Episode #1: Consumer Tech Napkin | Fundraising & Benchmarks in Consumer Tech

12 May 2026 · 45 min · 29 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

How European consumer tech gets funded now, using concrete engagement/benchmark metrics; AI as an “enabler” not a platform shift; when paid marketing helps vs harms; and how AI changes costs via inference and marginal unit economics.

Guests

  • Samir Singh, Partner at Speedinvest; focuses on marketplaces/network effects; previously at Abani.
  • Susan (Felix Capital), Partner investing across consumer and B2B; background as founder in social commerce (SoFillCommerce app, 2011).
  • Joe Siga-Dupoy, Venture at True; early-stage consumer and consumer enablement tech.

Key claims

  • Engagement per user rising over time is the top investor signal; retention is secondary; monetization last.
  • AI doesn’t replace product/discovery/UI; it’s an API/microprocessor-like input.
  • Paid acquisition is a “tax” unless paired with a defensible growth flywheel/moat; otherwise it can create artificial growth.
  • AI can reduce build and cost-to-serve, but inference costs can cap margins and runway.

Notable examples

  • Snapchat “snaps shared/open per user per day” as engagement proxy.
  • Aura as a case of routine/lock-in enabling performance marketing.
  • Revolut, Spotify, Klarna, Revolut/Apple/Meta/Google as consumer scale outcomes.
  • Function Health (founder brand/distribution boost).
  • Netflix as economies-of-scale example for variable cost reduction.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

The Importance of Engagement in Consumer Tech

0:00 to 1:06

Learn why engagement is crucial for successful consumer tech products.

“If you're trying to draw a common theme across all consumer software, its engagement is more important than retention, is more important than acquisition, is more important than monetization.”

AI as an Enabler, Not a Platform Shift

1:55 to 3:26

Understand the misconception that AI represents a platform shift in tech.

“We've got Samir Singh, partner at Speedinvest, focused on marketplaces and network effects, previously at Abani.”

Clarifying AI's Role in Product Development

3:29 to 5:06

Learn about the importance of integrating AI effectively into products.

“How do people get it wrong and think that AI is a product?”

Misconceptions About AI and Startups

5:09 to 6:12

Explore common mistakes startups make when incorporating AI into their products.

“a lot of companies have woven it into their products to deliver a much better product, more effective, more efficient in different ways.”

Real Opportunities Created by AI

6:14 to 7:58

Discover how AI can reduce costs and open new markets.

“So the underlying models can do that fairly effectively.”

Identifying Signals of Success in Consumer Tech

8:00 to 11:41

Learn how to differentiate successful consumer tech companies from the rest.

“But it's the sort of thing where when we think about when you talk about the why now and is AI a why now and in what circumstances, that's an example where we get really interested.”

Evaluating Pre-Seed Consumer Startups

11:41 to 14:00

Understand the key considerations when evaluating early-stage consumer startups.

“If you're trying to draw a common theme across all consumer software, its engagement is more important than retention, is more important than acquisition, is more important than monetization.”

Understanding Founder Expectations in Consumer Tech

14:00 to 16:12

Learn about the importance of founder resilience and early validation in consumer tech.

“Or would you say, no, no, sometimes you just go with the idea and the team and you're comfortable with that?”

The Role of Metrics in Building Consumer Businesses

16:12 to 19:18

Discover how important metrics and data-driven approaches are for early consumer founders.

“Because otherwise you don't like you're not cut out for it.”

Challenges of Raising Funds for Consumer Companies

19:18 to 22:20

Explore the difficulties consumer companies face in securing venture capital funding.

“I was about to say, at least it's been like that for many years, especially in Europe.”
Show all 29 chapters

The Impact of AI on Consumer Business Models

22:20 to 24:18

Understand how AI is changing the economics of consumer businesses and their profitability.

“or early teens, I think consumer became more synonymous with e-commerce and direct-to-consumer, which there's certainly been some great outcomes from that on being one example, as you mentioned, from Inure.”

Evaluating Cost Structures and Sustainability in Startups

24:18 to 28:00

Examine the importance of cost structures and pricing power in the sustainability of startups.

“There are a couple of small use cases, but then they don't have the scalability potential.”

Understanding Marginal Costs and Pricing Power

28:00 to 28:32

Discussion on the impact of marginal costs and the significance of pricing power for startups.

Moats in Software and Hardware Startups

28:32 to 29:26

Exploration of the types of competitive advantages or 'moats' in software and hardware.

“marginal and increasing costs um and samir how do you pressure test that uh the the pricing power the very basic test is that the number of moats that a software startup can have is like there about three or four.”

Challenges for Pre-seed Stage Companies

29:26 to 30:21

Insights into the difficulties faced by pre-seed companies in building brand and distribution.

“I do think there is brand and distribution mode, but I think, but I agree in the sense of, I mean, I think a lot of that is captured by the existing large players.”

Cost Considerations from Series A Onwards

30:21 to 31:32

Discussion on cost metrics and considerations as startups progress from Series A funding.

“When we're thinking about cost specifically, it's basically the same consideration?”

Metrics Shift in the Age of AI

31:32 to 32:25

Analysis of how AI is altering the important metrics for consumer tech companies.

“And now a lot of that's being increasingly automated by AI.”

Engagement Metrics and Growth Patterns

32:25 to 33:35

Examination of engagement metrics and their evolution in the current tech landscape.

“Just as we just now spoke about cost, we used to be thinking differently about cost.”

Unit Economics and the Importance of Marginal Costs

33:35 to 34:38

Discussions on unit economics and how marginal costs are affecting startup growth.

“But outside of that, that's more of what we're seeing in the metrics as opposed to which metrics we look at.”

