A (hypothetical) follow-up conversation with Bill Gates about Jeeon’
I had the good fortune of meeting Bill Gates 1:1 once in 2016, when he was already the leading philanthropist in the world. Thanks to our angel investor and friend Matt Dalio’s kind introduction and accompaniment, we had almost an hour-and-a-half of uninterrupted discussion where he asked us many questions and shared insights and advice.
Last night, out of the blue, I dreamed about a chance social encounter with Mr. Gates, where he was his usual curious and sharp self, and asked me some hard questions about Jeeon’s journey. Even while half asleep, I thought it gave me a great opportunity to reflect critically on the eventful 7 years past. Although I don’t remember anything from the dream conversation itself, here’s an attempt to recreate it. Hopefully it helps me reflect in reality as much as it helped me in my dream.
Me: Hi Bill! Do you remember me? I visited your bgc3 office once in 2017, with Matt Dalio. I run a social enterprise called Jeeon in Bangladesh that works with pharmacies. It’s so good to see you here!
Bill: Oh hey! I do remember you! I also vaguely remember that you were attempting something rather radical back then. Remind me what you do again with pharmacies?
Me: Thanks! Yeah, back then we were trying to develop a telemedicine service linking rural pharmacies with urban doctors, so that poor people could access high quality care nearby. But the big picture goal was to unlock pharmacies as a high-bandwidth and low-CapEx delivery channel for quality health products, services and technologies, given so many people across the world still rely on pharmacies as their first care touchpoint.
Bill: Oh yeah! Now I remember. I thought it was quite clever to try to utilize an existing health worker base and physical infrastructure, rather than build new ones as most others had done. How has it been going? I don’t think I’ve heard much about you guys since.
Me: Can’t say I’m offended! :) We went through a few pivots and tested different models, none of which ultimately succeeded in getting us to scale. So it’s been quite an uphill battle. Recently we were forced to convert to a non-profit, which is honestly kind of an admission of defeat, because we had originally set out to build a highly profitable model for delivering quality healthcare to rural areas.
Bill: I’m sorry to hear that. But it was an ambitious attempt, and I’m sure you’ve learned a lot. Tell me more about why the models didn’t work and what you learned from it.
Me: Of course. Umm…I guess I’ll start by going back to something you had said in our first meeting, which I did not appreciate fully then, but have since come to internalize through our failures. You said something along the lines of: “If you can make your services attractive and sticky for the pharmacies, that’s your golden ticket.” I think when you said that, you were implicitly telling us to treat the pharmacies as our primary customers.
I guess the biggest learning has been that when working with private sector pharmacies, we should have aligned our proposition with their business and social incentives first and foremost. For example, with telemedicine, we were trying our best to optimize value for patients, but it conflicted with the pharmacy’s incentives as the de facto providers in those areas. So in effect we were competing with our own channel.
Bill: That’s really interesting. I guess the complicated incentive structure of private providers is one reason why most government and non-profit actors have steered clear of them. How did the incentives piece play out in subsequent models you tried?
Me: You’re right, and there’s also the regulatory issue of these providers not having any recognition in national legal frameworks. Hiring your own health workforce, as most governments and NGOs have done, also makes them easier to control and manage, whereas here you’re working with a herd of wildcards. We were certainly ambitious and wanted to figure out this missing piece in global health, but we had no idea that we were biting off much more than we had the ability to chew.
Coming back to the incentives piece… they did play out in subsequent models as well. In 2018 we pivoted to an e-learning platform where we wanted to build their knowledge base, and monetize it by running paid promotions and selling products . Ultimately, we found that learning itself was not a sufficiently strong incentive to hold engagement in the long run — and by long run I mean beyond two months! I wish we had something like an accreditation lined up as a goal-post for learning, which would have made it a lot stickier I think.
We also tried an e-commerce model selling pharma products to mid- and low-tier pharmacies, but the pain point for procuring products was not sufficiently high in Bangladesh — given the highly competitive and aggressive local pharma industry employing thousands of medical reps to constantly knock on their doors for product sales — for us to be able to offer something they couldn’t live without.
