Overcoming the global identification challenge

Podcast transcription - 8th may

Alan Cowley:                    Welcome to The Invested Investor. I'm sat here with Toby Norman. Toby is the CEO of Simprints. Toby, let's just hear a little bit about your background and Simprints, please.

Toby Norman:                 Wonderful Alan, it's great to be here. Simprints is a non-profit technology company that came out of the University of Cambridge about three years ago. The challenge that Simprints is trying to solve is what's known as the identification bottleneck, the fact that today 1.1 billion people, that's one in seven people on the planet, have no formal identity whatsoever. They're functionally invisible in the eyes of the world, mostly in poor and developing countries.

                                           What this means is that it's incredibly difficult for organisations that are trying to fight poverty and global disease, so picture teams such as UNICEF, Mercy Corps, Doctors Without Borders, and others, to link patients or beneficiaries consistently to records, showing for example that they've received healthcare services or access to education. If you could imagine yourself as a health worker in a place like Bangladesh, and there's a child in front of you, and that child has no birth certificate, no vaccine record, just a name that potentially overlaps with hundreds of other kids in that particular slum, it's really hard to know, have we actually reached the correct child with health services? Which services have they received, which services haven't they received? That's fundamentally the challenge Simprints is trying to solve.

Alan Cowley:                    So how did you get to create and found Simprints? What was your background before that?

Toby Norman:                 My background's in global health. I've been in this space for about 10 years, first Harvard, then more recently at the University of Cambridge. I was actually doing my PhD research on frontline non-profit workers in Bangladesh. One of the contexts we were working with was frontline community health workers, who were going house to house in slums in Dhaka, trying to deliver basic maternal healthcare services. What we found, was that it was really, difficult for these health workers to be able to link patients consistently to digital records.

                                           So for example, in the maternal healthcare space, you're trying to see each mother about four times before they deliver their child. You're looking for anaemia, preeclampsia, the things that can potentially kill mums around the period of childbirth. Seeing this challenge up front was really where I first started thinking that identity is actually a hard and difficult challenge that the global health community is facing that we should try and solve.

Alan Cowley:                    Okay. So you came up with the idea at Cambridge University, then? Did you have a co-founder at that time?

Toby Norman:                 Yeah. I actually first had a couple of different rounds of co-founders. The final co-founding class I'd say was four of us, but initially, when we started working on this project, I never thought we'd build a company out of this problem. I attended a hackathon in 2012 that was talking about, how do we improve the way we deliver digital healthcare in developing countries, and that was where the first idea of using biometrics. So, for example, something like a fingerprint is that unique link between a person and their digital record, came about. We just started working on this really as a nights and weekends project. I was still doing my PhD at the time. I had accepted a job with Mackenzie, so I was thinking about what I'd do next steps, in terms of my career. Had no ambition whatsoever to be an entrepreneur, to actually ever build out a company.

                                           But as we got further and further into this problem, it really just became too interesting to let go and too important to let go, and so I think it really was at the end of my PhD, where I had to sit down and figure out what I wanted to do next, and it just became incredibly clear to me that the right thing to do was to turn down the job at Mackenzie and actually throw my weight behind trying, at least, to build a tech company to solve this challenge.

Alan Cowley:                    Was there four of you at that point?

Toby Norman:                 There was four of us at that point, three friends from the University of Cambridge and my brother, who had just finished his master's in computer science at the University of London. That was the founding team at Simprints.

Alan Cowley:                    Brilliant. Are all four of them still there today?

Toby Norman:                 All still here, which is surprising and a little bit unusual in the start-up world, to have four co-founders and have four co-founders still in the journey at this point. So I think I've been incredibly fortunate with my co-founding team.

Alan Cowley:                    How big is the complete team now?

Toby Norman:                 At the end of this recruiting cycle, we're going to be at 25. In three years, it's been quite a lot of growth for a non-profit tech company, which is actually part of the reason we're moving here to these new facilities in Bradfield.

