Why and how to start a for-profit company serving emerging markets

Link post

Wave (1) is a for-profit, ven­ture backed startup build­ing cheap, in­stant money trans­fer to and within Africa. Since launch­ing in 2015, we’ve be­come by far the biggest re­mit­ter to Kenya and Ghana, sav­ing our users and re­cip­i­ents over $100 mil­lion so far. Our biggest source of ex­pected fu­ture im­pact is build­ing mo­bile money sys­tems within Africa, which will have an or­ders-of-mag­ni­tude big­ger im­pact if it suc­ceeds.

Wave’s mis­sion is to im­prove the world, not to make money. De­spite that, we op­er­ate more like a tech com­pany than a so­cial en­ter­prise. Our in­vestors are ven­ture cap­i­tal­ists try­ing to make a high re­turn, and they hold us to the same stan­dards of growth rate and unit eco­nomics as any de­vel­oped-world startup. This might seem like a down­side (surely it would be eas­ier to di­rectly op­ti­mize for im­pact rather than have pres­sure from in­vestors to make money?), but for us it’s ac­tu­ally in­creased our im­pact in two ways. First, the pres­sure to grow quickly forces us to make our product bet­ter and scale faster, so we help more peo­ple by a larger amount. Se­cond, since we’ve done re­ally well by for-profit in­vestors’ stan­dards, we can raise much more money than a non­profit or so­cial en­ter­prise.

In my opinion, Wave’s path—im­port­ing the US startup play­book to de­vel­op­ing coun­tries—was pre­dictably high-ex­pected-im­pact ex ante. First, start­ing a com­pany gen­er­ally has high ex­pected im­pact: the so­cial benefit of in­no­va­tion is usu­ally a large mul­ti­ple of the pri­vate re­turn. Serv­ing an emerg­ing mar­ket adds an­other mul­ti­plier, be­cause the prob­lems you could work on are much worse. (Pro­vid­ing some­one $5 of value means a lot more when $5 is their day’s wages!) Fi­nally, there’s more low-hang­ing fruit for com­pa­nies to work on in de­vel­op­ing coun­tries, be­cause the sup­ply of skil­led en­trepreneurs is smaller.

On the mar­gin, then, more al­tru­ists with ex­pe­rience work­ing in the de­vel­oped world should try this ap­proach. I feel safe say­ing this be­cause very few peo­ple (al­tru­is­tic or not) cur­rently seem to.

This is sur­pris­ing, since lots of de­vel­op­ing coun­tries now have the in­fras­truc­ture to sup­port tech com­pa­nies. In big cities like Dakar, Nairobi, Ad­dis or La­gos, there’s mostly-re­li­able elec­tric­ity, de­cent in­ter­net, high smart­phone pen­e­tra­tion, drive­able roads and so on. But there aren’t many great star­tups tak­ing ad­van­tage of them. (Ac­cord­ing to our in­vestors who pay the most at­ten­tion to Africa, Wave is by far the most promis­ing.)

Why isn’t the space more crowded? I’d guess it’s be­cause cre­at­ing a great product re­quires two things: be­ing ma­ni­a­cally perfec­tion­ist, and deeply un­der­stand­ing your users. To be ma­ni­a­cally perfec­tion­ist, you need to be im­mersed in a cul­ture with re­ally high product stan­dards (for in­stance, Sili­con Valley). To un­der­stand your users in Africa, you need to live in Africa. The in­ter­sec­tion of these two groups is prac­ti­cally no one, be­cause most peo­ple who could live in Sili­con Valley would much rather not move to, say, a former tank base in the mid­dle of the desert (where many of my cowork­ers lived for years).

One way to think of Wave is as an im­porter of high stan­dards. For in­stance, in most mo­bile money sys­tems in Africa, if you try to make a large with­drawal, your lo­cal agent may not have enough cash—it could take them hours or days to come up with the money. At Wave, we re­al­ized this made users sad, so we started pre­dict­ing how much cash our agents would need and work­ing with them to make sure they never ran out. This was a lot of ex­tra work and risk for us, but led to mas­sive adop­tion from traders—with funds available in­stantly from Wave, they could of­ten turn over in­ven­tory liter­ally twice as fast. Every way in which Wave beats other mo­bile money sys­tems has fol­lowed a similar play­book: no­tice a prob­lem, then try re­ally hard to solve it in­stead of say­ing “eh, what­ever.”

