Making User Research in Africa Accessible and Iterative

Yann Le Beux at UXDX Community: EMEA Online Series. Video: https://youtu.be/fDXKHWOYAk0

Readable transcript: edited from the recording's captions for readability (fillers and false starts removed, punctuation and section headings added). Wording is the speaker's own. Timestamps are positions in the video. Names marked [?] could not be verified against the audio.

Why we built a research company in Senegal

[00:00:01] Hi everyone, and thanks so much for having me today. It's a pleasure joining you to share about what we're passionate about, which is doing user research on the African continent. I'm not sure you've seen a lot of people doing user research or UX in Africa so far, so I'll try to summarize the way we do it in our design agency, but also some of the learnings and the tools that we built along the way to help us do that and make it accessible. There is a lot. We've been in business for six years building this community from Senegal, so if we don't have the time to cover everything, please drop some questions, I'll be very happy and excited to chat with you afterwards.

[00:00:43] A bit about me, you can read it, but the most important maybe to remember is that I've been here for quite a while and always frustrated by how startups in the tech world seem to copy paste models that come from the Western world, and it actually doesn't work. It doesn't work because the usage, the behaviors that we see on the African continent are very peculiar and change a lot from country to country, or even city to city. So I'm pretty passionate about making tech more accessible and relevant for the African continent, and we built this company six years ago to solve that issue. You can follow us online. I still love to do UX research myself and write about it, but also write about how we build a design and research company on this continent.

[00:01:31] Why we created this company in the first place is to make sure that people can project themselves in the future, with their needs, their identity and culture represented in the technology that they will use on a daily basis. This is why for us it's very important to make research and to make design accessible, so that everyone can do it by themselves and more people can actually ask users for the right thing to build, and then in the end create products that are really tailored to the African context.

[00:02:00] We started like this, a bit of our story, because it touches a lot on what we want to make and why we want to make it accessible and iterative. But this is how it started. It's basically us with my co-founder over here that you see with the T-shirt, sharing with a group of people, talking about user research, talking about UX design, and this is how we started from Senegal six years ago.

[00:02:23] User research has been central to us since the beginning, and this is a picture of one of our colleagues doing user research with street vendors at a time. That fascinated us, to see that in the informal economy there are a lot of people that actually are very central to everything we buy and do on a daily basis, but the use of digital is totally unknown. And for us it was very interesting to start digging into this community of street vendors and understand what could be the potential needs for digital solutions.

[00:02:54] We launched the first app, that was a huge failure, never worked. It was for educational content. But building this first app we actually realized there was an opportunity to build a product that can have them, the street vendors themselves, doing user research for others. And this is the tool we built that is called Luca. I'm going to come back on it in more detail today, but this is how it all started, and for us this is really about making research accessible and in the end for everyone.

The company today, and who we work with

[00:03:23] Today we are a bit bigger than when it started, so we are a team of around 40, 42 full-time people, across different countries on the continent. As you know, this continent is very, very big, and covering it all for us has been a challenge, culturally and in terms of logistics, and so many of the things about building a company, but that's not the topic of today.

[00:03:44] We're basically very excited that we've been able to work with top brands like Google, Wikipedia, or Meta, Square, across the few years that we've been in business, but also a lot of big NGOs, for which we do projects on education, on healthcare and agriculture, so that we can also have a huge social impact doing all the methodology that we are passionate about. The third tier of clients we're very excited to work with is the local startups, and I'll come back a lot on that today, because actually if we want to make research accessible and iterative it's mostly for them. They are the ones who don't really have the means to pay an agency all the time. This is why building tools that could be tailored and adapted for them to do research on the continent has been part of our mission since the beginning.

Why research has to be accessible in Africa

[00:04:31] Why make research accessible in Africa? For us it's just a matter of understanding better the context. As you may know, there is almost no data whatsoever that exists on most of the users. It's not like you can go on Statista, just download it and you have beautiful data already. It doesn't work like that over here. It is very opaque and it's very hard to find reliable data. Literacy is very low, and digital literacy as well, so you always need to understand your user on a very specific basis, a case-by-case basis, otherwise you may totally fail building and launching a digital product.