Alan Cowley:                    Okay, brilliant. Tell us a little bit more about the product then.

Toby Norman:                 What Simprints does is it links people to their digital health or digital service records through biometrics like their fingerprints. Biometrics is a relatively mature market. It's been around, at least digital biometrics, since the 1970s. We looked at a bunch of different biometric modalities, so these are things like fingerprints, facial recognition, iris, there's a whole different number. The challenge we found, particularly in our work in Bangladesh and later in Sub-Saharan Africa, was that of the different modalities, fingerprints in theory should give you the best accuracy per cost, but the highest levels of cultural acceptability.

                                           For example, in many of the places we work, things like facial recognition are just culturally not appropriate. In Bangladesh, where you have Muslim women, sometimes who are veiled and are definitely uncomfortable with community health workers showing up and taking photos. But if you want to use fingerprints, one of the challenges we ran into immediately was the fact that the populations we work with are very different from the ones we have in the West. For example, if you're a farmer in a place like Kenya, and you've spent the past 20 years in agriculture, your fingerprints are actually worn down and scarred from years of manual labour.

                                           Similarly, for example, if you're a day labourer hauling bricks in Dhaka City, or you're just a housewife in a community in Nepal with scarred and damaged fingertips, because you've been cooking on open fires most of your life, this meant that when we actually started testing biometric technology, we found that the fingerprint accuracy was much, much lower than what we would see typically in the West. And so what we did while we were still at Cambridge was actually hand collect over 135,000 fingerprints from four different countries and optimise the hardware and software we were working with to be 228% more accurate than the next best technology on the market, with scarred and damaged fingerprints. That was really the technical challenge we had to solve as we began to develop Simprints, and led to a lot of the innovations that we continue to work on today.

Alan Cowley:                    You said hand collected 135,000? Do you mean that literally?

Toby Norman:                 I mean that literally. I mean hundreds of hours sitting in vaccination clinics, sitting in schools, sitting in hospitals. Literally myself and my co-founding team, taking people's fingers and manually placing them on different types of biometric sensors to figure out which type of hardware and software is going to give us the best accuracy. It still is to date the world's largest academic dataset of developing world fingerprints. It was an incredible amount of effort, but we had almost no resources when we started. This is really what it took to understand, how do you make fingerprint technology work at the last mile context like this?

Alan Cowley:                    Okay. What does the product actually look like now? You've recently been in Bangladesh. I know you've got a large presence there. What are people using?

Toby Norman:                 If you imagine the product is a combination of things. Actually, it's a biometric system, which includes software, hardware, technical configuration, and then support and analytics. So for example, what that means with our partners with BRAC in Bangladesh, which is actually, it's the world's largest NGO, is that BRAC community health workers are currently using smartphones to go house to house and collect patient data and transfer that data up to a central server.

                                           Simprints essentially plugs into that already existing digital healthcare system, and so the health worker is able to use a hardware fingerprint scanner designed by Simprints, so this is rugged, waterproof, shockproof, with a long-lasting battery life, that collects the patient's fingerprint, transmits it wirelessly to the smartphone, where the health worker is able to pull up, either online or offline, the patient's healthcare record, and then continue working with that healthcare record like they normally would throughout that visit.

                                           Simprints supports the full integration of both the software technology into this ecosystem, but also making sure health workers have the right training, the right technical support, and that we're supporting managers with the correct dashboards and analytics, to make sure the system's actually being used and the data is driving change. Because at the end of the day, you can have all the technology in the world, but if managers and health workers aren't using this biometric data to make better decisions in the field, then you haven't achieved anything.

Alan Cowley:                    You talked about Kenyan farmers and how difficult it was to recognise their fingerprints. What percentage do you have now? With the current hardware, what percentage of fingerprints are you able to recognise?

