The Cold Start Problem and Network Effects in Financial Services
When do you need them and when do you not?
I recently finished reading The Cold Start Problem by Andrew Chen, an investor at Andreesen Horowitz. The book is a deep dive on network effects in commercial (mostly tech) contexts, and doubles as a growth and marketing handbook. The book has instructive and fascinating case studies from various companies and sectors of interest, and I of course found the financial services case studies especially interesting. Chen’s financial services oriented case studies focus on credit cards, credit bureaus, PayPal, and coupons. In this post, I’ll summarize some of the key learnings I had from the book with respect to financial services. In general, I would recommend the book to anyone working in internet or technology-enabled businesses.
Chen defines network effects as simply “what happens when products get more valuable as more people use them.” The first learning I had is that financial services businesses that have many-to-many interactions, most notably any form of payment method or payment network, clearly benefit substantially from network effects. As more users are onboarded onto a payment method or network, it becomes more useful to anyone using the network. Think about Venmo; only useful if all of your friends and peers are using Venmo. Credit cards; only useful if the merchants you are shopping at use credit cards. Visa and Mastercard are both worth hundreds of billions of dollars because they have created global payment networks that practically every consumer and merchant uses.
The challenge with networked products is getting to the scale where the network becomes useful and valuable. Often, this scale needs to be substantial, but Chen illustrates that all you need to really get started is an “atomic network,” defined as “one stable, engaged network that can self-sustain,” and then you can build adjacent networks to the first atomic network. The example given here is Credit Cards, where Bank of America used Fresno, California to build its first atomic network, and grew from there.
Another takeaway from the book, more on the growth marketing side of the discussion, is that consumer products can use existing scale or financial incentives to acquire many customers quickly, but companies have to be careful that the value of those customers makes that customer acquisition worthwhile. Over the past few years, we’ve seen so many consumer fintech and crypto products offer extremely generous incentives to acquire new customers (think about crypto deposit products with wonderfully high deposit interest rates), and it hasn’t necessarily worked out well. Now as we enter a different macroeconomic environment, it will be especially important to estimate the value of customers through a full economic cycle to get a true view of their expected value over time and inform how much you are willing to spend on customer acquisition. This reminded me of a belief I’ve long held, that there’s really no substitute for strong organic growth.
The final, somewhat counterintuitive, takeaway I had is that there are many financial services businesses where you actually don’t need a massive network and/or scale to build a great business. Many financial services interactions are one-to-one, such as banking or lending relationships. While scale can help with diversifying risk and getting better data to inform lending decisions, you don’t necessarily need to have or want peer-to-peer interactions in these types of businesses. If you figure out how to effectively and economically target a specific customer segment in one of these types of businesses, large scale or rapid growth might not be optimal. There’s a reason there are a lot of happy community bankers in this country.