I launched the book on Amazon.
Brian Bell (00:18:19):
Nice. Did you go the self-publishing route or did you hybrid?
Attila Tóth (00:18:22):
Yeah, self-publishing. I didn’t write this book because I want to be an author. It’s not even in my LinkedIn. I know many people put in their bio their authorize. I don’t consider myself an author, I consider myself an experimenter who wanted to share the pain and learnings with people who want to understand this challenge and want to find a way how to get out of this trap.
Brian Bell (00:18:49):
So let’s talk about some of the learnings in the book. So we talked about the trap. Let’s talk about the escape plan in the book. Define what the trap is a little bit and then we’ll go to just how do you get out of it?
Attila Tóth (00:18:59):
Yeah, so probably in a nutshell, the trap is thinking that by going If you’re going on these platforms and spending on digital marketing, you’re doing the right thing. And you need to always spend more and more and just adapt to market conditions. If competitors are spending more, you try to increase your budgets. and if for some reason they are not covering a niche then you’re happy that you have a niche where your cost per clicks are super low and that’s it and that’s how most marketers behave because these platforms I want you to think like that. Now the escape plan. First and foremost, there’s no magic bullet. However, there’s one common thing that we identified in our tests and experiments. And that common thing is that each company sits on some level of marketing data. And I’m going to tell you like a super simple example. Let’s say you’re a company who sells tires, summer tires, winter tires, of course, right? Most of the companies who sell tires across Europe, US don’t utilize their own customer data. Meaning if I sold a tire to Brian yesterday, summer season is starting and you It’s highly likely that in the next four weeks you won’t buy another set of tires it’s highly likely I’m not saying 100% because you might have a second car but highly likely you won’t buy the same Size and type of tire now most companies don’t utilize that data that Brian just made a purchase yesterday so what will happen that Brian for the next two to four weeks you will see tire ads from the same company You just purchase your tires for the next couple of weeks.
Brian Bell (00:20:56):
It’s like, I just bought your thing and you’re showing me this.
Attila Tóth (00:21:00):
Yeah. And this is probably the easiest low hanging food to take your customer data and make sure you exclude from your campaign those people who recently made the purchase. Of course, it depends on the product, the seasonality. But in case of tires, you can take out a person from the campaign probably for quite a long time because maybe in the US, I don’t know, the average tire...
Brian Bell (00:21:26):
You know if they bought summer tires in April and therefore you don’t need to show them summer tires for at least a year or two, right?
Attila Tóth (00:21:34):
Yeah, at least two or three years. So this is like a really simple example. But the principle is that your company is sitting on first party data that’s yours. You didn’t buy that data set from outside. it’s based on how your customers are behaving right now and you should use that data actively and of course depending of the type of business what actively means can be different let me give you another example banks. Even in the modern era of fintech where we have Revolut, we have Wise and all these fintech companies who are revolutionizing how banks work. Many traditional banks are still stuck on not using I will give you a simple example one of the banks I personally use they keep sending me offers with loans and what they didn’t realize I never took a single loan on my property I always invest my money to different saving accounts or the stock market, and they don’t offer me products related to investment. They try to send me a loan, which is in my case, 99%, I’m not going to take a loan from them.
Brian Bell (00:22:54):
Well, maybe to steel man the other side, maybe they’re looking at it and they’re saying, hey, this guy has the ability to take a loan out. He’s a low risk. And he doesn’t have a loan. He should have a loan. Right?
Attila Tóth (00:23:05):
Yeah. But usually what happens, they just have an email database. Right. And the CMO cell, well, this quarter we need to increase our loan rate.
Brian Bell (00:23:16):
We need so many numbers that he locks by the end of the quarter. Go. Yeah. Yeah.