Toby Norman:                 It depends context to context, but when we first started evaluating the technology that was already existing on the market, we saw something known as equal error rates, so that's the balance between a false accept and a false reject rate. Just like normal science, you've got type 1 and type 2 errors, over 30% with a lot of technologies, which is completely unusable. Now in Simprints projects, our accuracy rates range anywhere from 100% to 98%, based on the project and the context you're working in. There's a number of things we can do to consistently bring that up to nearly 100% every time, based on the number of fingerprints that you're willing to collect in field and the accuracy requirements of the project.

                                           But basically, at the end of the day, this is a fingerprint technology that genuinely works at the last mile in these contexts and has been proven now, I think we're actually in 10 different countries at the moment, with deployments with over 11 different clients. Our largest deployment with BRAC in Bangladesh will cover 2.1 million mothers and children over the next three years. It's our first national-scale project, supported by the Gates Foundation, DFID, and many other great partners, including Arm, and we're incredibly excited to take this work forward.

Alan Cowley:                    You talked about being a non-profit earlier. Can you explain your non-profit status and why you went down this route?

Toby Norman:                 Yeah. We thought really hard in the early days about how we wanted to structure the organisation. And if anyone's interested in the topic of how do you legally structure a social enterprise, we actually open sourced our research on this topic in a white paper that we wrote with the local law firm Taylor Vinters. So if you go down to our website, to the resources page, you'll find a white paper called The Legal Landscape that looks at the different options for social enterprises.

Alan Cowley:                    We'll put that in the show notes for people.

Toby Norman:                 At the end of the day, I think there were a couple of key things that motivated our decision. First and foremost, we knew that we wanted to plant a flag in the ground, in terms of what type of organisation we wanted to be, and that was really being an impact organisation, an organisation that put impact first. And I truly believe that you can do this either as a for-profit or as a non-profit, but being a non-profit did give us a lot of advantages in the early days, in terms of signalling to the market and signalling to our partners that this is something we were serious about.

                                           However, we were quite determined that we did not want to be a charity, and actually, Simprints is a non-profit, but it's not a charitable organisation. It does not accept charitable dollars. And actually, it's working incredibly hard to make sure that the vast majority of its income comes from sales revenues. Clients such as Watsi, Mercy Corps, Doctors Without Borders, or other partners actually pay us for the technology and the services we provide, so that we're fundamentally accountable to this tech working in the field and making a difference. We've been on a transition since we went to market about a year and a half ago, to become a revenues-driven social enterprise organisation, a non-profit social enterprise. And we've made incredible progress on this front.

                                           About 18 months ago, I think we were 3.6% revenue-funded, and the rest was funded by grants and foundations. Today, that number is 36% funded by sales revenues, and the target by the end of this year is to get to 50% if we can. We're working hard to make this transition.

Alan Cowley:                    Brilliant. Obviously, you started off with grant payments. How easy was it to get those grants, and did you do it purely in the UK or globally?

Toby Norman:                 Yes, good question. It's incredibly difficult to raise grant capital in this space, and so I would encourage anyone who's thinking about going this route to choose the legal structure and the status that's going to work for you for the right reasons. Don't assume that it's going to be easier to raise money through grant capital than potentially taking on for-profit or equity investment. I would say we saw a couple distinct stages in our journey. The first stage is, you are really a storyteller, and you're pitching for, frankly, very small amounts of money. When we started, we actually pitched at student business plan competitions in Cambridge University.

                                           I think our first £5,000 came from the CUE, Cambridge University Entrepreneurs, business plan competition. And really, that was all the money we had to do a lot of our initial field testing, prototyping, a lot of that initial work. We started with very small sums of money for a very long time. About two years, Simprints was a student volunteer group, which was working on this project. What that phase allowed us to do, though, was start to actually generate data, showing that there was potential viable solutions here. And we were finally able to take enough of that data to larger grant-based foundations, specifically something called the Saving Lives at Birth Award, which is a combination of Gates Foundation, DFID, USAID, and other funding sources, and that was our first 250k of proper funding to actually start this full time.