Bill: Wow! You’ve certainly tried a lot of things. Do you think pharmacies only cared about financial incentives as micro-businesses, or more than that?
Me: That’s a great question. We realized that most pharmacies, at least in Bangladesh, have a dual identity — one as a drug salesman, and one as a primary care or family physician (sans the credentials). They consider this a respectable profession, which is why many of them choose to enter the business on the first place, compared to running a convenience store for example. Many of them also come from a family of informal “village doctors”, and learn the trade from their father or uncle.
So they not only care about their revenue, but also their reputation and social standing. As a result, anything that differentiates them from competition, or associates them with a prestigious entity or brand, or gives them a credential of knowledge or excellence, are also quite valuable.
Finally, there is the issue of their legal standing, or lack thereof. They often fall victim to being rounded up, or shut down, for practicing and prescribing without credentials. So any forms of accreditation or brand affiliations that gives them some legal shield is also a highly coveted incentive.
Bill: Great insights! How did you try building these into your subsequent models?
Me: To be honest, I don’t think we did a good enough job on the incentives front, despite the learning. None of our models had a strong financial value prop, for example one that would increase their revenue or margins by 20% or more. We did provide some trainings and accreditations for small side-projects here and there, such as a partnership with VisionSpring where we taught them to screen for eye problems and gave them reading glasses to sell, but we could not secure a government-endorsed accreditation for any of the core models. There were some reputational benefits from being associated with Jeeon and its doctors, but that was a weak incentive and clearly insufficient.
Now that you ask, it’s unclear to me why we failed to do something so basic.
In contrast, when COVID struck, we went all out to try to understand their pain points, and realized that they were operating in fear of their own lives while being forced to treat hundreds of symptomatic patients. So we launched an online course on COVID response and management targeted at pharmacies, and got it endorsed by the government, and in under 3 months we were able to train and certify 25,000 pharmacies. Clearly we had nowhere near that level of traction with any of our core business propositions.
Bill: It’s at least good that you realize this now. Better late than never! Also, you’re the first one I have heard from who has trained pharmacies to respond to COVID. That’s really fascinating. What else did you do during COVID? And what role did the pharmacies play during the pandemic?
Me: Yeah, we had a pretty unique vantage point when the pandemic started. Doctors were not in their chambers, and hospitals were turning patients away, so people were even more reliant on pharmacies than usual. And pharmacies were operating without any knowledge of the virus or protective equipment. No one else was seeing the problem like we were.
In addition to the online training, we secured PPEs for about 3000 pharmacies, set up a nationwide symptomatic surveillance through pharmacies with Yale University researchers, disseminated vetted information to pharmacies and fought misinformation through a Facebook channel, and helped develop a community engagement model leveraging pharmacies in one sub-district before turning it into an SOP for nationwide capacity building.
Within that broader community engagement strategy, pharmacies were expected to maintain a risk register of elderly and co-morbid patients, promote preventive behaviors such as mask use and distancing, screen symptomatic patients and advise them on home quarantine and isolation, monitor escalation of symptoms and bring anyone with breathing problems to the hospital urgently, and more. It was quite heartwarming to see how dedicatedly they volunteered in service of their communities, with no financial incentive and taking a lot of personal risks.
Bill: That’s pretty neat. Congratulations! Going back to your core business models, I noticed that you tried to apply technology for each of the models you attempted. Was that strategic, or an unconscious bias as a result of being a techie yourself?
Me: It was a constraint we had imposed on ourselves from the beginning — and our background as techies may have biased that decision. The rationale was that pharmacies are a highly decentralized and fragmented infrastructure, and the only way to bring them under a network that is low-cost and scalable would be to leverage the rapidly rising penetration of smartphones. In other words, if we could deliver something of value through a phone, and they used it regularly, we could use it to deliver many other products, technologies (such as diagnostics), and health services (like telemedicine and referrals) to people through them, at negligible marginal cost and therefore significant margins. Something like a super-app for upgrading pharmacies.