Learning from Blitzscaling and Moat Strategies

34:38 to 35:39

Insights on blitzscaling strategies and their relevance to current tech startups.

“So for that reason, like it will solve itself because we are sure that inference cost is going to come down in the end.”

Evaluating Customer Acquisition Strategies

35:39 to 36:52

Discussion on the evaluation of customer acquisition strategies and their long-term effects.

“does that actually apply to some of these companies?”

Future of Unit Economics in Startups

36:52 to 38:19

Exploration of future trends in unit economics and what startups need to consider.

“-time payment, even if their LTV to CAC is lower.”

The Role of Paid Marketing in Startup Growth

38:19 to 39:28

Analysis of how paid marketing impacts startups and their defensibility in the market.

“Like over time, they're going to be really valuable, even if you can't see it at the seed stage.”

Building Customer Loyalty and Routine

39:28 to 41:24

Discussion on creating customer loyalty and the significance of habitual usage.

“And that's not necessarily proportionally and size-satisfying of your user base.”

Internal Scorecard for Consumer Metrics

41:24 to 42:00

Overview of the internal scorecard used to assess consumer tech investments.

“Because when you become part of that routine, there is that strong lock-in.”

Understanding Portfolio Benchmarks

42:00 to 42:24

Learn about the benchmarks used to evaluate consumer tech companies.

“it's about sort of two thirds of our portfolio.”

Key Metrics for Investment

42:24 to 43:18

Explore the essential metrics investors consider for consumer tech startups.

“So looking at everything, obviously, from revenue, pricing, number of paid subscriptions, overall growth, percentage, if it is an annual versus monthly plan, what percentage is on annual.”

The Role of Customer Love

43:18 to 43:47

Discover the importance of customer love and engagement metrics.

“team is first and foremost a foundation of what every investor is investing behind.”

Surprising Instagram Insight

43:47 to 44:10

A surprising statistic about Instagram content sparks discussion.

“So I went to Gemini and asked, what should I know?”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00Sameer Singh:If you're trying to draw a common theme across all consumer software, its engagement is more important than retention, is more important than acquisition, is more important than monetization. The things I look for, in particular for any product with network effects, first, is engagement per user going up over time? That's the single most important indicator that something is working. It's a huge technological transformation, but it's a bit different from a new platform. Because as Samir said, the discovery and the way it engaged hasn't really changed. The user interface, if you will, hasn't necessarily changed from what's already here.

0:30So I think of AI as an incredible enabler. And we see that in our existing portfolio companies. I think if you're comparing a publicly listed company on something like Dow-Mow ratio, where they've already got massive scale established network effects, it's always going to be flatteringly high. So I think that can be a little bit misleading.

0:46Andreas Munk Holm:I want to throw you all a curveball. I saw a post a couple of days ago. lots of AI companies starting to experiment with paid marketing. So here's my take. Paid acquisition is a tax on your product's defensibility. The moment you can't outspend the incumbents and competitors, you die. Build channels that get cheaper, ask you grow, or you're just running your growth. I want to ask you, is that true? Is it not? Paid marketing, what's your take? Welcome to the podcast, everyone. So consumer is in a strange position right now. On the one hand, it's seen as a difficult, unpredictable, and in some cases, out of favor with investors space to be investing in.

1:19Andreas Munk Holm:On the other hand, when it works, outcomes are truly outsized. So instead of debating sentiment, let's make this fully concrete in today's episode. We're going to dive into what actually gets a consumer company funded in Europe, why it's not just narrative, but behavior metrics and thresholds that you need to be thinking about and knowing. And we'll try to structure this conversation into something that's truly usable for you. Following up on this, we'll end up doing something that looks a bit like maybe a napkin or something like that to navigate consumer, which is going to be exciting. But for now, let's look at our guests.

1:55Andreas Munk Holm:We've got Samir Singh, partner at Speedinvest, focused on marketplaces and network effects, previously at Abani. Samir, wave to the audience here. Hey, everyone. Susan, you are a partner at Felix Capital, investing across consumer and B2B with a background as a founder in social commerce. What social commerce firm? Well, it was back in 2011. It was a failed SoFillCommerce app, very early days of the App Store. So, yeah, it failed for many reasons, but humbling during. That was a tough time for everyone. And now, Joe, of course, Joe Siga-Dupoy, running Venture at True, focused on early stage consumer and consumer enablement technology.

2:35Andreas Munk Holm:Welcome to the show, Joe. Thanks, Ajax. So I thought we should start with talking a bit about where we stand today. And as I said just before, there's a lot happening in consumer right now. AI apps are really taking off. Social formats seem to be coming back, but not all of it translates into real business. Samir, I'd love to ask you where you see founders confusing what's popular on the one hand with what actually works.

3:00Sameer Singh:This is a long topic, but I'll try to keep it short. I think the biggest misconception people have today is that AI is a platform shift and you should build apps on it. Fundamentally, that is not how it works. AI is not like the browser or mobile. I increasingly think of AI as a microprocessor. Think of it as an input that enables your product, meaning you still have to build a product. And you can't just rely on AI to deliver the experience.

3:25Andreas Munk Holm:Samir, if you describe to the audience, what exactly do you mean? Give a concrete example here. How do people get it wrong and think that AI is a product?

3:34Sameer Singh:Okay, great. I'll start with why it's not a platform shift. That's where I think people really confuse it. The browser and the internet, that was a platform shift because as more people adopted browsers, it became more useful for developers to build new websites. And the more websites there were, the more people flocked to the browser, right? And so you kind of had a two-sided effect. Same thing with smartphones. As more people adopted smartphones, it became more useful for developers to build applications for the smartphones, which made smartphones more useful for end users. None of that is happening with AI.