                                           From then, we've been really fortunate to partner with a number of other amazing organisations, including support from Arm, organisations such as the Global Innovations Fund, Draper Richards Kaplan, Autodesk, and others. And really compiling those sources together, got much stronger at building grant funding applications, actually deploying our first projects. The final stage that I hope we're in now is actually the transition from that to really becoming a revenue-driven organisation, while still partnering with selected foundations or potentially philanthropists on exciting R&D initiatives.

Alan Cowley:                    You got any key tips for entrepreneurs that are looking to raise grant funding?

Toby Norman:                 This is a challenging area, and I think a couple lessons that we've learned over the past couple years, first, you need to be very clear what is the story that you're trying to tell? And often a combination we've found that works particularly well is making sure you have a qualitative human story that people understand and can potentially even relate to, or at least be able imagine themselves in that situation. But you need to support that story with a quantitative empirical story that's driven by data. And I think this is really important.

                                           Potentially, too often in some of the work I see in the charity sector, it's purely the human story and only the human story, which, while meaningful and important, I think if you're not willing to back that up with actual market data showing the scope of the problem, the potential of the technical solution, and evidence that what you're proposing can genuinely work, I think you're missing opportunities to get onboard supporters and funders who are going to be driven more by that kind of evidence.

Alan Cowley:                    The way Simprint's going, what does the future hold?

Toby Norman:                 Our focus at the, moment is growth and proving out the impact of what we do. What growth means in concrete terms is actually transitioning from the pilot phase to our first scale projects. As I mentioned, we're now in 10 different deployments all across Sub-Saharan Africa and South Asia, and we're about to start our first national scale deployment in Bangladesh with BRAC. This is going to be a massive challenge that we've never taken on a project of this size or scope before. I know we've got a lot of learnings coming down the pipeline, but this is also incredibly exciting, because if we can make this work, we can prove that you can deploy biometrics accurately, safely, robustly, and cost effectively at national scale, to improve the way that we fight disease and fight poverty at the front lines.

                                           In order to support that work, we're also going to need to very clearly build an evidence base that shows that this is physically changing people's lives. A big part of our commitment is actually research. It's doing things, for example, like randomised control trials, impact evaluations, and other hard quantitative evaluations of what's the actual difference using something like a fingerprint, for example, to improve maternal healthcare coverage or decrease fraud in cash transfers or try and fight teacher absenteeism in Sub-Saharan Africa? Can we show that those things have a concrete impact in people's lives?

                                           Because even if we achieve massive scale, if what Simprints is bringing to the table isn't adding value to individual humans, there's no point of the work we do. And that's why I think you really need to have both, what we call at Simprints is the impact equation, that combination of breadth or scale with depth, depth of impact and actual difference to a person. And if you can achieve both of those things, then I think you're really onto something.

Alan Cowley:                    That was hugely inspirational for our listeners to hear. You've built the business up to around 25 people that you said, and you're moving into a nice new space here at The Bradfield Centre. What are the lessons you've learned, and have you got any tips for entrepreneurs looking to start a company?

Toby Norman:                 I learned a couple lessons. I think we still have quite a lot to learn in this space, but a few lessons I think we've learned the hard way. First one, it really at the end of the day is all about team. I have been told this so many times, but it seems to be a lesson that I manage to relearn at least once a quarter, in one way, shape, or form. To make that more actionable, though, yes, everyone gets that it's important to have high quality, motivated, ambitious people as part of your team. How do you actually get that? How do you put that together?

                                           To give one concrete example, on the engineering front, one of the challenges we have as a non-profit technology company is that we're in Cambridge, competing with top technical talent from some of the biggest players in this space, players like Apple, players like Amazon. How do you ever compete on things like engineering salaries against those players? And so I think we've had to think very strategically, how do we actually get the best talent when those are the opportunities they're also looking at? I think what we've really learned is that you've got to think very strategically about what are your areas of relative strength that you can apply to their areas of relative weakness, in terms of the value proposition they're proposing to those potential employees?

