How to conduct an ethnographic study before launching your app ? Street vendors case study

The street sellers (who we may also call mobile vendors or ambulant sellers and many other terms) are the object of very little attention on the part of the technological community (and general population?). Despite the lack of available data on them, these entrepreneurs visibly represent an important economical activity in African cities and should, like everyone else, be more and more connected in years to come. We found it important to dig deeper on the matter. Our goal in sharing freely the results of this Human Centered Design study is to initiate a thinking on the development of innovations adapted to street sellers. In a more general initiative, we also want to share our methodologies so that a maximum of startups may use them in their respective fields.


When you spend a little time outside the African continent, what strikes you is how differently people use digital technology compared to the occidental world. Very few studies are done on the matter and it is actually quite hard to measure, but it sounds clear that in the majority of countries and cities on the continent, we daily create new ways to own digital tools and render them useful to life, survival, gratification, personal life etc. This is due to oral culture, to connectivity, to literacy, to the discovery of the web via mobile rather than PCs, to the size of cities and to the transport conditions, the important of familial circles etc.

This is one of the reasons why we launched UX a few months ago : it appeared primordial to understand the technological habits of Senegalese people in order to develop useful numerical tools adapted to our context.

After a first Human Centered Design study the sharing economy and crowdfunding in Senegal, we decided to focus on street sellers and to understand their needs, aspirations and numerical uses. Our friends from ByFilling immediately joined us on that project because they are very interested in the digital content that can educate and interest these people.

The subject was launched, it was now necessary to pin down the objectives, the perimeter and the methodology. This is how we went about it, it could be useful to you.

1)  Definition of objectives: quantitative or qualitative study…? And what’s the difference actually?

qualitative study ux design

The first question you should ask yourself is: “what am I trying to understand?”

  • OPTION SMALL A : I want (to try) to understand, objectively, the economy of street sellers and to quantify it: numbers and types of merchants, places, ages, revenues, number of coffees sold, number of oranges peeled, etc..


  • OPTION SMALL B : I want to (try) to understand their aspirations, frustrations, technological uses and needs in a more subjective manner. 

In order to choose between the two options, you then need to ask yourself the following question: “what can I do with those?”

  • WITH OPTION SMALL A : nice little infographics to put on your Facebook cover page (no judging, I’ve done it as well), a report to send to your president (well, I tried that, without success), a big excel file to train yourself to big data, sound interesting in conferences or during evenings with your numbers etc..
  • WITH OPTION WITH B : generate concepts of useful applications..

You’ll have understood it- in order to engage these themes, we chose option small B- it appeared more interesting, fun and original- but we sometimes like quantitative studies sometimes.


When you zone in on a subject, especially if you choose the qualitative option, you better study deeply a very precise population and limit your scope rather than try to skim over the totality of your target. We first listed all the street sellers and spread them in 3 categories:

  • Mobile Street vendors: credit sellers, sellers of diverse objects, Touba coffee
  • Semi-mobile sellers- set up their stands in the morning, wrap up in the evening in a close location: sellers of fruits, peanuts etc.
  • Fixed street sellers: fixed stand that stays in the same place even at night: tanganas, some fruit sellers etc.

Street vendors yux dakar

We then chose only 4 types of merchants, depending on the people we already knew or our interests: the fruit seller, the coffee seller, the Touba coffee seller and the credit seller. We also reduced the study to the zone of Dakar.

3) Narrowing down the methodology, the calendar, the budget

Once you have your perimeter, you have to put on paper the operational elements of your study. It is recommended to interview 5 to 10 people for each profile that you’re trying to analyze. In a zone like Dakar which is very diversified, it is of course important to choose people in different neighborhoods (residential, business, popular, chic, etc.) Then, you have several options for collecting information:

  • Field interviews called intercepts

field interview yux dakarImportant if you don’t know a sector because it allows you to understand the context and environment in which your target works. It was particularly important for us to understand the clients stream of our sellers, the relationship with other sellers, with automobilists, travel conditions etc..

  • Semi-directive interviews

icon2You invite the people you wish to interview in your office for a 1h-1h30min interview. The “semi-directive” means that you follow a precise set of questions but you can get out of that to dig into an aspect or an answer that sounds interesting. We’ll get back to that type of interview in another article. Here, just remember that it’s important to do that to at least 2 colleagues: one who asks questions and talks as naturally as possible with the person and the other one who takes down the answers (and says nothing, if he can 😉

  • Focus groups


Very useful when you already have an idea of an application’s prototype and you want to get feedback from a group of people and let them debate it a little bit. We haven’t done it at that stage but certainly later.


There are, of course, other qualitative analysis methods but the three above are among the easiest to access and the most pertinent in the first steps of a project. As far as we’re concerned, we did a mix between intercepts (25) and semi-directive interviews (8).

Depending on all this, you create your calendar and your budget. If your accountant is ok, then read the rest of this article…

4) Creation of the interview guide

questionnaire yux dakarThe heart of your study- spend time on that. The interview guide corresponds to the set of questions you’re going to ask. It is crucial to respect it from one interview to another and to have objective and representative results. The structure of your guide can vary but in general you’ll find approximately the same type of questions.

  • Profile, path, and aspirations of the person.
  • technological and media uses.
  • Description of the activity, value chain
  • Focus on frustration and needs.