4:06Sameer Singh:You don't interact with applications through ChatGPT or through Cloud. Well, developers have tried that. That's not really what's happening. What's happening is you're taking these APIs and you're sticking them into an app. And increasingly what's happening is that the end user experience is largely just tokens mediated through a thin app layer. And instead, what you need to do is think about this as another API you add to your application to give you powers that you couldn't have before. You still need to have some sort of multiplayer experience, a network effect, the economy of scale, something else on the product that actually makes the product the product beyond just the API.

4:47I think people sort of confuse that a huge technological transformation, but it's a bit different from a new platform because as Tamir said, the discovery and the way it engaged hasn't really changed. The user interface, if you will, hasn't necessarily changed from what's already here. So I think of AI is an incredible enabler. And we see that in our existing portfolio companies, a lot of companies have woven it into their products to deliver a much better product, more effective, more efficient in different ways. And I can go into details on that. But I think sometimes when you see AI native startups pitching, you see founders sort of confuse the two things.

5:24whilst I think that's all that's definitely all true I also think it's true alongside that to the optimistic side that it is fundamentally enabling new types of consumer experiences so even though the sort of rails if you like on it's not a platform shift from that perspective I think the capabilities to build really differentiated product is still there but you still have to do that right to Samir's point just and Susan's point just because it's AI is not enough right it's just it's a super powerful enabler but it has to come as part of a broader wrapper okay samir can you give an example of what is it that you see startups doing when they're doing it

6:05Andreas Munk Holm:wrong so to say when they're thinking that or founders of person what are they doing wrong when they're coming to you saying we're doing this because ai is a platform shift what is it

6:15Sameer Singh:worst the understanding i'll give you a very basic example like i don't want to name specific companies, but think about a company that generates audio or audio content or video content or image-based content that you then sort of end up sharing on social, right? So the underlying models can do that fairly effectively. And there's a couple of companies that have cropped up saying, you know, we'll allow you to create slightly better versions of this. But the thing is, because the underlying models are available to so many companies, there's a whole bunch of companies that can effectively create a very very similar experience there's not a whole lot your product is actually doing in terms of heavy lifting there's a couple of companies that i tried to build essentially a game where part of the input is ai generated so some of the elements that you interact with you can combine a couple of elements together to create a new ai generated element and then you continue onwards in your game the game experience is like a whole cohesive thing.

7:15Sameer Singh:And another couple of companies that have been trying to create some sorts of comic networks, but the panels are generated with AI. And that's where you have sort of both creators and consumers, or you have IP and consumers. So it's a cohesive product. You've created a product experience and AI is just sort of one enabler that's allowing the entire thing to exist as opposed to being the entire experience.

7:36Andreas Munk Holm:Instead of staying super negative here and saying the things that don't work, let's go to you, Joe. You mentioned just before, Or there are situations where you're like, this is a completely new thing that's now enabled or it's possible to do today because of AI. So say some of the hype-y stuff that is actually also real. Yeah, I mean, we've talked in the past around the opportunity to take markets where, for example, the supply side has been historically constrained by high cost of human labor. so you think about things like therapy or things like legal services or things like even private wealth management i think the opportunity to reduce effectively the marginal cost to serve the customer by doing that through software relative to a human means that that service can be delivered at a much lower price point and in theory open up the long tail of market that was not previously accessible to the sort of old operating model so that still requires a holistic product experience to be executed against.

8:37But it's the sort of thing where when we think about when you talk about the why now and is AI a why now and in what circumstances, that's an example where we get really interested.

8:48Andreas Munk Holm:Susan, I obviously got to give you the opportunity to also say, what are you excited about in this AI boom? No, I agree. I mean, I think, well, I think there's a lot of B2B use cases, but focusing on consumer, I completely agree with Joe that I think a lot of it is about democratizing what previously was incredibly expensive manual tasks. And if you think about what the underlying LLMs are best at, right, which is taking like different formats of data, you know, which can be very raw, can be handwritten, can be, you know, PDFs, completely unstructured and ingesting that, turning it into something that's very understandable.

9:19That is what, you know, that sort of that's the foundation. On top of that, you know, whether it's financial service advice, whether it's healthcare, I think health is an incredible area, both on the consumer side and the clinician and hospital side in terms of AI applicability and other areas, personalized fitness, therapy, et cetera. I think those are all incredible use cases.

9:40Andreas Munk Holm:I would love to ask the three of you, because one of my aims with this episode is that we're able to say a little bit about the fundraising market and what's important to be thinking about across the consumer spectrum. But the fact is that consumers, of course, everything from subscription marketplaces to social. So for that reason, I'd love to ask you a bit about how do you separate signal from noise within these segments? Are there in any way similar characteristics that make something look exciting across? But are there also, to a similar extent, are there specific things around, as an example, subscription where you're saying this is exciting, this is what really makes me tick when we're looking at subscription, whereas for marketplace, it's something else?

10:25Andreas Munk Holm:Susan, feel free to start. Yeah, sure. No, I think you're right. I think there are commonalities. Obviously, we actually have a – at Felix, we have sort of our own set of benchmarks that we've developed over the last 10 years of being in business across these different sort of business models. But I think when we think about sort of what's transversal and, you know, it can even extend to physical products or services, AI-enabled services, et cetera, I think one of the things we look for from very early on is signals of what we call customer love. In another way, you could say it's have you found product market fit.

10:58But what we like to see beyond sort of like product market features in terms of like usage or engagement or attention is really signs of customer enthusiasm, whether that shows up as the percentage of organic acquisitions or the rate of referral, word of mouth. with companies early on, we'd rather they go deep than broad. So even if you find the first 1 ,000 or 5 ,000 users or customers who are incredibly excited about your product, and it can in the early days seem a bit more like a niche, but I think broadening off from that, a very dedicated, excited base is much easier than spreading it across with everyone, but not really getting that much engagement or excitement back.

