June 2023: Subtle Differentiation (or why there are no 10x products in fintech)

Some times when you can't be 10x better at one thing, you can be 2x better at 10 things.

In the early days of the Cash Card, people sometimes asked why anyone would choose the Cash Card over a BofA card (in a lot of cases they’d also ask why anyone would choose Cash App over Venmo). I thought a lot about how to answer this, and within the 10x model for differentiation [1] (which was dominant in 2015, and I think remains dominant in software industry narratives to this day). I didn’t have a good answer. My answer at the time was that we weren’t going to be 10x better at one thing, because it just wasn’t possible. But we could be 2x better at 10 things. Basically no one bought this. 

To win in financial services you need to deliver differentiated money benefits vs the alternatives. By money benefits, I mean delivering a benefit that the customer can’t get if they tried to use another institution to store or move money. In financial services, you’re either storing money, moving money (between counterparties, or in exchange for goods or services) or delivering returns. Information products (eg analytics, insights, alerts etc) are nice to have, but no one has ever built a large, lasting consumer or SMB facing money business without doing one of those 3 things. Viewed through this lens there are only a few kinds of differentiated financial benefits you can create:

• Store money better (eg provide higher APR on deposits, increased/faster funds availability etc)

• Move money faster (eg settle securities sales faster, attribute paychecks or social security benefits faster, optimize withholdings better, etc)

• Save money for the customer (eg lower fees, sometimes on money storage, sometimes on transactions, sometimes both) 

• (Maybe) Reduce complexity [2]

• (Maybe) Increase access [3]

Every breakthrough financial services product has to succeed at one (and often, more than one) of those things. I have high conviction this is true for consumer finance and SMBs. However, for enterprise for example, you might be able to break through with something like an ERP integration that’s higher quality along some axis, or lower cost or better conversion, and for developer products there’s probably something about reducing complexity and giving better control, but I don’t have a strong view yet of what makes these work in a generalized way. If you look at all the breakthrough consumer and SMB financial products that have achieved mass scale over the past decade, you’ll see (not an exhaustive list): 

You’ll notice, none of these financial benefits are a “10x” benefit. Like none of them are literally 10 times better than the alternative. Paying $0 to trade stocks is definitely better than paying $7 per trade, but it’s incremental, and once you hit $0 you can’t sustainably go lower without destroying your unit economics. Giving access to paychecks 2 days early is definitely better, but once you hit 0 days, you can’t sustainably go lower without taking credit risk. Faster paychecks and lower cost trades are definitely better, but they’re not like “iPhone beats Motorola RAZR”[4] or “reusable rockets” better. (Also, I don’t mean to downplay the success of these products and companies - most of these products also won because they caught a wave[5], and did lots of other things right along the way).

2 ways to differentiate

I think there’s two ways to differentiate; 

• You can cut through the clutter by doing 1 thing 10x better [1]

• you can do 10 things 2x better (what I call subtle differentiation)

Differentiation in this context means doing something materially different (i.e. different in a way that moves the needle for your customers or your business) and better than others. 

Cutting through the clutter

Cutting through the clutter means you do 1 or more things 10x better than alternatives[6] for a particular customer segment and along a particular axis. 

One recent example that comes to mind is OpenAI. From the outside looking in, ChatGPT just straight cut through the clutter; literally ANYONE could use it, and it did things that were so obvious that you could not get from anywhere else, and it was so good at them that if you told people about it, they would think you were crazy then go try it themselves and be flabbergasted. By cutting through the clutter, you provide a level of novelty and utility that was basically unimaginable to your customers before they encountered it. 