Attila Tóth (00:23:20):
and they just push the same message for everybody and it’s not working it’s not converting and sometimes it’s even annoying like you get a message that’s completely irrelevant to what you have or what you need so the whole book is based on the principle of how you can utilize your own data and activate it in your digital marketing and growth campaigns. And this sounds simple but it’s not so simple to do. And the reason why because many companies especially the large ones sit on humongous amount of data which is uncentralized living in silos nobody knows where data points are nobody cleanses that data. So it’s a mess. If you’re a startup and you understand the principle, you can already build your company in a way like, I want to make sure that I’m utilizing my first party data strategically. But in most cases, sites or larger companies haven’t built their businesses around that principle. So when you say, okay, let’s activate some of your data points you’re having, like We don’t know where that data sits. Oh yeah, it’s in Salesforce somewhere, but we don’t actually understand where the Salesforce database is. It’s somewhere in the cloud. Okay, we do a CSV export, but that’s a one time and then it’s not working. And of course, today with AI agents, You can automate many things, but you understand the issue. Data is in silos. It’s messy. Most of the data is not useful and it’s not useful because it was collected in a way that some really critical parameters are are missing, like the date, let’s say, when there was an event, let’s say it’s an e-commerce platform, and somebody put in their wish list an item, right? And if you don’t have the date when that wish list item was saved, that data is almost useless. Because if that event happened two years ago has a different meaning compared to if it happened two weeks ago. Again, depending on the product data, whether it’s a B2C Yeah,
Brian Bell (00:25:36):
and this is why Marketo was so powerful when it came out, you know, 15 years ago to almost 20 years ago, is you could go in and start creating smart lists and static lists on the customer data. So I remember doing a lot of those CSV exports and merging and list creation and segmentation and what are you seeing in best practices? This is very much a startup podcast, right? Venture Capital and Startup Podcast. What are you seeing out there on the B2B front that’s working like in today’s marketing stack?
Attila Tóth (00:26:08):
I think it’s a stupid answer but it really depends on what type of B2B startup we are talking about because there are so many tools out there and I don’t think there’s a list of perfect tools how I approach things as we do a lot of digital due diligence whether those tools make sense for that specific company in the growth phase they are so if it’s series A it’s slightly different are different from a series B or from a fully grown corporate merger. So they need to have a different tool set. What I see, and this is, I think, thanks to, on some level to COVID, during COVID, Most of the big guys, Microsoft, Adobe, gave a lot of credits to different accelerator programs, venture capital companies, credits that startups can use to just buy a license, We set up basically free of charge a Microsoft Azure infrastructure, which sounds super sexy, like you get $60,000 of credits for a Microsoft environment, right? Sounds really good. But what happens when those credits and usually these guys don’t really have an incentive to help you set up your ecosystem in a smart way they have the incentive to give you those credits because they think okay if I don’t know from a thousand startups only a hundred will grow and become a serious company and maybe one or two becomes a unicorn then that the math is is worth it for them But the rest of the companies who are maybe growing slowly or don’t even have the vision to become a huge corporate will pay more than they should for services they don’t need. And that’s a trap, a different kind of trap. I see that many startups use these credits because they are free of charge, but once they are out of the credits and start billing and they have some revenue, they never question whether the tech stack or the tools they are the right fit for the stage the company is going to. So it’s hard. There’s no magic bullet here as well to say these are the five marketing tools I would use. What I would recommend in instead of thinking how can you capture your own data in a centralized way and if it’s a custom-made agentic AI that grabs from your website forms the data points and puts it into an air table it’s fine it doesn’t have to be necessarily fancy it has to work until you of course reach a limit of growth where then you say okay this was the bootstrapped version now we do the next layer but even the next layer shouldn’t be sales for 360 automation which again could cost a lot of money what’s a common
Brian Bell (00:29:17):
mistake you see founders making when or VCs when they’re doing due diligence on interpreting campaign performance
Attila Tóth (00:29:23):
Looking at the first mistake that’s really common, let’s say this is something actually we just finished an audit on a series A startup and they had a really good customer acquisition cost. It made a lot of sense financially speaking and they were super happy.
Brian Bell (00:29:43):
So your CAC, your customer acquisition cost was way lower than your lifetime value. You’re paying back very quickly. The VCs are happy.