11:41Sameer Singh:If you're trying to draw a common theme across all consumer software, its engagement is more important than retention, is more important than acquisition, is more important than monetization. So you're broadly trying to follow that ladder of proof. The things I look for, in particular for any multiplayer product, any product with network effects, first, is engagement per user going up over time? That's the single most important indicator that something is working. So as an example, if you're looking at a Snapchat, is the number of snaps shared open per user per day going up? That's your first indicator.

12:18Sameer Singh:Second, is the retention improving over time and is the retention flattening at a high enough level? The first one is more important than the second, but broadly that's what you're looking for. Because again, as there's more users on the product, the experience should be getting better and therefore newer cohorts should be retaining at a higher clip than older cohorts. third again customer love so organic acquisition should be uh the biggest chunk and not a simply paid acquisition which can be noisy especially early on and sort of the last signal is monetization which i wouldn't even look at unless all the other signals are there i think that the most important thing is to to see improvement in those engagement metrics i use the trend rather than the absolute level that i think is the the most high information signal for an investor i think if you're comparing to publicly list a company on something like Dow-mile ratio, where they've already got massive scale established network effects, it's always going to be flatteringly high.

13:14So I think that can be a little bit misleading. I think the key thing is understanding how things are evolving over time, both temporarily when you look at how the next cohort is coming in, how they're performing on those key metrics, but also if you can segment the user base to understand what are the characteristics of the customers that are disproportionately engaged or disproportionately retaining, that then is a much stronger signal for us that we really like to see.

13:42Andreas Munk Holm:You're all saying the earlier the stage, the less data. If we're looking at a pre-seed startup and it's something that hasn't yet gone to market, how do you think about this? Are there any metrics that do you then try and say, then at least do some analysis and test customers? Or would you say, no, no, sometimes you just go with the idea and the team and you're comfortable with that? I mean, I think the brutal reality is it probably depends quite a lot on the founder's background prior to starting the company. Like clearly pre-seed, pre-launched consumer deals are getting done and will continue to get done.

14:19One of the trends and the shifts that we've seen in the last 12 months is that the theoretical barrier to getting started and testing something, whether that's vibe coding to mock up a prototype to get some level of feedback, some level of initial proof points around your fundamental hypothesis and your insight on what business you're going to build is. I think the bar is higher. And I think rightly so as well. I think, you know, as an institutional investor, we're investing other people's money, which we take very seriously. And so I think there's right to have a relatively high burden of proof and a threshold around that.

14:53And I think almost the first step is, you know, does the founder show the resilience and the resourcefulness to get some early feedback, some early signal from customers? Clearly in some segments, you know, we've invested in companies that, for example, are going through regulatory processes and things like that. Clearly you can't ship something quickly in spaces like that. But I think as a broad comment, I do think the expectation is there to try to have some initial science more than you know just a pitch deck and a narrative at any stage you need either narrative or numbers and if you can have both then it's always going to be better than just having

15:31Sameer Singh:one in today's day and age there's generally not much of a reason for you to not put a a basic mvp out there and get some validation often when that happens it's because it is easier to raise on the narrative than it is to raise on the reader. And it does happen for some categories. One of the challenges with consumer is that consumers tend to surprise you. They don't always do what you expect them to do. And so surveys, interviews can be deeply misleading. And the only real ground truth is behavior. And so especially if you're a founder who doesn't have a background, who doesn't have as many VCs chasing them, it's a great idea for you to go get some quick validation as soon as you can.

16:11Andreas Munk Holm:I'd love to ask the three of you, now you're sitting here throwing around acronyms and numbers and stuff, and I imagine there's a lot of pre-seed founders that haven't yet raised that this nomenclature isn't just normal for them necessarily. Is it so ingrained in to being successful as a consumer founder that the second you start thinking about product, you also got to start thinking about metrics and growth and how you track engagement and all of these things? Because otherwise you don't like you're not cut out for it. You don't have the right focus if you don't also have that in you by already at the start.

16:55I think exactly as Samir said, I think the only ground truth is consumer behavior. And we've seen that with both very early stage companies where they've done surveys with 500 or 1 ,000 people when they think it's in their ICPE, but the actual behavior turns out to be wildly different than what they expected. But also it happens for a lot of established companies too, where they launch a new product or launch a new feature and it doesn't go as planned. In the current development environment, and especially when you're looking at the whole suite of DevOps tools available where it's so easy to test things and to put versions out there, I think it would surprise us if founders didn't want to do that.

17:37And I think the other thing is consumers can be tricky to monetize. And I think if you're thinking about long-term, how do I build a sustainable business, you need to think early on. I don't think you need to monetize early on, but you need to start thinking about what are the North Star metrics for you, whether that's engagement or retention or something else, key factor, et cetera, and really focus on those. And that needs to be a data-driven exercise.

18:00Andreas Munk Holm:You're all nodding. So I know that you agree with Susan. Are there any adjustments, nuances we should add to it? To the previous point, one of the good things about consumers is data early. And so I think it is important that founders are using that data as a mechanism to learn very, very quickly on the extent to which they're actually solving the problem that they're setting out to solve. so I think it's not it should be something that they are actively leaning into absent any fundraise just to know whether the business is working and whether the product is resonating there's obviously a use case for metrics running up to a fundraising process where you want to have you know the your metrics clearly articulated and ideally be in the top quartile and everyone always ends up in the top quartile somehow but I think the broader step back point is those metrics should also be a useful way to almost have a health check on the business.

18:53They don't have a shelf life that runs out once the fundraising process is over. And I think that having that familiarity with what your Northstar metrics are, what is driving them, shows as part of the fundraise, the depth of understanding you have about the problem you're trying to solve and the customer base you're serving, which is valuable beyond just getting an investment or not.