                                           So for example, if we're going toe to toe with Amazon for engineering talent, there's no way we're going to be able to compete on salary, so don't spend your time or waste your money coming in 20, 30, 50% below what Amazon could offer, because it's not going to actually seal the deal. Where you can compete, though, is you can absolutely compete in areas like purpose. Is there something meaningful to the work you're doing? Autonomy, in terms of control over people's lives, giving them the freedom, the flexibility, and the responsibility to have ownership. For example, one thing Simprints does is we do unlimited paid time off. We screen very hard for people who are so passionate about making the world a better place and doing great engineering that they're going to choose when they need to take a rest, refresh, and come back to work.

                                           We think really hard about culture and team, and these areas, that actually it's tough for Amazon to compete on. No matter how exciting and huge the organisation is, at the end of the day, it's not going to be making the type of difference to, for example, really poor children's lives in Bangladesh, that an organisation like Simprints can make. And so if we compete on those areas strategically, we've got a much stronger chance of landing the type of talent we need to solve the technical challenges we face. And those technical challenges are massive. Very exciting, very cutting edge, but very difficult.

                                           The second thing I'd focus on is fast feedback cycles, particularly if you're thinking about working in the non-profit area. Something that we've seen very consistently in our work and that we find incredibly important is having really clear feedback loops for the different stakeholders in your ecosystem. You could imagine a normal for-profit context, essentially your customer and your payer are the same person. If I am selling an application on the pay store that does some function, and I'm selling that directly to a consumer, if the application is good, the customer will buy it, will pay for it. That's a pretty tight feedback loop. If no one's buying my product, I know I've got to fix something.

                                           One of the challenges we have in the non-profit sector is sometimes the user and the payer aren't the same person. So for example, our work, our project could be funded by the Gates Foundation, but the actual implementation is being done by Mercy Corps, and it might be a manager who selects Simprints to use as their biometric frontline tool, but it will be a health worker, potentially in a village thousands of miles from there, who's actually at the end of the day the one using Simprints. And a beneficiary who has nothing to do with any of that, except their interaction with the health worker, who's actually receiving the service.

                                           So in order for us to stay effective, we've really got to think, what are our feedback loops with every single one of those stakeholders? Obviously, with the payer, it's a financial feedback loop, and that's often the strong one that a lot of people in this space focus on, but I think there's a real risk if you just focus on that financial feedback loop, you are actually going to miss the hugely important user feedback loop to the health worker and the impact feedback loop to the beneficiary. And if you don't have those things, you are not going to be learning, and you're not going to be continuously improving the work you do. And I think that's fundamental, because this area is too complicated for anyone to be able to predict how everything's going to go first time round. You have to have those feedback loops, so you can learn, so you can improve. That's what it takes to create impact at scale.

Alan Cowley:                    Although a lot of companies won't be based in Bangladesh and then have the manager somewhere else, that is really, useful to a lot of companies. Well, Toby, it's been absolutely brilliant having you on The Invested Investor Podcast. Hugely insightful, particularly about non-profit fundraising, and I know you've already won a number of awards, and I know Simprints will continue to do brilliant work. Thank you.

Toby Norman:                 Fantastic. Real pleasure being here today, Alan, and you can check us out if you're interested in our work, www.simprints.com, or follow us on Twitter. Thanks a lot.

Alan Cowley:                    Brilliant. Cheers.

Peter Cowley:                  Thanks for listening to another Invested Investor Podcast. You can subscribe to all future podcasts via our website, investedinvestor.com, or via a number of podcast platforms online. Remember, you can order our book online, and be sure to follow us on Twitter, LinkedIn, and Facebook to get the most up-to-date, interesting, and insightful content from The Invested Investor.