Attila Tóth (00:29:51):
exactly and nobody questioned whether that cost is sustainable and by sustainable I mean for those audiences they are attracting in this quarter are those audiences big enough and will will it cause the It’s the same when you scale for a different niche or a different beachhead market because of course in the startup world first you want to conquer your first beachhead and then you go to the next one. The reason they’re not asking this question is because they are thinking only short term and we have to survive the next 12 months or the next 18 months and then we prepare for the second round of investment. And that mindset, of course, works well when the economy is booming and everybody’s throwing money at VCs and startups and VCs have larger funds. But in the current circumstances, I think there’s a lot less capital deployment on the market, especially for Series A and below Series A, then you shouldn’t think of your first round or your second round as just a breath of fresh air to survive. You should think a bit ahead at least two years three years and to understand the market dynamics because I saw at least so in the last six months I saw five companies who were super happy with their CAC were not looking looking wide enough on the market and one of these companies already basically reached the audience that’s capable of reaching and now they’re asking okay for the next frontier we have to go here and here that’s a different country they started in the UK and now they are attacking the US and the US is way higher in terms of customer acquisition cost and the ROIs are really bad yeah so that’s that’s a common common thing that
Brian Bell (00:31:52):
I don’t know if you agree with this statement but I’ve noticed that CAC doesn’t tend to go down over time. Have you seen it actually go down over time? It tends to go up, right? So that’s an interesting truism I think in demand gen circles. And something I think as you analyze companies that you should be aware of, if the CAC is such that it takes a year to pay back right now and they’re at like seed or series A, it’s not going to get better than that. You know, maybe they’ll be able to cross sell and upsell and get more revenue out, but the CAC is not going down, right? and so you got to really consider that are they do they have this is something we care about a lot now and I think it’s become the bottleneck in venture capital or startups I should say is it’s distribution right because it used to be like oh like I could I could build something that’s amazing and now it’s not so much about that anymore like anybody can build stuff and engineers are 10x more efficient or whatever it is with cloud code and codex and and others so you can build whatever customers need the real real challenge is getting them to care and care fast enough and pay for it
Attila Tóth (00:32:52):
Exactly. And here’s another thing that I saw it working that many also VCs and startups completely miss is the digital brand value. So branding, I think, is one of the most misinterpreted word out there in marketing if you ask a designer what’s branding they will say it’s the logo it’s the colors if you ask a CEO it says it’s the company values so everybody has a definition of branding and on some level they are all true but I think the most important element of branding is connection real connection not fake connection real connection to that audience and I like to use this example with with Apple so a couple of years ago I think Apple messed up with one of the MacBook Pro versions they removed all the different inputs from HDMI SD cards and they were just going to full flat USB-C however since there was a connection to the brand with their audience they survived this strong decision and then I think three four years ago they brought back these You can see these different input options for MacBook Pros. But if a mistake of that level is done by a no name company in the market, that company can go back up because there’s no connection to that brand. What startups and we see sometimes miss is investing in the brand and understanding how you create that connection which is not this color or logo which of course are important but it’s a more meaningful connection And this is something that many startups don’t focus on at all, like, we want to get product market fit. Okay, but what happens when you start with a product Saturated product market fit. What’s going to happen there? Branding, especially digital branding, it’s a tough thing to crack. But if you don’t start early enough, then you will be just another company in the market. best case yes you do an exit to a PE or a big corporate who can put your company strategically in their ecosystem but you have no brand value that’s something that’s difficult to explain to investors and most founders don’t know how to explain that and that’s the only thing I saw from a data perspective where brand value was measurable and visible CAC I’m not saying it it it become like 50% cheaper to get the customers but it survived rough seasons
Brian Bell (00:35:41):
without going into crazy heights yeah we call that in our our framework of evaluating startups we we call that category defining meaning like can this brand can the startup define the category or own the category in which it’s playing where people synonymize them with that That product or service right and then you get the brand halo effect and that’s yeah it’s hard to predict that as a VC when you’re looking at a startup but it’s it’s something that we do ask ourselves can this define the category will this become the apple of of this this you know b2b segment absolutely there’s a simple
Attila Tóth (00:36:18):
question and this is something I I learned from one of my mentors when you want to Understand a category or a brand. Even if you know your audience and you ask a couple of questions, sometimes they don’t know the answer because they think limited in the box of a different category And that’s also why defining a brand in a completely, or even if it’s not completely new, because let’s just pick Slack messaging apps where they are before. for, right? But Slack found a specific category in which they dominated the market. And if you would ask before Slack people in the corporate world, in startup world, what are your pains in communication? They would have probably different answers, but none of those answers would point to a solution what Slack has done. So it’s a tough thing to... Yeah, I don’t think anybody would have said,
Brian Bell (00:37:21):
you know what, I need an instant messaging app. you know with channels and and I need to be able to hashtag and at sign people inside of channels I don’t think people would have would have said that and then you had teams that came along after that and decided oh you know what we’ll make the calendar the channel and the chat you know integrated with like OneDrive and and so like every recurring meeting can have its own like drive or every channel can have its own drive. And I think that was like an innovation that teams had.