19:17Andreas Munk Holm:Joe, I want to ask you, because as you said in the beginning, there's always a bit of a bias against consumers. I was about to say, at least it's been like that for many years, especially in Europe. How do you see that actually showing up when companies are trying to raise? Yeah, I think, like you say, I think consumer has become a bit of a dirty word over the last sort of five to 10 years, which is ironic because, I mean, there's not great data on the funding side because consumer is not really kind of mutually exclusive. but their estimates is sort of 15-20 % of venture funding but it's 60 % of the economy.

19:49So it feels like it's underserved and underweight when it comes to venture capital. And just to jump in, I mean, again, the sort of the largest outcomes in Europe, whether it's, you know, Spotify, Klarna, Revolut, et cetera, all very consumer. The largest companies in the world include over a trillion. I mean, you know, Apple, Meta, I mean, Google to some extent as well. All of these have a very large consumer business.

20:14Andreas Munk Holm:So why is this? If we spend 30 seconds here, why is it that consumer has a hard time in the venture scene? I personally think it's because it's the power law, but it's even more accentuated as a power law. I think the loss ratio or the rise-off rate is probably higher. And there's some data that for us, but I've shown that. But don't you just portfolio your way out of that? Like you just do a 50 portfolio? I think you can in theory, but I think that comes on to my next point, which is at the early stage, you know, unless you're funding these companies to profitability, you have to believe that the market is there at Series A as well.

20:50Continue the show, keep the show on the road, right? And we're now doing an exercise of mapping basically all of the preceded Series A funds that we know of in Europe. And so far we're at like 40, which is not a lot given that there's like a thousand, well over a thousand venture funds across Europe. So I think it trickles down because people know that the challenge is going to be there on the subsequent funding rounds. It increases the risk and increases the risk of like an early write-off, right? If a company runs out of runway after its pre-seed or its seed. The US market is a little bit more risk on in general.

21:22And so we often actually hear founders just getting on a plane and going to San Francisco to do their fundraise, which is like, it's obviously heartbreaking for us as a European consumer investor. investor and just as European, we shouldn't be having to outsource our capital markets to the US. So I think there is a, you know, one of the things. Hopefully we can collectively change that. Well, that's the thing. I think there needs to be a little bit of a, like, a kind of like rallying cry around consumer. It will be helped by some great outcomes, right? I think Susan's point, like Revolute is a consumer facing business and it's an awful hundred billion, even on coming out of Switzerland's consumer business probably listed in the US, right?

21:58So there's a fuel that got bought this week by Danone for a billion. These are great case studies that I think prove that consumer can deliver outlier outcomes in Europe. And hopefully that means that there's capital with its attractors into the market here in Europe to support these founders, because I do think it's underserved at the moment. Yeah. I'll just add one more nuance to that, which is I think that in the 20 mid-teens, or early teens, I think consumer became more synonymous with e-commerce and direct-to-consumer, which there's certainly been some great outcomes from that on being one example, as you mentioned, from Inure.

Read the full transcript

22:35But there have also been quite a lot of challenges there. And I think there were probably, not to name names, but a plethora of larger venture peers that funded a lot of businesses in that bit. They confused, I think, a business model with sort of other important attributes. And I think that, you know, direct-to-consumer business model in some ways got overfunded. And there were businesses that were, you know, fundamentally lacked sort of the great, you know, product market fit and the ability to acquire customers organically and retain them and to show sustainable unit economics, which then led to, you know, some write-offs, which made, I think, people overreact to that and then move away.

23:14And I think, you know, whether it's in consumer tech or in consumer, you know, even physical products, I think there are some fantastic outcomes. But you do, you know, you need to realize, you know, some of those business models don't scale in the same way as others. And that, you know, you need to be thoughtful, particularly around the marketing piece.

23:31Andreas Munk Holm:One of the things that you mentioned also early in this conversation that AI is truly changing is, of course, the economics of building and consumer. Are you seeing that consumer, the whole problem that you described before, Joe, around you needing oftentimes to fund the business very far down its path before you get attained profitability? That's kind of what we've heard about all the AI companies, that the beautiful thing is that you can get to profitability really, really quickly with a lot of models that you could never get to profitability with before, or at least it took many rounds before you got there.

24:06Andreas Munk Holm:So are we seeing more models now, business models, I mean, that you can actually get to market and get profitable with because of AI? Or do we still have the same problem in consumer? There are a couple of small use cases, but then they don't have the scalability potential. I mean, my perspective on that is probably two angles of getting to profitability, right? So one is like, does it cost as much to actually build a team, ship a product, et cetera, get to market? I think we're probably all seeing examples across our portfolio where people are doing much, much more with much, much less. So that's good.

24:41I think in terms of the can you very, very quickly ramp up the revenue side, we'll probably talk about monetization later on. That is harder in consumer. So I still think there is a constraint there. There are obviously examples of things, particularly prosumer, which has a lot of shared characteristics with consumer where you're seeing that happen much more quickly. but on pure consumer I still think that is a constraint but I'd be interested to hear what the other guys think yeah I think that I I agree that like the the building part of it is is there's certainly I think a lot of costs have come down I think they're the mark the acquisition cost which is can be a large chunk of of um of consumers companies you know cost base particularly at when they're they're in sort of hyperscale mode I think that is we haven't I mean I think there are a lot of different AI tools that are helping to you know make make content creation make make make marketing ads much easier to create.

25:35I think in terms of how does that impact actual CAC, it's, I think, still unknown. Or I think in some cases, probably still like a net-net sort of zero effect. But I think, and the other thing is, in terms of the cost to serve, obviously, I think one thing people sort of forget about is right now is given all the foundation models of competing with each other, the token costs have obviously dramatically been falling over the past one to two years. I think there is, at some point, there may be a reckoning where some of those costs get adjusted, especially as some of these companies go towards an IPO.