[00:05:08] Connectivity is still an issue in urban cities and rural areas notably, but this is why you need to really take it on a case-by-case basis and do more research to understand if there is any barrier to adoption of your service. And last but not least, in general people here don't trust organizations. This is why you don't have a lot of room for error when you launch your product. People will take you very seriously, and if you fail satisfying them they will put you on the side very rapidly and not ever come back. This is why you need to get it right quite rapidly. But more importantly for us, as I mentioned in the introduction, it's about giving a voice. It's about making sure people's cultures, behavior, identity is actually reflected in the products they will use on a daily basis.

The state of user research on the continent

[00:05:54] So to understand a bit better about what we actually do, we did, for the second year this year, a big study that is public, for free, on the state of user research. And it is from those types of studies that we can actually understand what are the things that are needed for the design and UX research community on the continent.

[00:06:13] A few stats, so that you get a bit into the head of a UX researcher in Africa. How often do companies do user research here? As you can see, it's not often. Most of the companies do it less than monthly, and most of them don't do it more than quarterly or three times a year. Naturally, it doesn't mean this is very accessible and very iterative, so there is really something to solve over here to make it a bit more continuous.

[00:06:43] How is user research done? Basically, what methods are UX researchers or designers doing on the continent? And I'd love to get your feedback on how it compares to your different markets. But a lot of people focus on usability tests or individual interviews, which is great, but for them it's been really a challenge organizing those. And actually surveys and quantitative studies are done, but I'm not sure they're done very well, because in the study we also have the tools, and most of the people use Google Forms and are not necessarily trained about the right way to do quantitative research, and so a lot of wrong data is actually coming out of this research.

[00:07:18] So what are the main challenges people face doing research? As you can see, it links directly to the accessibility and the fact of doing research on a continuous basis. So first is sourcing the right participants. As you can guess, finding people who are reliable, available, connected, of different profiles, is really, really hard on this huge continent with so much diversity of profiles and behaviors. Second is the time availability. Companies here are resource constrained for most of them, and so they don't have a full stack research team available. So it's mostly one researcher, or sometimes the UX designer that has to do everything by himself, maybe with the help of the product manager. So time availability is definitely a constraint. And this is actually following some of these constraints that we noticed four or five years ago when we started the company, that we decided to build different tools to make research more accessible and iterative.

Luca: surveys, a data hub and a panel

[00:08:15] And this tool is Luca. Luca for us is really about making research into the hands of people who are doing research, be it product manager, be it researcher or designer. We want to make it simple, easy, but also more importantly reliable, to collect data on the African continent. As you can see, some of the top clients we work with are actually using Luca for doing their quantitative research.

[00:08:42] So what is it in a nutshell? It's a research platform, so it's a SaaS platform dedicated to Africa. Basically you have a survey component for your quantitative research. What we call Luca Hub is where we can find existing data that we collect on a recurrent basis, because we notice people are asking often the same questions, so why wait for them to go and ask the question? We can ask for them beforehand and give it ready for them to use. And then after, there is the panel if you want to organize your qualitative research.

[00:09:10] I'll get a bit more into detail and then after I'll show some examples on how we used it. So basically, Luca survey is face to face. And why do face to face? Just because in Africa you have a lot of people who are not tech savvy, who don't even have a smartphone to which you can push online forms. And having a network of 400 surveyors, like we have in 18 countries, for us has been key to help our clients easily understand what's happening in the streets, the behaviors. And having those surveyors really knowing the context has been also very key to make sure the data that comes out of it is really reliable. Outside of that we also have the capability to do online, through the panel that we have available, and pushing a form to the ones who are able to use a smartphone.

[00:09:59] Naturally there is an app, there is a tech component into that. So you need an app for the researchers that are on the ground, so that it's easy for them to collect the data, but you also need an easy way for people who are not experienced researchers, or maybe Excel users, to be able to correlate the data and do some basic presentation of this data, and this is everything we've built inside the Luca platform.

[00:10:26] As I mentioned, the hub is about, you find a report that interests you, some specific type of data, you just buy it and you're ready to go. In a few minutes you have the data that you need to inform the first step of your product development. And if you need more detail, or tailored, or a different geography, then you can go for a more tailored type of survey.