Attila Tóth (00:37:49):
One thing nobody actually solved, and I think it’s a good niche since we do a lot of M&A work. and have a lot of different companies to audit. Let’s say the mother company is on Microsoft, the corporate and the scale-up they’re investing in or the startup they’re buying, it’s on AWS or Google Cloud and those systems just don’t talk to each other. Nobody is solved yet in a good way to synchronize your Microsoft Outlook calendar with your Google Amits calendar. And it’s an interesting niche and I don’t know who will solve that, but definitely if you would ask a couple of people today, they couldn’t define the category, but there’s a problem that’s existing.
Brian Bell (00:38:36):
What other things should VC, early stage VCs in particular, like us look at when we’re evaluating a startup’s distribution and paid marketing and acquisition channels?
Attila Tóth (00:38:47):
One of the things that are most of the time hidden on plain sight is looking at, we call it the market’s digital footprint. Let’s say you have a B2B startup that is in the automotive sector. Post VCs will look at the economics from potential market size, country regulation, so on and so forth. highly likely miss is looking at market dynamics I’m going to give you an example autonomous driving it’s a really cool thing and it’s a category that’s that’s growing however there’s a difference in how Audiences are on one end expecting autonomous driving and the other end trusting autonomous driving. So if you go to a German market and look at the data how people search for cabs, Uber, Bolt and different solutions is different compared to how people search for those services in Japan. And those slight nuances open up niches and sometimes those niches are Really easy beachhead markets where a startup can enter and grow really fast. And I think this lens of looking at the market’s digital footprint is missing from most of the VC checklist. not because they don’t have people who understand data it’s considered secondary because they never saw relevant data in action and I think that’s something where VCs need to Be open-minded and look at some investments happening on a global scale where these
Brian Bell (00:40:44):
data points are accessed Next two questions are kind of related, but you’ve been in marketing for a while, you’ve seen a lot of changes What’s changed the most in the last five or ten years in your mind? And what are you looking forward to in the next five or ten years?
Attila Tóth (00:40:58):
Let me share a funny example. This is an example from US. So there are three different companies in the same segment, competitors. I think until 2022, all of these companies had different brand positioning. So if you looked at their marketing, one is really red, the other one is blue and the other one is yellow. But not just in colors, in messaging, positioning, everything. Now, if you look at those three companies, the branding is different but the visuals messaging is almost identical and of course that’s because Claude ChatGPT and basically this safety and I like to I may be a bit aggressive with this word but I think it’s laziness of using AI tools just on their default mode like okay I’m running a campaign for this audience in this country blah blah blah for this product Yeah. And then of course, how LLMs work, they connect different data points and generate an answer. And if you don’t have anything specific, like you don’t really put in effort and using the LLM to really...
Brian Bell (00:42:15):
You just like take whatever, whatever it’s outputting on 5.4 or 5.5 or whatever it is. Yeah.
Attila Tóth (00:42:20):
Yeah.
Brian Bell (00:42:22):
That sounds good.
Attila Tóth (00:42:24):
And that’s... that’s something I I’m seeing that’s happening and I think in the next five years it will get worse meaning even for a SME like even if they didn’t have a CMO or startup didn’t have a CMO they had a friend or they asked an opinion or like we are building this product could you let’s sit down for coffee I invite you for dinner like you know growth hacking get getting ideas for their marketing now what happens
Brian Bell (00:42:56):
Or even going to a conference.