26:11So I do wonder how some of the inference costs will shake out in the long run. One of the things I've been tracking is the underlying hardware, the cost of the infrastructure to generate that inference.

26:26Sameer Singh:And that seems to be going up fairly rapidly. So I'm expecting inference costs to start ramping up. You're only seeing GPU rental costs to go up quite significantly, in particular for B200s, on the last couple of weeks. So the question of, it is true that AI companies generate revenue quicker in their lifecycle, but often that revenue can be passed through. And so the margins are fairly tiny and they're very sensitive to the underlying costs. And so there's a big question about kind of what happens on that front and if companies will need to work out new models to generate revenue. We saw with all the scooter companies,

27:08Andreas Munk Holm:the problem of having just inherently unsustainable business models and inherently unsustainable cost structures. How do you think about that at the early stage? If we look there first, if we look at the seed stage, does that matter at all? Or are you like, no, So we know that inference is going down in price. And now I'm assuming that most of the companies you'll be looking at at seed stage are probably AI companies. So that's probably a big driver. But let's talk about cost and the sustainability of business models. Through the stages, if we start at the seed stage, is this a problem there? How do you measure it?

27:47Andreas Munk Holm:What do you care about? And then if we go on to the later stages as well.

27:51Sameer Singh:at least from from my viewpoint one of the things i look for very closely is one how high are your marginal costs uh so there's lots of different kinds of costs if costs are one time i think that that's largely fine if the costs if you have very high marginal costs that's a problem especially if you don't have a clear moat of pricing power and the combination of that can really bite so if you don't have the ability to raise prices and you don't have the ability to lower costs like that's something that's going to break at scale and we kind of saw this with some other scooter companies we saw this with the 10-minute grocery delivery companies a lot of costs that were assumed to be one time or the quote-unquote declining costs were actually marginal and increasing costs um and samir how do you pressure test that uh the the pricing power the very basic test is that the number of moats that a software startup can have is like there about three or four.

28:46Sameer Singh:That's about it. There's network effects. There's economy of scale, which don't really work when there's too much capital sloshing around the ecosystem. There's genuine switching costs where once someone starts using the product, leaving it is either painful or creates a massive degradation of the user experience. And that's pretty much it. Brand is not really that much of a thing in software. At least I'm not a big believer in that. For hardware, I think it matters a lot more. But software, just because the products are so easy to create, it just sort of does not give you enough time to create a brand for that to be a primary mode.

29:20Sameer Singh:And so really what you're trying to work out is, are there fundamentally any switching costs or network effects here? That's about it. I disagree slightly on the brand thing. I do think there is brand and distribution mode, but I think, but I agree in the sense of, I mean, I think a lot of that is captured by the existing large players. So for you to be a, you know, to be a pre-seed stage company today and try to build that, unless you have some sort of unique advantage where that could be a celebrity founder, for example, or someone who has a massive own following and distribution relevant for that specific area.

29:55So for example, on the health side, we see Function Health that's scaling really well in the US. They have Mark Hyman who was their founders and that gave them a huge initial boost. Obviously, you need to figure out how to compound that over time, but that is one thing. But I think if you don't sort of have any sort of unique, trying to build that from scratch is harder in a world where there are a lot of existing combats who have that massive distribution.

30:20Andreas Munk Holm:Does it look the same at Series A and Series B and onwards, guys? When we're thinking about cost specifically, it's basically the same consideration? I think for Series A onwards, then you start getting a bit more. I think you get I get more tangible numbers so you know a lot of the for a lot of the the cth companies I look at and a lot of their I mean I looked at one the other day which said their their costs were zero and I like I was like guys you're I mean they they're they're leveraging a lot of LLMs I'm like certainly you have like inference costs and they were like and hosting costs and they're like no no we're getting all that free credit I'm like okay but that is not the right way to think about it so um by by series by series A you certainly like hope their uh the metrics are more bottomed up and more robust.

31:07For anything on the app side, there's still the app store taxes. So that's certainly a consideration as well. And then, and we have seen companies, by the way, quite smartly redirect that by actually acquiring more users on web or on desktop versus on mobile so that the blended COGS is lower because it overall does make a pretty significant difference of five to 10%. And then I guess the one where we've seen marginal costs go down quite a lot with AI is things that you previously, whether it was customer support or some sort of human touch point, if that was part of the product experience. And now a lot of that's being increasingly automated by AI.

31:53We've seen that in a lot of digital health apps and services, for example.

31:56Andreas Munk Holm:I don't want to ask you, and I promised in the opening of this episode, they would be talking a bunch about metrics, and we have already. I would love to ask you, and I just thought about this. It does make sense to ask you to go through all the different metrics that matter at each stage. Anyone can ask an LLM to find that. There's a bunch of articles about that, so no reason to go into that. I think what's more interesting is to ask you, what metrics have you seen important movements on in this AI shift? Just as we just now spoke about cost, we used to be thinking differently about cost. Now there's an inference cost that you really got to be thinking about and know how to measure and know how to put up against your different dynamics in your business.

32:40I mean, I think for us, the key sort of engagement ones, I think that the metrics themselves haven't changed. We are seeing sort of, I'm trying to think of like, has the benchmark shifted? And I think, you know, maybe I think in terms of growth, we're seeing more companies that are growing faster. But I think, you know, in terms of underlying metrics around, you know, we, depending on, obviously, on the product, you know, we look at engagement measured by either sort of like while-mile, down-mile percentage of active versus percentage installed. We'd look at time spent on the product, how many sessions engaged and sort of passively versus actively engaged.

