GOTO 2019 • Frugal Innovation • Betty Enyonam Kumahor



[Applause] hello everyone okay it's a pleasure to be here I actually started my career in Oak Park interestingly enough for client that was in Milwaukee and I might be dating myself a little bit but we were actually building middleware and a health insurance company so it's a pleasure to be back here but what I'm gonna talk about is very different from that experience but I'll show you a little bit more about myself I run a company that's based in Ghana and West Africa we have anywhere from 15 to 30 people depending on the projects we have and the size of them it is primarily a software development company but we do consider ourselves an advisory firm primarily because we found that building software has a lot less to do at least in our vironment has a lot less to do with the actual technology then it has to do with the problem that we're trying to solve so let me just tell you a little bit more about the environment I moved back to Ghana in 2009 I missed home that was pretty much the long and short of it and by that time I had a role in a global product manager project manager design role for an accounting firm and had been spending the last sort of seven eight years before I left the building technology that was used in 160 countries around the world and so I was used to enterprise software development we would build custom from scratch and I thought at the top left here that that would be a similar environment to what I experienced in here in the States and in London it wasn't at all like that the PI private sector in Ghana is a lot smaller they use a lot of cots packages which don't really fit the environment and so I found that instead of doing software development people were doing software implementation and it just wasn't a great fit for where I wanted to go on what I wanted to do at the same time I started looking up statistics because I could see that the number of users we had were thousand two thousand and a thousand two thousand and that's not really kind of what I was interested in because I realized that 70% of our population lives in the rural areas and that rural areas like the one on the bottom right and sometimes in sort of crowded areas in the city where most of the people in the city live like the bottom left and then I just put a picture on the top right so you could see how gorgeous it is so I realized that I really wanted to be solving the problems that are experienced by the majority of not just gun Yun's but Africans and that really caused a change in my career so I left the big four and went ahead and joined a company that was just starting out I'm not sure that this is and so what I do today most people often consider as frugal innovation I tend to speak on this topic a fair bit and we've realized that frugal innovation is essentially building software cheaply but ending up with a product that is also relatively less expensive right so it's those two things there's a process by which we do it is actually doesn't cost as much and the product that you end up with doesn't cost as much and the reason is that the end users can't really afford expensive things and what we found about frugal innovation is that the definition is really sort of understanding the domain understanding the context in which the software is going to be used but also making sure that you account for all the constraints and so I'm going to do that into that a little bit more so I tend to think of of frugal innovation as constraint driven development and the thing about constraints is that we have a ton of them on the continent and this is just a few so you've probably heard of some of these things like power and electricity is one so in some cases you're not really even getting electricity the lights will go off for example in Accra the city in which I live most people who can afford an apartment or house have some kind of generator if you can't afford to generate so you probably have an inverter and power is intermittent and I'll tell you show you some examples of how we've actually worked around that connectivity is a huge challenge as well and I mean the reality is that we didn't really have many copper wires we didn't have many landlines and so when mobile came it's something like ninety nine point ninety eight percent of the continent had no connectivity until we had mobile and so that's a huge consideration as well even today even though we've got a lot of people who actually have mobile phones it's still something that we have to think about literacy is also consideration so for example in Ghana literacy is relatively high at about seventy something percent but even if that's the case it doesn't necessarily mean that when they see you instructions on a screen that they will understand everything that you're writing and so we have to think about ways to actually communicate with our users that doesn't necessarily involve a lot of text it could be more image picture based it could actually be voice based and so those are all considerations that we have to put in the address system I always say that when you're giving addresses so now part of the world it's always like okay you go right when you see the woman selling mangoes don't take that left take the next left and when you see the tree ask there for the place it's like how does that work if you're trying to deliver I mean Amazon how's that gonna work so a lot of plots have numbers so they don't have house numbers not necessarily street numbers and that could change in five years when the per area does get developed so building for that is interesting as well peripheral devices is another big consideration even though there's more and more smartphones on the markets we still have a large segment of the population who can't really afford anything more than a feature phone so that's basically numbers no whatsapp no no smart screen any of that so we still have to build for that so SMS ussd still big in our part of the world cash is another big consideration we've been trying to push mobile