You go meet people and talk with them. Not a lot of methodology here, to the contrary. You have to bet on your relationship skills to reassure people and earn their confidence. Don’t ask your questions by looking at the questionnaire under your eyes, don’t wear suits, talk about anything and everything at times etc. You see what I’m saying right? And most importantly, if you’re interviewing the coffee seller…. have a coffee! Even if it’s 6pm and it’s your 12th!


Invite people with whom the atmosphere was best. Choose people who represent the typical profile but also don’t hesitate to be surprised by original profiles. In our case, we knew a very creative coffee vendor and also met a young woman, with a diploma, who sold fruits. Design likes to feed off what it sees in “extremes” (we’re not talking politics of course).

Important detail #1: think about the language in which your target is most comfortable. In our case, interviews were done in Pulaar and woloff depending on profiles.

Important detail #2: Think about recording your interviews in order not to miss any detail and take photos in order to help colleagues and readers understand context.

Important detail #3: an interview for a UX study must be paid for. Unless, of course you’re only getting a few minutes from the person or it’s one of your clients or partners: the amount depends on the profile and time requested. In occident, there are even scales for that, but not here. We have, in our case, paid 10.000 FCFA for each interview that lasted between an hour and 90 minutes.

Important detail #4: think about the place where you’ll be leading your semi-directive interviews. If you’re a big boss with classy offices, the seller might be uncomfortable and it could influence his answers and his openness towards you. The important thing to remember is that you need to make people comfortable, think about giving them a drink and a little food, always appreciated.

7) results analysis

analysis yux dakar

The most complex. You have got to be very meticulous, re-listen to all the interviews and not let cognitive biases guide us. You have to clean up all answers and share them with your team so each can acknowledge them. It is recommended to work in teams through workshops. The deliveries, at that stage, can be multiple. Since we’re very oriented towards sharing and results and the creation of applications following this study, we bet on a restitution under the form of “personas” and “experience map”.

8) Creation of deliverables.

Personas street vendors yux dakarPersona: Fictional character representing a target population and helping to think about products and services to develop for these people. The persona is made following interviews of real individuals and can represent either the average of the people met or an interesting extreme. It usually describes the aspirations and frustrations of the person, his/her path, activity, capacities and technological uses.

What takes time, here, is to define the structure and elements that you’re going to present to personas. You have to be precise but remain synthetic…

experience map street vendors yux dakarExperience map: The experience map in French describes the path of a user in contact with a service or system. It can also trace the typical day of the latter. His/her graphic representation, under the form of a chronological diagram, gives a detailed view of experiences and emotions lived by the user at different moments.

We hadn’t expected to do them early on but when making the personas we realized we were missing a few details on certain points of the day of the merchants. We therefore called them by phone to ask them a few precisions and were able to create experience maps for the profiles.

Important detail #1 : Persona and experience map are not cards of a person in particular-they are supposed to reflect a typical profile– so several persons interviewed with common traits. Nevertheless, we found an interesting profile (the fruit seller with a diploma) and made a persona just for her. It could come in handy later.

Important detail #2 : make it niiiiice ! It’s not easy with the ton of information you have to put in, but do your best to make the deliverables readable and agreeable. 


In order to then get started on ideas of solutions, we regrouped problems and frustrations experienced by the sellers in several categories: business management, work conditions, training, information, dialogue with the authorities and value chains.

We did not classify them by order of importance because even if 2 or 3 aspects come out often, it is very difficult at this stage to know which problem has a bigger impact than the others in the eyes of the merchants.


The rest is pretty straight forward. Take the problems and brainstorm on digital solutions that can help to solve them. We generated 30 application ideas. It is important, at this stage, to not wonder about the technical, economical or political feasibility of your solutions. It’s just a list..

UX design process YUX Dakar 1

Then, we decided, depending on the subjects of interest for us, of our mission, constraints and of what could be useful to the merchants, to create one of the listed apps. You can see the results in a few months but this is the steps that we went through.

UX design process YUX Dakar 2


  • Language : Surround yourself with colleagues speaking the languages in which the people you want to interview are most comfortable.
  • Openness of targeted individuals: You’ll notice that in our personas, you can’t find the Touba coffee seller identified earlier on. We were not able to find the keys so that the people we met would open up and tell us about their day to day lives.
  • The fear of authorities: Most of the people we met were suspicious in the beginning. Some are sometimes in illegal situations in Senegal, others might have had bad experiences and are therefore naturally wary of public authorities.
  • Wariness towards corporates: most merchants we met felt wrongly considered, sometimes exploited, by big businesses. They do not want to share their secrets at the first time of asking.
  • Expressing emotions. This is perhaps the most striking conclusion of our study: it is very hard to gather strong emotions, positive or negative, from the people we met. There are multiple reasons for that: culture, humility, religion education, trust etc. You have to know how to read between the lines and pay attention to the smallest detail when a person answers a question: a look, movement, smile, silence etc. Some “I’m good” can sometimes say a lot”.



One of the frustrations of people in the informal economy in Senegal is that they are often questioned and interviewed for large scale consumption products or things they don’t really need. We are trying, with this project, to have a co-conception approach in which the merchants are implicated throughout the whole creation of the applications in order to produce something that could be useful to them. Then, only then, will the thinking around the business model. If you conduct a study of the sort, do your best to give it a following, and produce something by keeping the human aspect at the center of your thinking.

Thanks for reading. Download the full Human Centered Design study for free here (in French):




2 thoughts on “How to conduct an ethnographic study before launching your app ? Street vendors case study

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s