Attila Tóth (00:42:58):
Or even go to a conference. Yeah. Now what happens, they sit in front of their laptop. Best case, they have a paid version on one of these LLMs. And that’s the best case. and there will be many similar campaigns many similar messages coming out which will automatically destroy customer acquisition costs these will go up conversions will go down so I think in terms of marketing things will get worse first and then of course because there are always innovators people realize okay we need to do deeper work and come back stronger we can use the tools but maybe use again coming back to your own first part data use the customer behavior data you have and feed it in and create your custom GPT not just you could iterate
Brian Bell (00:43:51):
so much faster now like we used to have to manually create ads and manually set up campaigns and manually just adjust everything kind of by hand really and you know you had a tool you had a SaaS tool and then you’d you know run your Marketo campaign with different subject lines or whatever different body messages or different headlines or calls to action and you’d run a multivariate test and it was a lot of work to set it all up and run it and get the results and now I would imagine with the app you can like iterate your way to better messaging perhaps have you seen any of that in
Attila Tóth (00:44:21):
the field yes I know in the sense of laziness we see that most people even if they can iterate they won’t iterate because initially it saves them time just to put out something and the market is not yet crowded with AI slop meaning it’s it’s not yet we are not yet in the worst phase of marketing so it’s still working and in some niches even these basic prompts are working but those will stop working and then yeah so in that sense but yeah answering also to to your question in iteration I saw a couple of Good things as well, specifically in sports fashion. So when you advertise, let’s say, footwear, running footwear, now everybody’s talking about Adidas and the new word record. And of course this is a good advertising moment for them which will certainly pass but what I saw some brands do really well is to use data points from events happening let’s say in in Switzerland near Geneva there’s a trail marathon in the Alps right and they use that event to promote a specific running shoe in that area because they know this event is happening I don’t know mid-June it’s attracting these type of people so they’re using that event as a wave they are they are basically surfing on to create content around that and personalize their communication with different iterations before, during and after that event. Not directly, because of course, this would mean you have to go and sign partnerships with all the events, but indirectly, but just by having a photo from the mountain that’s in the event, or there are like really smart ways, creative ways, like You get a data point and you can creatively think about what does that mean for images? What does that mean for messages? What type of people will be running here? And those kinds of personalized elements can be iterated really nicely in campaigns and drive up conversions temporarily. But that’s the beauty, as you said. Now you can run the same mechanism for a next event that’s happening in Italy or in Spain or in the US and you don’t need you just set up the framework use different data points and generate completely different outputs while giving a personalized experience to the audience it’s not having the same standard message in all the countries but using these little insights to customize and iterate. That’s something I saw. It’s working really nicely with one of the footwear companies that I’ve been advising.
Brian Bell (00:47:16):
Yeah, I’m sure marketers will figure out how to make it like minority report where, you know, it knows exactly my blood type and everything about like how I slept last night. And, you know, you know, because I wear, you know, I wear the Fitbit and, you know, I didn’t sleep well last night. So I’m sure they’ll be like, you know, didn’t sleep well last night. You need some Monster Energy drinker.
Attila Tóth (00:47:34):
That’s another topic in the book, in the last chapter, which I think we should be talking about more as people, what happens with our data. So in an ideal world, I think we should own our own data, as you said, your Fitbit data, your location data, everything. And in an ideal world, you should have a vault, a data vault, where you say, okay, I’m going to share my data with I don’t know this private clinic McDonald’s or the companies you are highly likely interested in what you do and you either share it freely because those services are relevant to you or if you want you can sell that data and you then you are aware I’m selling my data to XYZ company
Brian Bell (00:48:24):
I think I started trying to do this. It always feels too early. It’s like, okay, and here’s what we’re going to do. We’re going to go around all these products and services and aggregate data. And I’m like, okay, great. And then what? Well, we’re going to help people use it places. Okay and then what and it’s just there’s no business there I think it’s kind of like I think it’s there for defined categories like sharing my calendar right I need to share my calendar with certain services for certain reasons but like if you start like taking all my data and putting it into something There’s just so many variations of how that data is stored. It’s a tough one.
Attila Tóth (00:49:01):
It’s a tough one. Even in the book, I... Have you ever downloaded your data from a service?
Brian Bell (00:49:07):
Like I downloaded my Fitbit data. I was interested. So I was like, okay, I’m going to download it and I’m going to upload it in the chat. I have this project for all my personal health and wellness stuff. And so I downloaded it. It’s going to be great. I’m going to download it. It’s going to analyze it. And the way that it was constructed, I expected like a big CSV, like a log, like a table, right? That’s what you expect. It was literally just like plain text files and folders. And so I’d have to run some sort of ETL transformation, which I probably could code up with Codex or Cloud Code or something. But it’s just like, that’s what I’m talking about. It’s just like, there’s so much data out there stored in all kinds of, and then I’m wondering, did they just make it hard on purpose? I think so. To obey the law. Yes, you have data portability, but we’re gonna make you download it in such a nefarious way that it’s almost unusable like my sleep data like you should have it it’s like on this date this many hours this much deep sleep blah blah blah no everything was like a little plain text file like every cell on a CSV was a plain text file in a file folder system. I’m like, well, who does that? Why’d they do that? Anyway, I’d like to wrap up just a few wrap up questions for we run out of time. What’s a question you wish investors ask more often of founders when evaluating their digital marketing strategy? In your own practice, what you’ve noticed, what has been a signal or metric that has become significantly more predictive of startup success?
