Can Tokens Solve The Cold Start Problem? An Interview with Sameer Singh
Sameer Singh is the creator behind breadcrumb.vc, a blog providing investors and founders with resources on everything network effects. He is the creator of the Applied Network Effects course, one of the highest-rated courses on Maven as well as an Atomico Angel who exclusively invests in Pre-Seed and Seed-stage startups with network effects.
Sameer will be hosting a special masterclass at The Marketplace Conference in Berlin on October 11th where he will discuss whether or not tokens are a viable way to solve the “cold start problem” on Web3 networks. Leading up to the session, I sat down with Sameer and learned why tokens aren’t necessarily the magic solution for every Web3 marketplace.
Listen to our full interview or read the highlights from our chat below.
(This interview was lightly edited for clarity and brevity.)
As part of the Atomico Angel Program, you invest exclusively in network effects businesses. What are the main criteria and KPIs you focus on when evaluating these businesses?
I used to call myself a ‘horizontal’ specialist versus a ‘vertical’ one, meaning I am very agnostic on the idea per se or on verticals, while I am very very focused on the horizontal aspect of it. What is the long-term potential of this network effect? How is this network effect being executed?
You can look at it is from three angles:
- Is this network effect viable?
- Is this network effect scalable? As you build this company and as you scale, do your unit economics actually get better or are they likely to get better?
- Is this network effect defensible? Once you scale, what is the likelihood that somebody can clone you and effectively compete with you?
What is the one piece of advice that you would give marketplace funders who are trying to solve the ‘cold start problem’?
The one piece of advice to solve the ‘cold start problem’ and your goal on day one is not growth, but to build a Minimum Viable Marketplace. This is what Andrew Chen calls an Atomic Network in his book, The Cold Start Problem.
You want to have the smallest possible marketplace where you have liquidity. That is, transactions flowing back and forth. So assuming you have 100 users and they're doing 1,000 transactions, that’s fantastic! On the other hand, if you have 1,000 users and they're doing 10 transactions, that’s terrible.
The scale of your user base does not matter. Your GMV does not matter. All that matters is liquidity. You want to see strong engagement and retention upfront and, once you crack that, then you grow.
Other two key pieces of advice are:
Track everything. When you're building a business with network effects, your users are the product, while what you are building is simply an enabler. What this means is that your metrics really show the story, and therefore, it’s of utmost important to track them on a granular level. In particular, it is important to track what I call ‘Retention to core action,’ meaning what a user is meant to be doing on that product. How often do they come back and do that action again? What's the propensity of them to come back next month or next week and do that?
Be careful about incentives. Incentives can often skew the behavior of users. In the words of James Currier, you want to find your ‘White hot center,’ meaning those users who are starving for a product like this, because it solves a key problem for them. Fundamentally, incentives are not a great way to find those users because incentives will attract users who are not as married to the idea of solving this problem, which will reflect in poor engagement and retention, and make you unable to build a viable network or market.
If we apply this to Web3, tokens are seen as a way to solve the cold start problem, but it’s not that simple and the fundamental rules of networks do not change. Without spoiling the topic of your live talk too much, what are these fundamental rules?
If you stick to finding your white hot center, the most underserved users for your network, then tokens aren't necessarily the best way to attract those users.
To better explain that, let’s look at the idea of a token. It’s a financial incentive for someone to join a network early on because they will get a big upside. Now, a financial incentive appeals to everyone, and therefore not necessarily to the most underserved users. So let's say you use your token as an incentive and manage to attract 100,000 users. What are the odds that x percent of them are the most underserved users? And what will happen then?
Let’s say this is a marketplace and you have 50,000 sellers and 50,000 buyers. They’ve all signed up to your product. But if they're not the most underserved users, then they’re not deeply engaged with your product no matter how much they’ve bought and traded the token.
On the other hand, let’s look at the most underserved users who come onto your product. For example, a buyer. They're looking through these listings from these 50,000 suppliers and trying to engage with them. They might not be able to have a successful or a positive interaction because most of the people on the supply side do not care. They are not the most underserved users. Therefore, the odds of them actually finding a supplier who is underserved and who really cares about solving this problem, is pretty low. Essentially what you've done is polluted your product with an awful lot of users who shouldn't be there right now.
Of course, there are certain types of product where this might not be a problem and where you don't need a lot of involvement from the people you're trying to attract. In those cases, it's fine and the token will help you bootstrap the network. But be very judicious about what type of product you have in your hands and the way you are using tokens to bootstrap your network.
The Marketplace Conference 2022
Want to hear more from Sameer on how to think about liquidity on Web3 networks and why using tokens to solve the “cold start problem” is not as straightforward as it appears? Join us at The Marketplace Conference and sign up to Sameer’s Masterclass ‘Why tokens are not magic’ LIVE in Berlin on October 11th at 10:45am CET.