The “lo­cal con­text plus high stan­dards” the­ory sug­gests a sim­ple (though not easy!) strat­egy to build a high-qual­ity busi­ness that helps the global poor:

  1. Move to a de­vel­op­ing coun­try to un­der­stand your fu­ture users.

  2. Learn the startup play­book (for in­stance, by do­ing Y Com­bi­na­tor).

  3. Start a busi­ness whose users are in the place you live.

The re­main­der of this post fleshes out this strat­egy. I’ll re­strict my dis­cus­sion of lo­ca­tion mostly to sub-Sa­haran Africa, since that’s the re­gion I’m most fa­mil­iar with, but I’d guess it would work in other de­vel­op­ing coun­tries too.

The first im­por­tant de­ci­sion in this strat­egy is where to move to. Here, I’d (roughly) rank the most im­por­tant crite­ria as:

  1. Min­i­mal lan­guage bar­rier so that you can un­der­stand your users and coworkers

  2. Good enough in­fras­truc­ture to sup­port a tech startup

  3. Some­where you like liv­ing, to min­i­mize your risk of burnout

  4. Strong enough in­sti­tu­tions that your ex­pro­pri­a­tion/​cor­rup­tion risks are low

  5. A large ad­dress­able market

For an English speaker, the top few places would prob­a­bly be La­gos, Ac­cra, Nairobi, Kam­pala or Dar es Salaam (though I’m not too con­fi­dent in this be­cause I’ve never lived in any of these). If you’re con­sid­er­ing this strat­egy, I’d recom­mend re­search­ing them all, vis­it­ing mul­ti­ple and choos­ing the one you’re most ex­cited about.

You should prob­a­bly spend some time just learn­ing be­fore try­ing to start a busi­ness. While you’re do­ing that, it will help to work some­where else—ideally some­where that will ex­pose you, not in­su­late you, from what daily life is like for most peo­ple. For this, NGO or “vol­un­tourism” type work might ac­tu­ally be use­ful. For in­stance, Drew, Wave’s CEO, started think­ing about money trans­fer af­ter run­ning a non­profit sel­l­ing hand-carts in Tan­za­nia for a few years, and ex­pe­rienc­ing the pain of try­ing to send money to his own busi­ness.

The sec­ond im­por­tant de­ci­sion is what to work on. If you talk to a lot of peo­ple about their prob­lems for a few months, you’ll be more of an ex­pert than me on this—I’ve lived in Africa for a cou­ple years, but I was already work­ing for Wave at the time so didn’t look too hard for other ideas. That said, I’ll give some of my pri­ors.

The biggest busi­ness gaps in the de­vel­op­ing world are the ones where prod­ucts built in the de­vel­oped world don’t gen­er­al­ize well to new con­texts. For in­stance, credit/​debit cards barely ex­ist in de­vel­op­ing coun­tries, be­cause most cus­tomers don’t have bank ac­counts and most mer­chants can’t af­ford a card ter­mi­nal. You’ll find many other in­stances of that kind of gap.

Here are a few I’ve no­ticed. They’re skewed to­wards fi­nan­cial ser­vices, but should give you a gen­eral idea of the types of prob­lems.(2)

  • Most peo­ple don’t have com­put­ers, so lots of com­puter-based busi­ness soft­ware in the de­vel­oped world should be re­built to work well on smart­phones. In Wave’s ex­pe­rience, for in­stance, most small busi­nesses don’t keep any sort of fi­nan­cial records. An easy-to-use mo­bile cash reg­ister or book­keep­ing sys­tem could help small busi­nesses run much more effec­tively.

  • There seem to be way fewer chain stores in Africa, and es­pe­cially few that also op­er­ate in the de­vel­oped world. I’m not sure why this is, but I’d guess that the play­book for op­er­at­ing suc­cess­fully in Africa is very differ­ent and doesn’t trans­fer well. But chain stores still should have ad­van­tages in economies of scale and brand­ing.

  • Many more peo­ple in de­vel­op­ing coun­tries are in­for­mally em­ployed or self-em­ployed. That means they have a lot of prob­lems that de­vel­oped-world tech com­pa­nies have yet to solve, like in­come that’s very differ­ent from month to month, or need­ing to co­or­di­nate with other in­for­mal en­trepreneurs in the same line of busi­ness. Prod­ucts to help with these challenges would help a huge num­ber of peo­ple.