Subtle differentiation

Prior to the last 10 years I only really thought of differentiation in terms of being 10x better. The iPhone was so materially better than the Blackberry and Nokia, Google was so obviously and materially better than Yahoo, etc. As a user of these products, it was blindingly obvious. In practice, at both Cash App and Carbon Health, I’ve seen a more subtle kind of differentiation where we’re 2x better at 10 things, than our competitors. At Carbon Health:, 

• We’re vertically integrated from tip to tail; combining an in house clinical, operational, and software team, and a single stack that manages almost every operational (revenue cycle, practice management etc) and clinical function (labs, imaging, prescriptions, charting etc) required to run our clinics

• We have patient scheduling built into the EHR. Most EHRs don’t

• We were the first EHR that had a single patient object across multiple practices (this is why you could go to UCSF and then across the street to Sutter but still require separate MyChart accounts, even though the app is the same)

• A patient’s full electronic health record is in their app

• We’re fast (we ship multiple updates daily, incumbents update every couple of years)

• We calculate your full patient responsibility up front, and tell you before you leave the visit

• And now, we’re hands-free with GPT4 [7]

At Cash App, we were also 2x better at a lot of things:

• We were the only way to instantly send money between consumers bank accounts, for free, from ~2013 - 2017 (you could use Venmo, PayPal etc for balance to balance transactions instantly, but from bank account to bank account it was just Cash App)

• We issued you a virtual card instantly when you opened an account

• We enabled you to instantly provision an Apple Pay card digitally when you opened the account (I think we were the first fintech to do this in 2017, and only the second issuer behind US Bank)

• We charged no fees on the Cash Card

• We instantly settled funds from btc and equity trades, etc

• We provided a beautiful, customizable card, that for our demographic, was a more desirable indicator of status than a Green Dot card for example (to be clear this was not a financial benefit, but it worked. I still don’t have a good explanation for why, even though some of my colleagues predicted it!)

These individual benefits are/were better than the competition at the time, but incrementally better, not by a step function. None of these, by itself, is so head and shoulders better than the competition. They’re all incremental, but instead of mattering all at once (because customers instantly understand the benefit and sign up), they compound over time as you penetrate the awareness of more customers. 

The problem with subtle differentiation

A huge advantage of cutting through the clutter is the benefits of your product are obvious and customers instantly get it, before using it. Without this, you run into a few big issues:

Knowing whether you’re right can take a very long time

When you build something that cuts through the clutter, you quickly know if you’re right or wrong, because the right customers, upon hearing the benefit, adopt your product right away as the benefits are so obvious to them. When you build something that’s subtly differentiated, you spend a lot of time explaining to customers and partners why it’s better, and even then there’s a high likelihood that the specific benefit you’ve built is tangential to them (because each of these 2x improvements probably only applies to a subset of specific customers, and you likely only launch with one). As such, they have to try it to believe it, and even when they do, it does not result in a step function/order of magnitude improvement.

Each 2x benefit only appeals to a niche

A simple example of this; at Carbon Health, we’ve built tooling that makes it super easy for providers to do medical coding, and for us to audit and automate it. This matters a little bit to patients (eg because on the margin your providers are more productive so they can spend more time with you as a patient) but matters a hell of a lot to providers, as it makes a tedious part of their job much easier, even though it doesn’t directly improve patient experience. Similarly for Cash App - there were customers who came for peer-to-peer payments (because their network was there) and didn’t care at all about debit, stocks, or bitcoin.

At scale though, the aggregation of multiple 2x benefits does compound. I think this is because each 2x benefit attracts different niches, or enables you to win smaller channels that competitors don’t focus on, with a more modest promise. 

Customers have to try the product to get the benefit

Instead of just making an eye popping statement that clearly communicates why your product is better, customers have to try the product to get the benefit. 10x products, when they work, have the benefit of being so obvious that they are impossible to ignore. For everything else, you need active sales, marketing, and customer education.