[00:10:47] And lastly, the panel is about providing you the tool for your qualitative research. Qualitative research, as you all know, could be one-on-one interviews, could be focus groups, could be diary studies. But this one, sometimes you can do the research yourself, but you struggle really finding people to do the research with. As we've seen in the data, recruitment is the key issue for UX research around the continent, so having access to a panel of reliable people has been tremendously useful for most of our clients.

[00:11:20] So how does it work? Basically we find people in the panel that keeps growing every month, every week, people registering, applying to use the panel. Second is profiling, and this is really the peculiarity about running a panel like this on the continent: you not only have a form, but you have to do almost small interviews of each person to make sure they are reliable, otherwise a lot of people will just apply to the panel to get a bit of money and will not really fit the profile you asked for, and it actually requires a lot of time. Second is the scheduling, and scheduling over here is actually such a challenge as well. You need to make sure people are on time, you confirm, and you always need to have options for them in case you have a no-show of one of your panelists.

Client examples: Google, Wikimedia, Bitcoin and Chipper Cash

[00:12:07] So, a few examples of some clients and companies that do this type of mix between qualitative and quantitative research. The first one we worked with recently, this year, is Google, that wanted to understand better the users in Nigeria and South Africa and other markets. For them it was key to have this two-sided approach: the quantitative approach to understand what's happening in the market that they don't know at all, and also the qualitative approach to inform what could be the ideal features or pain points to solve for the users. And it was really exciting when we finally were able to see the users of Google and go with their teams on the ground in Lagos to interview users, to see the reaction and differences, and also the complementarity between the different types of data we collected before.

[00:12:55] The second type of client we do this type of work with is Wikimedia. Wikimedia is an interesting example because we've been working with them for more than a year right now, and what I really like with Wikimedia is that they don't try to understand everything at first, at the same time. They really focus on various tiny features, like the registration process for content creators for instance, or how you read specific pages. So it's very small chunks of research that you do on a very iterative and continuous basis. And for us that's the ideal case where you can mix quantitative and qualitative data, because you don't have to ask thousands of questions or dozens of questions, and it could be very straightforward and you can go very fast iterating on each element.

[00:13:42] We did a very interesting study on Bitcoin usage in Africa, and this one actually I took because it's a case of non-iterated research. We only did it once. But what's fascinating with Bitcoin usage is that it's grown so fast, every six months it almost doubles. So it's impressive that for some topics you really need to have a continuous approach and be able not to do a huge research once a year, but be equipped with the different tools and processes and teams to be able to understand how fast the continent is actually moving. And I hope Square understands it, and actually recently they recontacted us to use Luca to do their research by themselves on something very specific as well, and we're excited to be able to keep working with them.

[00:14:26] And recently we also worked with a unicorn called Chipper Cash, based all across the continent, and they wanted to understand how it was working in francophone Africa. So for these tech giants coming from the anglophone world, understanding how technology and usage may be in Senegal and Côte d'Ivoire is actually very scary. It seems to them very, very different, not only in the language barrier but the cultural barrier and the digital behaviors overall.

[00:14:52] So what we did for them is a mix of quantitative research and qualitative research. Quantitative to understand what were the consumer behaviors and the profile of participants, kind of work on the segmentation and understand what was happening in those two markets. Then we had qualitative research to better understand the pain points, the behavior, the frustration and the needs, so that it could inform the design of the Chipper product on the continent. Then after, ideally what will be done is back to evaluate all those products and concepts through another round of quantitative research that's much more focused on the ideas that you've been able to generate in the qualitative phase, and you can finally come up with very tailored recommendations for the product that has been tested for the different markets.

[00:15:44] Some of the deliverables: it's basically using the Luca platform to do correlation very rapidly. You can see differences between Abidjan and Dakar, or Côte d'Ivoire and Senegal, and really inform key insights but also follow-up questions for your UX research. And that's the beauty to be able to mix quantitative data with qualitative, is that quantitative will show behaviors but will not tell you why these things are happening. So you can note all these questions and then after ask them in your UX research or qualitative phase. And using the platform Luca for making easy correlation like that on the fly was for us very, very convenient.