  • Things that rely on cus­tom hard­ware in the US could be re­built to use cheaper/​no hard­ware plus a smart­phone—for in­stance, an ATM or cash reg­ister could talk to a phone over Blue­tooth in­stead of us­ing its own screen.

  • Lots of sys­tems in the de­vel­oped world (ship­ping, for in­stance) rely on well-known street names and ad­dresses, which don’t ex­ist in many Afri­can cities. But they were also built be­fore GPS was ubiquitous. “Lo­gis­tics, but with GPS” is prob­a­bly sev­eral differ­ent com­pany ideas.

  • To­day’s e-com­merce is built on a lot of as­sump­tions that are false in Africa, like street names men­tioned above but also easy elec­tronic pay­ments, a func­tional postal sys­tem, rel­a­tively easy credit, well-known large-scale brands, etc. One Afri­can e-com­merce com­pany, Ju­mia, re­cently went pub­lic, but they don’t seem to be ex­e­cut­ing par­tic­u­larly well on product. I think it should be pos­si­ble to do bet­ter.

  • It’s ex­tremely hard to find good skil­led trades­peo­ple (plumbers, elec­tri­ci­ans, etc.) in many cities I’ve lived in. My guess is that the US solves this with some com­bi­na­tion of high-qual­ity vo­ca­tional school and oc­cu­pa­tional li­cens­ing, nei­ther of which ex­ists (or is en­forced) in many de­vel­op­ing coun­tries.

  • Most “API for X” com­pa­nies don’t serve de­vel­op­ing coun­tries, or treat them as an af­terthought. For in­stance, Africa’s Talk­ing is “Twilio, but for Africa,” and at Wave we’ve built our own micro-Twilio in a few differ­ent coun­tries where Africa’s Talk­ing doesn’t op­er­ate yet. Th­ese busi­nesses de­pend on a strong tech ecosys­tem, so I’m not sure how vi­able they are yet, but they also en­able that type of ecosys­tem, so are es­pe­cially benefi­cial if they do work!

Last, a few thoughts on how the “Sili­con Valley startup play­book” needs to change in a de­vel­op­ing-coun­try con­text.

If you’re not lo­cal, you will make a lot more mis­takes be­cause you don’t un­der­stand your users or your con­text. It’s way harder to build a good product when your users can’t read, or when you have to spend two weeks chas­ing a tele­com to get per­mis­sion to send text mes­sages. That means it’s even more im­por­tant to learn quickly from your mis­takes, and you should ex­pect to iter­ate for a long time be­fore find­ing product/​mar­ket fit.

Even if you’re in an An­glo­phone coun­try, you’ll need to be “bil­in­gual” be­tween lo­cal and tech-startup norms. At Wave, our in­ter­nal cul­ture em­pha­sizes hon­esty, trans­parency and au­ton­omy, which is very differ­ent from a typ­i­cal, say, Sene­galese work en­vi­ron­ment. Some of our most im­por­tant hires were peo­ple who were ex­cep­tional at “code-switch­ing” be­tween the two—they helped us work much more smoothly with lo­cal part­ners, and made a huge differ­ence to our product and strat­egy.

If you’re in a coun­try with weaker in­sti­tu­tions, the risk of cor­rup­tion or ex­pro­pri­a­tion hurt­ing your busi­ness is much higher. Com­pared to the de­vel­oped world, it’s much harder to “fly un­der the radar” of ei­ther the gov­ern­ment or your com­peti­tors, and they will also prob­a­bly be less scrupu­lous. Start work­ing on miti­gat­ing this risk (by hav­ing pow­er­ful, well-con­nected in­vestors or busi­ness part­ners) ear­lier than you think you need to.

If you’ve read this far and are in­ter­ested in start­ing (or work­ing for) a busi­ness that helps peo­ple in emerg­ing mar­kets, I’m happy to talk more! Just get in touch.

Thanks to Eve Bi­gaj, Drew Durbin, and Lin­coln Quirk for read­ing a draft of this post.

  1. The web­site only men­tions in­ter­na­tional money trans­fer, be­cause mo­bile money users don’t re­ally browse the web. You can find the mo­bile money team’s min­i­mum vi­able re­cruit­ing web­site here.

  2. Th­ese are weakly held ex­am­ples—you should val­i­date them for your­self, thor­oughly, if you’re in­ter­ested in one.