payments for a while we're still not getting there I think everyone here has probably heard of an Pesa right m-pesa m-pesa comes out of Kenya passer was owned by Safaricom which is the large telecom in Kenya and I don't know the current statistic but certainly a few years ago something like 20% of the GDP at least was going through this mobile payments platform and met and it's been around for a while at least ten eleven years something like that and the reality is that we have 53 to 64 countries on the continent depending on who's counting but all of those countries have tried to replicate what empresa did and no one has really gotten there and so we're still trying to figure out how to do mobile payments because cash is just difficult to handle particularly when you're trying to do technology systems so we need to get something that's electronic and so those are a lot of the constraints of just the environments that we end up having to build for almost regardless of what solution what industry you're building for and another consideration I get this question when I'm talking about frugal innovation which is well as the tech you build really any different and the tech that we build tends to be influenced by the constraints that are up on the screen right which are skills availability availability of the engineers and basically the whole team it's still quite a challenge we definitely have had some players over the last I'd say five to eight years that are changing that but that takes time and it takes a while and we're also faced with the fact that for a number of people and makes more sense to leave the continent and continue their careers elsewhere so skills availability is is high it's low and then you also find that that the tech stack that people are familiar with is older right so we're still on PHP for a lot of the applications around PHP version 5 maybe so we're not using some of the newer technologies for everything and being able to get the resources to understand new technologies is is a challenge and so it takes a while actually to bring the bring the software up to where it might actually work in other parts of the world our life cycle is pretty much the same except you'll find that most people don't do things like pair programming again that sort of related to skills availability so you'll find a lot of people sort of coding by themselves testing for instance we don't do a lot of TDD necessarily we still have a lot of manual testing and it makes sense if you think about we're still a highly labor-intensive economy so a lot of times rather than find somebody who understands how to sort of test software electronically it's easier to hire someone and have them test manually so I think we tend to describe our software industry as sort of being behind the various continents so being behind the Americas maybe five years being behind Europe maybe four years things like that and so that just gives you some idea of the context in which we're in so that's constraints and the next thing I wanted to talk about was the context the domain and I'll do that by telling you a story so I have an uncle who left Ghana when he was a teenager came to the u.s. New York and studied ended up going to think was NYU became a doctor a medical doctor and decided after his first few years of residency that he wanted to go back and give back to his home and quite a number of people got excited he actually got a number of other doctors and nurses who all wanted to go along and the interest really caught on to the point where they actually got funding to be able to take everybody and so instead of him just thinking about doing something in his village he decided to do a bunch of vigil jizz all around where he grew up just fantastic they ended up settling on doing something around family planning teaching women how to take the pill because in Africa we still have the situation where you might have families that are too large to feed to clothe to send to school and so if you can get in early enough so that people don't have such large families then you can actually have people who were well educated well fed and brought up and become very productive members of society so they thought this was a great thing they learned all about monitoring and evaluation we're going to go measure this we're on point right so they go and they actually have figured out that they're going to take the women out for a two-day workshop and they've asked permission from their husbands and they're going to run them through everything that they need with practical exercises everything was just perfect and then they leave and come back a year later what should happen right then the incidence of pregnancy should go down incidental childbirth should go down that's not what happened it didn't stay the same it went up in both cases and so that they have to start figuring out what the heck happened so they start going around and asking everybody like so women did you come to the workshop the full two days yes did you understand everything we said yes well your husband's mad at you and so they didn't support you when you go home no everything was good so what happened when you went home and so when they started asking that they will say that their husbands told them to sit down and tell them everything that happened which they did and they said okay so how was your husband oh he was fine everything was good he understood he liked it but but like what's the but he said I am the man of this house if anybody shall take the pill is shoving me context context and so the thing that you realize is that even though my uncle was from this region spoke the local language he missed that one little thing like how are you going to actually figure out that one little thing until you're actually out there trying out the solution and seeing what the results there and we find that that's very much the same thing when we're building software if you miss one little thing the whole thing goes kaput because it's really more about the context than it is about the technology