[00:16:25] Second, you can definitely have personas, user profiles like that following your qualitative analysis. I'm not going to come back on that, it's probably a tool that you guys use a lot. But what's interesting is that after that you can mix both qualitative and quantitative data to address a very specific theme. So for instance if we talk about cross-border money transfer, that is actually not very well documented, or you have no data on Africa on that, you can directly mix the quant and the qual to generate key takeaways and actually have to brainstorm on the opportunities. And at the end you have a happy user. I love this picture coming from Chipper, because at the end of the day, if we do all the things and we have all this process, it's for that. We want to make sure people, whatever type of product they have, can benefit from it and can grow themselves.

Key learnings

[00:17:20] So, key learnings doing what we do, and I think after that I'll open up to a few questions. As you probably understood, for us it's very key to do research often, but to reduce the research scope. So focus on specific features, specific themes. Do it as soon as you have a doubt, as soon as you have a question, be ready to start some research, and don't make it over complicated. And ideally mix quant and qual to always have the mix between the macro perspective and the micro perspective on your users.

[00:17:53] Work with local researchers. Naturally, as you can understand, it's not me, Yann, as a Frenchman in Senegal, even if I speak a bit of Wolof, I don't have a lot of value doing research myself unless it's in English. But you really need people who speak the local languages and, more importantly, understand the context. This is why you need to have a network of local researchers. We just listed here a few of the languages you can have in Nigeria, but you have hundreds of them, same in Kenya, same in South Africa. And this is really one of the key elements about doing research in Africa, is that you need to be able to have people who speak those languages to understand your user. This is a challenge, but there are solutions for that.

[00:18:35] Train and select for quality. A researcher for us goes through a very selective process, and that goes through training, because you just don't need people who will fill a form, you need people who will create the empathy with your user like a qualitative UX researcher will do, even if they fill a form at the end of the day. And for us this is a core focus. If you want to make sure you have data of quality, this is where you need to put your focus.

[00:19:01] Verify data automatically, ideally, so that you can scale, but don't remove the human aspect of that. For us it's always been very important to have people looking at the data and trying to find inconsistencies in the answers that we have, and we always have data analysts making sure that behind all the automated processes we have, you can also verify the relevancy of the data that comes up from the data collection tool.

[00:19:30] And that's it for me. I hope I wasn't too long and I wasn't too draconian[?], and I could enlighten you and also encourage you to do research on this beautiful continent. It's challenging but it's exciting, and there are so many opportunities. As you know, this is the continent of the future, so I hope all the tools that I shared today would encourage you to build more things that are relevant for the continent. Thank you so much, guys.

Q&A: scoping generative research

[00:19:54] Rory: Thank you, that was a really interesting conversation, and kind of looking at some of the challenges that I had perceived, like the lack of data that you can search quickly and look up when you need it.

[00:20:10] Rory: Just on questions, we did have one from Alicia asking what are some recent research questions you've worked on, but that was before you went into all of your examples. She has said that you answered it, so thanks for that. I have a question, and please, if there's other people watching as well, please do ask your questions as well. You mentioned specific questions and you gave the example of Wikimedia and their login flow, or registration sorry. That's quite easy I think for evaluative kind of work, but when you're talking about generative, can you give some examples of ways people can be specific when they're starting to look at some generative work?

[00:20:57] Yann: Absolutely. So one of the key topics I've been working on recently, and actually wasn't in this slide, was agriculture. And agriculture, as you can guess, it employs something between 70 and 80 percent of people on the continent, so it's not something where you can just talk to a farmer and understand everything. You have to understand the entire value chain, so the scope for generative research is like super broad. So where do you start?

[00:21:20] Yann: And so we divided the work in different what we call use cases, the different use cases in which the farmer is involved. The first one is, how do you profile a farmer, and how do you make sure the person you have in front of you is actually a real farmer? If you want to provide a service to this person you need to identify that. Then after, you have to make sure this person is actually producing some crops and procuring, so bringing his tea for instance to a collection center, that's the second step, so that you can actually have a farmer that produces something. And we did a first round of research only on those two elements, to understand who are they.