33:16And then we also spend a lot of time on unit economics, I think. But I think those are still sort of the key ones we look at. So I don't think that they've changed materially.

33:27Sameer Singh:In some cases, we've seen lots of companies that have faster growth and lower retention. Like that's been a consistent pattern. But outside of that, that's more of what we're seeing in the metrics as opposed to which metrics we look at. I think what is very different relative to previous consumer technology businesses is that marginal cost of inference is quite a big deal. So I think if you haven't figured out a way to monetize early on, then actually fast growth, if you're at negative contribution margin, is just kind of driving you quite fast into a cash runway issue. So I'd say that is quite different to what we've looked at and cared about in the past.

34:06Andreas Munk Holm:How do you, though, think about the unit economics side? Because we had the hyperscale time where we just move fast and break things. It'll solve itself in the end. And then we all learned just before COVID, I think it was maybe right after COVID, sorry, with the tech reset, we learned maybe this isn't that healthy. And now it kind of seems, at least in some areas, to be coming back that we're like, it's so astounding what's happening. The growth is so astounding with each of these startups. So for that reason, like it will solve itself because we are sure that inference cost is going to come down in the end.

34:45Andreas Munk Holm:or as you said, Samir, it's going to be a hardware acquisition you do and then you don't have to run it on the big LLMs. How are you thinking about this today?

34:55Sameer Singh:On my end, I think often, I think a lot of folks learn the wrong lessons from the 2010s. I always recommend reading the book Blitzscaling. Some of the assumptions behind the model become clearer. One of the first rules of the book is do not blitzscale unless you have a moat. And it's not like the one thing people immediately forgot during the Zerper or during COVID. I think that's kind of happening again in some cases. Where you need to have, the idea of blitzscaling and going inefficiently is if you are the first one to scale and you have a moat, it's impossible for someone else to catch up. So if you're Airbnb and you have all the supply, if you're Facebook and you have all the users, that use case is not, it's impossible for someone to replicate that use case with that network, right?

35:38Sameer Singh:And so you kind of have to think about, does that actually apply to some of these companies? And in some cases, the answer may be, Yes, in a lot of cases, the answer might mean no. To Samir's point, during that era, we saw a lot of people sort of relying heavily on paid acquisition. And that works if you really have confidence in your cohorts and your long-term retention and you know that your LTVD CAD will be above X number. But I think a lot of people hadn't really confirmed that before they started pouring millions into meta ads. I think for us, we do look a lot at both payback in terms of, on the economic side, we look at overall acquisition breakdown by different channels.

36:26Of course, it could sometimes be hard to do attribution, but best guess. And I think certainly on a blended basis, when you look at it over time, you can get a good sense of what portion of that is driven by organic. And then we look at both payback in terms of timing and then also all TV to CAC. I think one of the things we've learned the lesson on actually is, you know, we used to be a bit more dogmatic on wanting to see LTV over CAC over a certain ratio. But actually, you know, we've seen in some cases, because it really can depend on product, is that, you know, for products that are able to monohize very effectively with a relatively high upfront, you know, whether it's annual subscription or upfront, you know, one -time payment, even if their LTV to CAC is lower.

37:11So let's say it's two rather than three, because they have that instant payback, they're able to recycle the cash. And so long as your TAM is large enough, you can still scale very effectively. Then hopefully as your product gets better and better, you'll even push the retention curve up or get a smiling curve. And ultimately, LTVD cash should also move upwards. But I think we've learned to not be overly dogmatic on that metric. One of the ways we think about it is like if your unit economics aren't working today, what do you have to believe over time changes, right? So either you're going to have incremental pricing power over time.

37:48So to some point you're establishing network effects, maybe brand is a contributor to pricing power or whatever, or you have to believe that your variable costs are going to drop over time, whether that's inference or whether that's scaling a fixed cost element of your cost to serve across a bigger user base, like the Netflix example, right? massive beneficiary of economies of scale. Or you have to believe that you can monetize that same customer in other ways. So there's like an LTV build argument. So yes, maybe on this initial product, it doesn't look great. But over time, we're going to layer in additional products to those same customers.

38:18And that's probably the Revolute story, right? Like over time, they're going to be really valuable, even if you can't see it at the seed stage. Or you're basically believing that you can fundraise better than everybody else and effectively outlast everybody else. and eventually then go back to the first point and have price and power because you kind of killed the competition through capital warfare. Like if one of those four things isn't true, then there's no reason to believe that your economics are going to change over time. And so you kind of need to have a pretty sharp hypothesis on which of those is your banking on, ideally several of them.

38:52Andreas Munk Holm:I want to throw you all a curveball. I saw a post a couple of days ago by Andrew Chen, who we of course all know from A16 said Speedrun. He wrote, Lots of AI companies starting to experiment with paid marketing. So here's my take. Paid acquisition is a tax on your product's defensibility. The moment you can't outspend the incumbents and competitors, you die. Build channels that get cheaper as you grow or you're just running your growth. Of course, all of it written without any capitalization as any good Silicon Valley VC does. I want to ask you, is that true? Is it not paid marketing? What's your take?

39:28Andreas Munk Holm:I mean I think there's nuance in the argument

39:31Sameer Singh:first I think that paid marketing and the moat are two different things if you have a moat and you have a sustainable growth flywheel you lay your paid marketing on that's a great way to scale lots of companies have grown that way but in the absence of one and you're trying to scale purely based on paid marketing like that's when it can hurt you because a lot of those costs scale as you scale those costs don't necessarily come down over time especially as the category becomes more obvious more competition comes in and so you can see a lot of artificial growth early on because of that and to be honest that's not just paid marketing it's also sort of viral content marketing which can really scale up to a point because you've got to figure out you need to be able to scale the percentage of your user base right and your content growth tends to scale based on how many views TikTok wants to give you.

