AS LEARNED THE HARD WAY
Margins, markets and regulators
Late in 2019, I founded Jupiter Save with my friend Avish Brijmohun. Our mission was to make ordinary people addicted to putting money away instead of spending it.
Our strategy had two legs: a delightful app that made saving money easy and fun, packed with games, nudges, confetti and rewards; and bundling small saves into large ones, to give ordinary savers the interest rates and liquidity terms only available to rich people. I had built a user-facing platform, Grassroot, to several million users, with very high growth and referral rates; Avish had a decade’s experience in banking, including reward programs, high-income wealth and investment products, and treasury management.
We launched in March 2020, as the pandemic struck, and built strong traction for six months (~1,000 savers, almost all organic and no press, balances climbing 15% MoM, avg >1.5 saves per month). We had data, and the data showed behavior was changing. The 300+ experiments we ran were starting to pay off.
Then late in 2020, the South African financial services regulator (FSCA) shut us down. It’s a somewhat tortured story, but it came down to: (i) their licensing division was unwilling to give any ground on impossible (for a non-incumbent) administrative and capital requirements, and (ii) their “fintech” or “sandbox” team supported us, but lost the turf war.
There are a lot of “what I learned” posts out there, so I won’t repeat the most common lessons. They’re all right: ship early, experiment quickly, focus on getting to product market fit, start your raise and networking early.
Instead, I’ll focus on a handful of lessons that I haven’t seen as often, even if some are quite specific to fintech. Some might seem a little obvious in hindsight, the result of a little naivety, but hopefully they’ll help others avoid some painful lessons. I’ve kept it to the top three: on margins, markets, and regulators.
Focus is nice, margins are better
Our strategy for Jupiter always involved, at some point, adding higher margin financial products to our core money market / saving offer. But early on, we decided on a narrative that we would scale up first using the vanilla product, then add the higher margin ones later.
Early feedback reinforced our choice — people who otherwise were not fans of the deck would say “that slide on focus really resonated”.
Later on, the idea of focus became baked into our self-conception of our strategy. We thought about changing it a few times. Each time that early feedback and the oft-repeated advice of “keep laser focus until you scale” reinforced our natural desire to just do savings. And, structurally, savings is a low margin business. So we stuck to “we will focus on getting this core product to scale, and take low margins on huge scale”.
In retrospect, that was a mistake. Many potential raises, including our final round YC interview, ran aground on this point. “Looks like a great team, your numbers look great, but those margins, in emerging markets — it seems just too hard”. That was a common phrase.
The low margins put off more investors than the focus point attracted. This wasn’t universal and eventually we did get several investors close to signing onto our seed, but it took too long — just long enough for the regulatory hammer to fall, and scrap the round. The focus point helped, but it didn’t keep as many investors in the funnel as the margin point made drop out.
Moreover, we could have told a margin and product diversification story right away, and operationally kept a laser focus in the near-term — or long-term. Sequoia recently made public DoorDash’s Series A deck. At the pitch’s core was a plan to add lots of other products to delivery, which is low-margin, by becoming a local logistics platform. That never happened. DoorDash just focused like crazy on getting to scale in delivery in the suburbs, on a (very very) low margin core product. Their investors aren’t complaining.
Of course, if DoorDash hadn’t executed so effectively, then their investors probably would have insisted on diversification. But having the multi-product, higher margin play on the table not only made a more compelling story, it added some robustness — if the delivery margins had been just terrible, or the suburbs hadn’t worked, adding other local services would have provided resilience through optionality. Focus is great, but not enough to sacrifice your margin story or your options in a crisis — no matter how resonant the slide.
If you’re trying something hard, start in a big (or at least a good) market
Jupiter was always intended to work across many markets. The strategy was to start with South Africa as a “test market”, achieve the scale to raise a Series A, then enter much bigger emerging markets. We knew South Africa, were living there, and the model of testing in a small market before entering big markets has been used by some of the larger tech companies.