However, in retrospect, I feel that maybe we tried to put the cart before the horse. If we had figured out the incentives piece first, and created something of tremendous value to a pharmacy’s business that they couldn’t live without, we could later have figured out how to make it more scalable using technology. Indeed, that’s how some of the more successful models emerging recently, such as mPharma in Ghana, have approached it.
Bill: It certainly looks a lot clearer in hindsight, doesn’t it! And people don’t always realize that the Silicon Valley model of rapid growth and scale may work in areas with high density of capital and potential acquirers, but not so much in emerging markets. Would you agree?
Me: Absolutely. We were certainly seduced by the Silicon Valley blitzscaling mindset, and were optimizing towards that in subsequent pivots. But if your goal is not the fastest and biggest exit, and rather deep and lasting impact, starting slow and building strong foundations of value creation at a small scale may be the more prudent approach. Kind of like BRAC’s Fazle Hasan Abed, who used to pick an important problem, start small experiments and tweak rapidly, but design for scale from the very beginning.
Bill: Yeah, Abed was a rare breed. I miss him. Looking at your pivots, I also can’t help but notice that you were moving further and further away from your goals of impacting health outcomes. What was your rationale behind that, or was it simply due to external pressures to monetize, or like you said, the need for rapid scale?
Me: That’s a great question. I have wondered that many times myself. I think when our telemedicine model, which was a direct path to impacting health outcomes, did not pan out for lack of strong incentive alignment, scalability, and a solid revenue model, we realized that we needed something that would be stickier with pharmacies and easier to scale and monetize. With e-learning, there was enough early promise of traction and monetizability, until we found out it was not sticky long-term. By then, we had already spent 4+ years with the business and really needed something that would start to generate revenues in the short term. E-commerce was the best plan we could come up with, as drugs are the core business of pharmacies with over 90% of the transaction volume, and since we would make a margin from every unit sold on the platform.
We convinced ourselves that we would go back to a health impact focus as soon as the business side was stable and generating cash. But it’s clear that even without any undue pressure from our investors, there was mission drift simply as a consequence of our own legal and incentive structure, and maybe additionally the Silicon Valley mindset.
Bill: That’s hardly uncommon. While on the topic of legal structures, I am curious why you want to continue Jeeon as a non-profit if you feel that your mission has already failed?
Me: Again, great question. It’s true that we feel our original mission of developing a profitable model for rural healthcare delivery has failed. However, the problems with rural healthcare remains pretty much as it was when we started. For one, people still rely on pharmacies and informal providers as the gatekeepers of the health system. Yet, the formal healthcare system and regulatory environment does not even acknowledge, let alone utilize, their presence and popularity. So there’s still plenty to be done in the realm of designing and developing models of formal-informal collaboration, research and evidence generation on health outcomes, and policy advocacy. All of these are better suited to a non-profit model.
Bill: Okay, makes sense. What use cases are you thinking of where pharmacies could play a significant role?
Me: Of course infectious disease surveillance and management, a la COVID, could be a big one. But I think an even bigger opportunity lies in non-communicable disease management, especially diabetes and hypertension. They are fairly simple to diagnose and monitor, and the advice and treatment are fairly algorithmic. In fact, we are already collaborating with the Bangladesh Diabetic association to test a model where pharmacies can use an app with embedded clinical decision support algorithms to triage, monitor, treat and refer diabetic and hypertensive cases. And finally, I think there’s a lot to be done to prevent them from over-using antibiotics and contributing to the growing AMR crisis.
Bill: Listen, I have to run shortly, but before I leave you, I wanted to understand what key lessons you think the global health space, and organizations like my foundation, can take from your experience? Does it vindicate their current position of staying away from private providers and trying to build suitable alternatives, or no?
Me: Wow! That’s quite a big question. Let’s see. I don’t think ignoring pharmacies and informal sector is the answer, because billions of dollars poured into alternative infrastructure over the past 40 years has been able to make almost no dent in their popularity. If only a small fraction of investments could go into testing different models for leveraging them towards global health goals, I think it could go a long way. The ROI on any successful models developed would be tremendous, and much higher than existing models, given there are so many of them already serving so many people each day.
Thanks so much for your time and questions, Bill. You might only be in my imagination this time around, but it was still extremely helpful.