40:26Sameer Singh:And that's not necessarily proportionally and size-satisfying of your user base. Again, there's a lot more nuance to that, which as we discussed earlier, in today's world, it's hard to put a content out there with nuance because we know it doesn't necessarily get the level of engagement or views that you'd like. But I think, yeah, to some point, I think there's a lot of great companies where if you can build, if there is an enduring mode, and I'll give one example again from a great consumer company in Europe, which is Aura, right? They, you know, they've scaled very impressively. I don't know if they're doing$500 million or a billion in revenue now.

41:00But obviously, a lot of that throughout the acquisition has been paid along with a lot of organic. And, you know, they have obviously a very visible product. But the great thing is, you know, there is a huge lock in once they have your data, once, you know, you rely on the data into a daily habit. And, you know, that's what we say for a lot of our consumer businesses we look at is, you know, can you really become part of someone's routine? Because when you become part of that routine, there is that strong lock-in. And then if you add on proprietary data and personalization on top of that, then it becomes very difficult.

41:33You create that strong switching cost and you have that customer. And so in those cases, performance marketing can absolutely be a great way to accelerate growth.

41:43Andreas Munk Holm:Susan, I think to really take us out of this conversation in a way that gets super succinct for everyone, I thought it would be cool if you would pull up the internal scorecard that you guys are using at Fili. to think about consumer? So we've been, as I mentioned earlier, over the last 10 years, as we've invested in a lot of consumer companies, it's about sort of two thirds of our portfolio. We've developed our own set of sort of benchmarks for different metrics. And obviously there's a ton of nuance here depending on the stage of the company, the business model, is it a marketplace versus a consumer subscription app or something else, physical product.

42:20But roughly sort of that, we look at similar types of buckets And so we have, you know, traction where we're looking at, you know, again, this is sort of more and late C to Series A stage, which is where I typically invest from a venture fund. So looking at everything, obviously, from revenue, pricing, number of paid subscriptions, overall growth, percentage, if it is an annual versus monthly plan, what percentage is on annual. If it's a freemium model, the paid subsidy of MAU, et cetera. Then we look at unit economics, which we talked a lot through before, everything from blended CAC to one-year LTV, fully loaded gross margin, including inference costs, hosting, payment, app store costs, et cetera, LTV to CAC, payback time.

43:02And then on the customer love side, obviously, we look at things from mal, dal, wow ratios, retention over time, percentage of organic acquisition. And then obviously, very importantly, the other elements, which here looks small, but team is first and foremost a foundation of what every investor is investing behind. So incredibly important. And then we also do look at sort of market and category just in terms of how competitive it is and how large it is.

43:31Andreas Munk Holm:There's so much we could have spoken about. Now on this one, you have a customer love, Instagram follower base, and you have Instagram and TikTok engagement rakes. just last night, I was fed up with social media. There was no more Denmark election campaigns that I could follow up, you know, that I could lie there and scroll about. So I went to Gemini and asked, what should I know? And what should I be reading? And what it told me was, did you know that 77 % of all Instagram content that's put out now is actually made by butts? And I should have brought that for this conversation because that made me really think, especially as a creator, of course.

44:10Andreas Munk Holm:So we didn't make it to cover that. I think we could have done so many more episodes on each one of these. But Joe, Susan, Samir, thank you so much for joining me today. I think there was a lot of learning in this episode for all the amazing, beautiful consumer founders out there. Thanks for having us. Thank you so much for having us. And for all the founders out there, we're rooting for you. Come talk to us. We're excited.

44:36SILEN Gemeente

From the publisher

What actually gets a consumer company funded in Europe? Fewer things than most founders think.

This is one of the questions the first episode of Consumer Tech Napkin explores, with Andreas Munk Holm joined by Sameer Singh (Partner with Speedinvest's Marketplaces and Consumer team), Susan Lin (Partner and Investor at Felix Capital) and Joe Seager-Dupuy (Director, Investment at True).

The conversation covers engagement as the real leading indicator, why a decade of cheap capital let weak products hide behind paid acquisition, what behavioural signals actually move investors and why AI is not the defensibility play most founders assume it is.

If you're building consumer, this one's worth your time.

Key highlights

  • Engagement is the strongest signal in consumer software, not monetisation
  • Growth can hide weak businesses when retention and organic acquisition are missing
  • AI alone is not a moat and thin wrappers are easy to replicate
  • Consumer is underfunded in Europe despite producing many of its biggest outcomes
  • Blitzscaling only works when companies already have defensibility

Timestamps

  • (00:00)⁠ What actually gets a consumer company funded in Europe?⁠
  • (02:40)⁠ Why AI is not a platform shift
  • ⁠(04:15)⁠ Thin AI wrappers and defensibility in consumer software
  • ⁠(07:30⁠) AI-enabled consumer opportunities in health, therapy and financial services
  • ⁠(10:00)⁠ What investors actually look for in consumer startups
  • ⁠(11:20⁠ Why engagement is the strongest signal that something is working
  • ⁠(15:10)⁠ Why behaviour matters more than surveys or narratives
  • ⁠(19:00⁠ Why consumer remains underweight in European venture
  • ⁠(27:00)⁠ Marginal costs, pricing power and scalable business models
  • ⁠(34:30)⁠ Why blitzscaling only works when companies already have a moat
  • ⁠(39:00)⁠ Paid acquisition, defensibility and sustainable growth
  • ⁠(41:30)⁠ Felix Capital’s consumer investing scorecard

Subscribe to EUVC, the home of European tech, for more insights: https://www.eu.vc/subscribe

More from EUVC

All 626 episodes
Episode #1: Consumer Tech NapkinEUVC · 45 min
Listen in VO