In retrospect, that was also a mistake. The top 10% of South Africans earn 70% of South Africa’s income (it is the most unequal country on the planet — yes, even including wherever you are), so even when we were getting traction with ordinary savers they just didn’t have much to set aside. The funding pool in SA is also quite shallow and some ostensible competitors had already reached the one or two funds interested in risky fintech at the seed stage. Likewise, when the regulatory bus hit, and we had to look for partners, the pool was extremely limited — the handful of big banks with no real interest in reducing spend and raising savings (again, whatever they like to say in public), and a handful of asset managers.
If we had decided to launch first in a much larger market, we would have needed to do a lot more learning upfront, but we then would have had a much better option space in both funders and partners. We would have been significantly more resilient. We did try to find another market after the regulatory incident in SA, having talks with people from Malaysia to India to Brazil, but without the prior lessons and without the ability to travel, it was just a bridge too far. If we had started in a more attractive market, with larger scope, more partners, and a good funding landscape, it would have been much easier to pivot, whether to a new model (B2B), to a license-borrowing model, or in many other ways.
Watch what the regulator does, not what they say
Talk is cheap. Just because the regulator says they want to do whatever your product does, that doesn’t mean they actually will facilitate it. Some regulators really will act in a way that matches what they say; others won’t; what they say or write, to the press or to you, provides no reliable evidence either way.
So when trying to judge how a regulator will act, focus solely on what they do and have done. If they give a license to every payday lender who walks through the door, let obvious fraud continue for a year, and almost never grant a license for a new type of asset administration — well, it doesn’t matter how much they say they don’t like predatory lending, they very likely will license your debt product. Conversely, it doesn’t matter how nice they are in meetings about your savings idea, they won’t license it. Watch action, not words.
One particular trap to avoid is to think, “well, that regulator over there acts in a way that will stop us, but this regulator here, my regulator, is different”. Regulatory bodies swap personnel all the time and exist within the same political, social and economic structures. So they share the same regulatory culture, even if with slight nuances, and culture largely determines action. Observe how the entire regulatory complex acts, filter out talk, and plan accordingly. This lesson is probably obvious to many people who’ve dealt with regulators, but startup founders are so wired for optimism, and regulatory cultures are often so set up to make it look like you’ll get permission, that a lot of traps get laid — even with the best of intentions all round. If you have the ability to choose where or how to launch, do so on the basis of regulatory action, not stated policy goals.
In sum, we knew we were trying something very difficult. We knew there was a reason only a handful of large-scale attempts to bring good savings products to middle and working class people had succeeded, and only really in the US and China.
We thought the way to tackle that problem was to keep laser focused and start in a small market with weak competitors and with regulators making supportive noises, and build from there. In the end that left us fragile and hard to pitch. In retrospect, we would have been far more alert to regulatory possibilities in reality; should have started in, say, India (even if it’s crazy competitive); and should have been aggressive in thinking about and talking about a future product roadmap full of high-margin, extensible options, even while executing on focus in the near-term.
A last word is that I remain absolutely convinced of the opportunities in this space, perhaps more so. Jupiter’s run brought us into contact with a bunch of incumbents, in a bunch of markets. It was eye-opening. If you’ve ever encountered a traditional bank’s IT department, or its management, or the product or “innovation” or “digital transformation” people at the old banks, you know Ant and Square and all the others coming are no accident.
Jamie Dimon knows what’s coming, but even he can’t do much about it.
Most other banks will still be saying “we can do that if we tried, but it’s not profitable, why would we bother” or “these guys can’t scale” on the day their last valuable client leaves. In small, low-growth markets with captured regulators, that day may never come. But in good markets, as a partner at a16z put it, “there’s never been such energy in fintech”, and for good reason. Periodic holding action by regulators or injections of free money and bail-outs of one form or another will slow the assault every now and then, but not for long. Truly innovative fintech will be a lot of fun in the next few years, and help an enormous number of people, and be very lucrative, and that doesn’t come along often.