Attila Tóth (00:50:31):
That’s an easier one. If we are talking about business categories that exist, it’s basically digital trends. So what’s the sentiment around it? How many people are looking into that topic? Is it a topic that’s growing in terms of their trend or decreasing? from simple search terms to complex things like now you can analyze LLM sentiment so let’s say what’s the sentiment around a special kidney disease or you have so many interesting insights that you can capture and utilize as an investor but also as a startup that could help you define and faster product market fit can help you define your brand these are things that are out there it’s almost free of charge and people are just like they’re not using it
Brian Bell (00:51:26):
I would describe this as in our framework is timing. Why is it the right time for this business? And part of that timing is the quantification of what you’re describing of market readiness. Like, hey, there’s a trend happening right now where it’s the right team with the right idea at the right time and the right market conditions. And there’s like a tailwind here of like just market interest in this category.
Attila Tóth (00:51:49):
Yes, and I think why timing is probably the best word to describe it is because even in like really traditional assets like hotels and real estate, I’m just going to give you one example. This is already out of NDA so I can share it we had a client interested to invest in some luxury hotels in in East Asia and wanted to pick a location and most of their advisors were coming with different offers like here the land is cheap or or look at this island with the views and so on. So they were like really emotional and not objective parameters. And what we did is we looked at how many people are searching for different destinations. in East Asia and back then this is like more than five years ago we saw a massive interest spike in the area called Lombok and Lombok was basically underdeveloped there was nothing there I wouldn’t say it’s Spectacular but since there were a couple of people influencers who went there either for diving or hiking or whatever reason they started creating a micro trend which turned visible in google searches and with with all the investors well people are wanting to go here there’s no infrastructure there are no hotels yet so you can just ride that trend Yeah, exactly. I think timing is the best way because now if you look at Lombok, it’s already full. It’s not the right time to invest. But yeah, timing is the perfect word.
Brian Bell (00:53:26):
All right. So last question. Back to the first question. What is the question you wish investors asked more often of founders when evaluating their startups and digital strategies?
Attila Tóth (00:53:36):
It’s hard to pick one question, but if
Brian Bell (00:53:39):
Or two or three, yeah.
Attila Tóth (00:53:41):
If I had to pick two or three, one would be how is your digital strategy compared to similar, we are now, many people call it SaaS Apocalypse and I think it’s not but we’ll see but if I would ask a startup how would you compare yourself to a similar software service startup yes you are now an AI company but what’s different in your digital strategy what sets you apart and usually the answer will be they don’t know if they are on They say something, but in reality they won’t know.
Brian Bell (00:54:31):
I don’t ask it exactly like that, but I say something around what’s your distribution wedge? How can you acquire the next 100 customers better than your competitors? Something like that, some variation of that, depending on how much context I have and how much I know. yeah and I just kind of listen to see what they say you know because I’m always interested because I’m a demand gen person right I want to learn right oh you’re like oh I haven’t heard of that tool what is that tell me more you know sometimes they’ll have something really creative they’ll say something really creative I’m like okay this this person’s like a growth hacker and I think that really it’s a really important skill now it’s really important to like absolutely hacks of growth
Attila Tóth (00:55:08):
I think it’s more important than ever because today most marketing again will be based on cloud or chat GPT output and if you don’t have anything else that’s outside of that and you have basically nothing that sets you apart I have one more in my pocket. Yeah, go for it. The second one would be if a similar company appears in six months, how will you react? And that again tells you on one end the mindset, but whether they have enough preparation and enough data in the background to know, okay, even if we are new and fast now, but competition will come and then what?
Brian Bell (00:55:49):
Yeah, I like that question. I’ve never asked it quite like that. I usually ask it like, who are you worried about? Oh, we’re not worried about anybody. But like, okay, let’s imagine that’s the good follow up. Okay. Imagine somebody did come along and they’re doing exactly the same thing.
Attila Tóth (00:56:01):
Yeah. Yeah.
Brian Bell (00:56:02):
How do you react? I love that. I really enjoyed the conversation. As you can tell, I could probably talk about this for another hour or two, but we’re out of time. Where can folks find you online?
Attila Tóth (00:56:10):
LinkedIn. That’s the, I think only social media platform I actively use LinkedIn slash in slash creative Attila. Awesome.
Brian Bell (00:56:19):
Thanks Attila for coming on. I really enjoyed it.
Attila Tóth (00:56:21):
Thank you for having me.