[00:21:57] Yann: Then after, we followed up and focused a bit more on the payments aspects and the financial services. It's only once we've been able to understand that we actually had farmers that will procure crops, tea for example in this case, then you can understand how they want to be paid, by mobile money on their apps or not, go to the bank, and then also what financial services you can build for them, like basically loans and insurance and everything.

[00:22:21] Yann: And we really divided that step by step, because otherwise you just overwhelm your farmers with tons of questions, and one hour in-depth interview is very hard for everyone, but for a farmer it's even worse. So you need to keep it short as much as you can, otherwise you can be sure that the guy is going to just answer whatever to get rid of you, and we had that exactly at the beginning.

Q&A: sharing interviews, cameras, and notes

[00:22:43] Rory: Great, well that's a great one. And then the other question that came up was, I often find value from having multiple people look at an interview, because different people pick up on different things. So are your interviews recorded, by any chance, or is there any way that, let's say, the team who didn't do it would be able to watch the interviews and just see if they can pick up something different?

[00:23:16] Yann: Yeah, absolutely, and even cooler. What we did for Google, we actually did almost like a happy hour where people were watching the users' videos. So we will record most of the interviews we do remotely, and in this case we used Google Meet to have the users in Nigeria and South Africa, and then after we had this kind of happy hour where people from Google all across the globe were just watching the video we edited from our users. So the hard work over there was to edit the video to make sure we take the most interesting elements that these users were sharing. And I think that that works a lot. We also insert those videos into the reports right now, so that it is really live.

[00:23:53] Yann: Naturally, that's when you can do remote interviews with people who are a bit tech savvy. If you interview people on the ground, generally we try to avoid using a camera, because it really breaks the entire relationship. So for instance with farmers we never show up with the camera. We may record just the voice, but even if they agree to that, sometimes we prefer even not to do it, to not risk to have any bias or problem on the side of the farmer. Then actually all the notes that we have are shared. Most of the time we use tools like Notion or Google Docs, and recently a tool called Notably has been very useful for us to share those different notes to other people and part of the team, and then after do the analysis on Miro, where everyone can see.

[00:24:40] Rory: Excellent. It's a really good, nice mix. And it is always that challenge of, do people start playing up to the camera?

[00:24:48] Yann: Yes, no, which is always weird, like the way people behave in front of the camera. I mean, they will prepare a ceremony even.

[00:24:56] Rory: Yeah, yeah. Now, there's many challenges.

Q&A: letting surveyors follow a tangent

[00:25:00] Rory: Great. I'm just checking to see if there's any other questions from people, so now is your last chance if you want to ask Yann anything. And I'll go with one last one. You mentioned that you prepare a kind of a script for your interviews and then you have the trained interviewers ask them. Do they ever pick up on tangents, or do they always kind of stay with the script? So if somebody gives an interesting answer they weren't expecting, are they trained to maybe pull on that thread?

[00:25:33] Yann: Pull on, absolutely. This is actually a great feedback because we didn't think of it at the beginning, but actually our surveyors were really frustrated because they came back with an entire notepad of follow-up insights. It was supposed to be quantitative research and just filling a survey, right? And so [?], but then naturally some of them followed up and they came back with a lot of things.

[00:25:55] Yann: So what we invented is that there is a voice feature in the application where you can even record extra insights via your voice, or even just by typing the text. So you have at the end of your questionnaire an open text field where the surveyors can add the extra insights, and we train them to do so. Because what we really, really want is to make sure the quantitative research we do is more and more qualitative, so that it goes into the realm of UX research and qualitative research more and more. And also because that's why it's super frustrating, if you just fill a survey without having a real conversation you're actually going to get tired very soon.

[00:26:35] Rory: Great. Now, it's great to see that you've covered that use case as well.

[00:26:39] Yann: Yeah, thanks for the question actually. You need to be an expert to ask this one.

[00:26:45] Rory: Well, I think that's actually all we have time for, so thank you very much, Yann. I really enjoyed that, and thank you for sharing.

[00:26:50] Yann: Yeah, very cool. Thanks, Rory, for having us, and I'm open to any question that can come afterwards. Thanks guys.