we found so the domain in which you're working in absolutely makes all the difference I will mention one other thing about coding and I guess this is both a constraint and a domain thing on the screen on the left hand side you'll see the picture of Africa with just one single line around that was a Sat 3 cable so this is circa 2004 I think up to 2008 we basically had one submarine cable that serviced the entire continent you can imagine that's not getting anywhere and see the landing points there's only a few countries that it even lands on now fast forward to 2014 which is on the right-hand side that the submarine cables around the continents increased by over a thousand percent all right there was a lot of funding that sort of came in from the World Bank and another and that really had people pumping a lot of submarine cables into our seas seabeds with more connection points and some that were actually connecting the inter locking countries as well now this of course changes things because even with the South three cable we had the South three cable had more connectivity than the copper that was in the ground so you can imagine what that has done in terms of connectivity on the continents and specifically in terms of mobile so we had nothing sort of mm and look at that by 2016 were already at 80 percent mobile penetration throughout the continent and the interesting thing is that most people have more than one SIM card all right so I will have I have one for every network in my country because maybe my aunt is on MTN so if I call her there's cheaper my dad is on Vodafone so if I call him that's cheaper so there is a little bit of inflation in those numbers but the reality is that there is I mean when I go to the village the only thing anybody ever wants from me is a phone and SIM card that's how how much it is valued so this has really changed the way that we build technology because it's not really people say we're mobile first continents we're not mobile first we're really mobile only so even when we are building websites we have to make the mobile responsive because that's how majority of people are actually going to access it so let me run you through a few stories just to kind of demonstrate what I mean by domain and context and all of that stuff and constraints so the gentleman there is called home in January see if you never heard of him he's kind of considered the father of software development on the continent Hermann kind of developed this software he has a bunch of products and I'll show you another thing that he developed as well and hae-joo law is a word in Ghana I don't know what language actually we have we have 75 languages in Ghana but five semi national languages and but Julho is used by everybody and what happens is if I go to the market and I say somebody stole something for me and I say jeweler jeweler jeweler that person stole something from me let me tell you what what happened everybody around me will drop their stuff pick up your trousers or their skirts chase that person pull them down and beat him up totally I mean this is like there's no police yes and the police will come and join sometimes they don't ask any questions this beats you up it's actually not good we actually need to change it because they do it can become great aggressive that can actually put a tire and light it up and it can get too aggressive and then people get really hurt and some some people are passed with it but we do need a justice system but Herman was thinking about well why can't we pre implement technology in a security system and rather than importing something from the West why can't we use some of this social phenomenon that happens and so what he did was he basically recreated that and you can use a simple feature phone press one basically speed dial or you know however you do your heart key to call the hydrilla number you would have signed up for the service and when you signed up you would put up to five people in your home in it up to five people five of your neighbors in the phone and when you press that number it basically will call all those people and it would call the radio stations and it will call the police and what would happen is the radio station will start saying house number seven two five five yes the one by the mango tree they have been robbed go go go go go go go and people would show up and it's very simple very simple application there was a dotnet application but just by using SMS he could activate all that and what they found was they actually created stickers out of that thing on the on the right hand side what they found was just by putting the sticker up with Haeju law and people knowing what it does crime went down so they didn't even get too many calls I think well the first one was a guy who was being beaten up by his wife who knows what he did and the other one was actually emergency call somebody was actually having a stroke and so that was actually helpful to to get some help there so that's one example that the bane and using the social phenomenon makes all the difference this one is a story about constraints so the gentleman there is drained off rained off very smart young lad at the age of 14 he was already starting University classes so he started early very bright and he was just ready to absorb and learn as much as he possibly could one of the things that he found was that everything was available online also software engineer actually worked for me at that works and he he was ready to absorb all the information but it was too costly to buy data to get all the information so what he ended up doing was he built a browser called an insane on github if you want to look at it and it's a it doesn't have a whole lot of features but the important thing it has is that you can actually store your favorite sites and when the data usage data cost is low which was typically between 12:00 p.m. and 5:00 a.m. it would actually go out to those sites and download the latest copies so you don't have to disturb your sleep to get all there it's just information so when