I’m not sure why its hard

It is super hard to build something that’s 10x better in financial services and in healthcare, and I’m not sure why. A few reasons I’ve thought through so far:

First, I think that because these industries are regulated, you have to invest a lot in getting the table stakes right (like if you’re going to provide loans to borrowers, you need a lender, that entity is regulated and requires you to be compliant or risks losing their charter, etc). In order to be compliant, you have to invest in table stakes, so a lot of that VC round you just raised goes to building stuff that incumbents and competitors already have, and doing those things well might differentiate you vs. other startups, but it won’t stand out against incumbents[7]. 

Second, a lot of types of competitive edge  are quickly copied, or spur competitors to otherwise close the gap. For example, to increase the returns on risk free deposits (which are currently around ~5%) by 10x, you’d have to find something to invest in that will consistently yield 50% annual returns at scale on a functionally infinite deposit base, with no risk of principal loss. Even if you did this (and you should look with a skeptical eye at anyone who claims they’re doing this), one of 2 things would happen: either you’d quickly acquire 100% of deposits on the planet, or your competitors would bid away the return and pay more for that asset until the return on capital invested dropped to the right risk adjusted rate. Not happening. The rule of thumb I used in fintech was that no edge lasts more than 18 months; ie from the time you launch your differentiated product that has customer traction, you have a maximum of 18 months before your competitors catch up (and from what I can tell that remains true today).

Lastly, in healthcare broadly, the core atomic unit of value you provide is healing a patient; you’re either providing care, or supporting someone who is. As a result, a 10x improvement will involve either 10x the healing, or reaching 10x the number of patients. The closest analog to a 10x improvement over the alternatives is probably biotech (like inventing a blockbuster drug), but to do that you literally need a scientific breakthrough. And even if you had a scientific breakthrough, the skills, discipline and capital required to bring it to market differs wildly from that required to build software or to deliver healthcare services, and anytime you’re mixing multiple capitalization models, you take a lot of risk (because your investors have literally different expectations, timelines and hurdle rates to deliver against).  But if you’re not commercializing a breakthrough, subtle differentiation is probably what’s left for you.

Thanks to Nico Chinot, Jaren Glover, Audrey Kim, and Aaron Frank for reviewing this in draft.

[1] Not sure if this is the origin of the 10x model for differentiation, but it’s the one I remember: https://bothsidesofthetable.com/your-product-needs-to-be-10x-better-than-the-competition-to-win-here-s-why-6168bab60de7

[2] There’s something around transaction complexity that I haven’t yet been able to articulate; for instance I think some of Coinbase’s early success was creating a “sanctioned” crypto trading product so regular consumers didn’t have to learn what a wallet was. Similarly from the outside looking in, some of Robinhood’s success was unlocking options trading on mobile which is a fairly complex interaction in trading, and often gate kept. Not sure how to generalize this yet, but in general, there are not that many financial interactions that remain highly complex for consumers (mortgages and foreign exchange come to mind as two possibilities).

[3] Increasing access typically means things like being able to underwrite customers that the rest of the industry cannot, or has ignored, or taking existing financial instruments and fractionalizing them (Eg buy a slice of Apple for $1). I think of these as tools, not so much benefits (mostly because its not a given that increased access has been unilaterally better for the consumers.

[4] Dating myself here, but: https://en.wikipedia.org/wiki/Motorola_StarTAC

[5] I’ve written previously about the concept of wave hunting: https://kunle.app/june-2020-wave-hunting.html

[6] “Alternatives” here refers to whatever customers would use to solve the problem if you didn’t exist, which may not be, strictly speaking, a competitor. If you consider your value prop exclusively against your  direct competitors it’s easy to miss adjacent things you could do to deliver value to customers.

[7] For what it’s worth, it’s possible vertically integrated LLMs might result in 10x benefits for a period of time for the first movers, but this isn’t true today.

[8] I think this is one of the sources of appeal for the concept of “embedded [insert category name here]” (eg embedded fintech, embedded accounting, etc). It’s appealing to utilize an embedded fintech provider because in theory, they bring you all the table stakes in an API call, and you can focus your resources on differentiation. In practice, this only works “some” of the time.