you get up at 5:00 or 6:00 your news is there all the latest lessons and class are there and so that's an interesting solution to solving a constraint and I think that's one of the things we found is that constraints if you don't have the constraints you just don't come up with these ideas there's something about living in the space where the constraint happens that makes these solutions happen I had a staff person who had before she joined us had joined one of those competitions hackathons with a team and they had actually won and the prize was too big to get space in a hub for a year where they could actually have people who would help them kind of evolve the idea and their solution and it was interesting because she said when she got there because it was a nice space air-conditioned all that she said with all the air conditioning blowing in her face she forgot the problem she was trying to solve so your path is that they haven't evolved much so one of the things that we think about when we're actually building the software is we introduced the constraints we can't sit in the city and think we're going to build something for the village doesn't work you have to go to the village and experience it and figure out that men are not going to take the pill and so this is another interesting ones is this is also by Herman's company and this has even sort of evolved because now we have a lot more mobile payments but this is circa five years ago and the challenge was how do we get people to look how do we get past the cash challenge and so what they did was you know like phones we have a lot of prepaid cards and so he said well let's just get prepaid cards of like 10 CDs we use our currency is called CDs and so they would sell say 10 CDs on a card so you go to the vendor you give the vendor cash you get your 10 CD card which has your code on it you would actually then scratch the card when you're ready to use it put in the number but I don't know if you've used that card before or I need to find out exactly what the value is and so I would actually send an SMS or the so the user would actually type it in type in the number it goes by SMS to a processing server explain the serve in a second it goes to a processing server where you can actually get information well you so go to processing server right from the processing server I need to validate that against the entire list of numbers that I've issued so it goes to that validation server and then from the validation server it tells me yep you still have 10 Ghana Cedis left you could credit whatever the the product you're trying to buy is now will coming from the West we probably would think about literal servers or server space at least some kind of security as a cell authentication middleware something not in our part of the world the service became laptops the laptops had SIM cards so basically the first processing laptop was calling the validation laptop right so to basically sending an SMS rather to get that validation back and all they did for security was they would put the laptops in different locations under your auntie's bed in a drawer somewhere so very simple and frugal this is an interesting story out of Nairobi where Nairobi has the fourth worst traffic in the entire world Lagos has a second and in Lagos they actually have been instances where the traffic was so bad people left their caste and took motorbikes home and came back the next day to pick up their cars that's how bad traffic gets and so this that came out of the IBM Research Lab in Nairobi they decided to get cameras to put cameras around or actually be able to calculate and see the rates of traffic and where the traffic was was a really expensive so they took low fidelity cameras that as you can see on the bottom left could at least see the dark spaces where the roads were and so using that the black and white images of how dark the spaces were they would sort of see how much was being covered and uncovered and at what rates and they use that as an approximation of traffic flow and then just by sending an SMS to say I'm going from point A to point B it will tell you what is the most effective way to go and so by using that they were actually able to redirect traffic more efficiently this is an interesting project that we did starting about three years ago Ghana has a largest waste dump in West Africa which is significant because it has been a dumping ground for a lot of industrial companies and countries and it's this huge expanse of a space when we first started and we registered the users who were there registered the people who were there because there's a lot of people some of them even barefoot who are trying to make a living from the dump by finding old mobile phones that they can fix up and things like that within I think was 48 or 72 hours we registered 600 people who made a living in the sea waste em and so what we decided is we partnered with an architecture firm they built a building do they also were the ones who designed that structure that you see there which is a makerspace Eko craft which you can actually take down and move and the idea was to make sure that the people who were working in the e-waste dump could safely find and take apart the basically the products that they were trying to use to create new products that they could get designs basically for the products that they wanted to create and that they could trade with one another and so that application was really more about using pictures so it had to be a smartphone first of all using pictures and using voice you could call another person say I'm looking for this parts look at the picture do you have it let's trade so that was a very interesting use as well and I'll tell you one more because to me this is another this is an example of it takes more than one innovation so kind of put some of these solutions together because a lot of times the entire process that you're trying to enable or the chain that you're using you don't have all the things set like electricity is not necessarily there I need to find a solution for that it's connectivity is not necessarily there I need to find a solution for that so this was a perfect example and for this one what we really wanted to do was we wanted to organize the paper in hospitals and clinics their medical records have actually trained medical information so it's a person I worked with on this and he will tell you that you know what we talk about connectivity and big e Amar's and all that but Pippa cannot go through wire so we have to find another way to organize the paper and then a way to transcribe what is on the paper into electronic form so the first thing we needed to solve was how to organize paper and as you can see it's highly disorganized on the right-hand side you'll see sometimes it spills out of the Mehcad room it's sort of all over the place and people can't find your records we found that the average wait time in OPDs was literally three hours and which is way too long so you actually see a doctor a lot of people would just leave and so what we ended up doing was first we cleaned up all the records in some cases we actually found that people would have 14 folders that's because every time they would go to the hospital they can't find the old record so they would just buy a new folder and start up a new record and records anyway were destroyed in seven years so your medical history doesn't really exist unless you take care of your folders and so that's what we found people would take the folders home to make sure that they had their medical history whenever they needed to go back in so the first thing to do was to organize the organizer medical the files and deduplicate and you can see on the bottom left what that started to look like so you can easily find where you're going to find that record and then on the right hand side we bar-coded everyone and we only captured eight pieces of information your first name your last name your first mobile phone number your second mobile phone number your national health insurance ID your next of kin their number and I've forgotten what that what the eighth piece was but that was all the information that we collected and we put into a system so when you come in which you can just search by your name most of the time it was actually a phone number and we would find your record swipe it and you're good to go so remember that three hours wait time that I mentioned that's what it became ten minutes just by doing that and now we can build an ear mouth EMR on top of that which the government is actually doing now one of the interesting things we found is that the mobile phone number was really interesting because my name I have five names betty Rachel in your name ABBA now come home but I've only one last name but I can go anywhere as Betty kuma ho and your como ho it's completely legal there are some tribes even within my country where your last name is not even static so I can show up one day as quick who Mensa and the next day I can be Nana before it's perfectly fine and so the name we can't trust that the name is actually going to allow us to uniquely identify you but people don't mess with your phone numbers so we literally found like one guy who had been sending and we only found him because of the phone number that was tied to the to the record he was using his national health insurance for eight people and so we we actually were able to use that to uncover some fraud within the system as well so I'll tell you one more story because it's interesting when I give this talk because a lot of times this it's a very different environment from Chicago I think so people are like how does this apply to what we're doing and I don't know the answer to that question yet I think it'll be interesting to start working a bit more together and figuring out what can port and what can be learned I'm certainly doing more thinking and thoughts around what makes frugal innovation work and how do you introduce innovation into more mature environments and companies and there's definitely some lessons learned around that and some of that we've already seen right it's it's fail fast keep evolving the model it's introduced constraints that's a big one if you want something that's cheaper make sure that the budget goes down all of those things but this was an interesting one because this is a project that we contributed to from Uganda when I was working in the company there and it it is actually sponsored by the UN rapid FTR stands for rapid find trace and family trace and unification so in areas of natural disaster or conflicts children get separated from their families and so what rapid FTR does is with a trained professional you can come in and take information about the child and use that to actually find their families and this was interesting because we had actually first implemented it in the Congo when there was a conflict there and what we found was that the the paper format the carbon sort of paper format that they had used before would have taken 11 months to register I think was about 10,000 children and this system it took three weeks to actually do that so your reuniting children with their families a lot quicker and when we found out about the Philippine feel of Philippines typhoon I think that was Bobby five years ago or so we were ready bleep laced to help with that so what we ended up doing was we literally just went out into the store bought some Android phones contacted the people we had worked with at UNHCR and within a few days they landed in the Philippines and with rapid FTR actually had about 600,000 children identified within the system and mapping them with your children so for me this was really the one of the examples where we've actually been able to build something on the continent that has that is exportable that we've been used in other environments and so I would like to see more of that and I'm looking forward to that so and that is it for my talk you [Applause]

One Comment

  1. ATESCM Rock Player said:

    Nice video

    June 30, 2019
    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *