Interview Methods

This research project is based on semi-structured interviews with 27 micro-entrepreneurs. In this note, we will describe how we selected them, how we interviewed them, and how we analyzed the results.

Sample selection

Like most qualitative research, our sample was purposive rather than strictly representative 1. This approach allowed us to identify a wider range of experiences and themes. This qualitative work will be paired with a representative, quantitative sample of micro-entrepreneurs in 2019.

That said, we were guided by the “2016 Micro, Small, and Medium Enterprises (MSME) Survey” issued by the Kenya National Bureau of statistics 2. This report suggests there are 7.07 million non-agricultural enterprises with fewer than 10 employees in Kenya, the bulk of which are sole proprietorships, or businesses consisting of a single employee. The top categories for MSMEs are retail trade and apparel. Roughly 81% of microenterprises are informal, meaning they do not maintain a registry with the registrar of companies.

We began with a baseline to mirror this distribution, but then reallocated some of the interviewees away from the retail field. Our final distribution looks like this:

  • Wholesale/Retail (e.g., retail trade, auto mechanics, wholesale trade): 9
  • Accommodation & Food Services (e.g., hotels, bakers, food delivery): 5
  • Service activities (e.g., computer repair, sports activities): 5
  • Manufacturing (e.g., apparel, food products, fabricated metal products): 3
  • Other (transportation, financial and insurance activities, education): 4

We worked with a Kenyan researcher and recruiter to identify prospective interview candidates. Interviewees received 2000 KSH (22 USD) as an honorarium for participation. The goals and methods of the study were fully explained at the time of recruiting and again at the interview. Participants indicated informed consent via a signed document. Participants also agreed, separately, to include their photograph or not in the study. All of the participants’ names have been changed, apart from the five persons who agreed to have their videos. and details about their businesses, on this microsite.

Study design

For the interviews, an experienced researcher was paired with an enumerator/translator to capture data. The interview was a mix of semi-structured and open-ended questions about each entrepreneur’s history with their business(es), their current operation(s), and aspirations for the future.

Additionally, about half the interview was conducted using a “digital day” reconstruction/elicitation technique. In one sense, this is as simple as asking people what they did yesterday. However, Caribou Digital has modified the basic technique to fit mobile-centric and mobile-only users in the global South 3.

Here is an example of a completed digital day from the study:

And the digital version:


Analysis proceeded along two tracks: one structured, one more inductive.

The structured component mapped key platform practices to the “FiDA platformization lens” 4. That lens identifies four core functions of marketplace platforms for fragmented, inefficient markets in the global South: convocation, aggregation, and gatekeeping; transaction support; credibility and governance; and ancillary value creation via data aggregation. We scanned interview notes for the platform practices—that is, the user behaviors carried out in reaction to platform functionality—corresponding to each of these four elements, which are discussed in the four “platform practices” sections linked above.

The inductive component of our analysis allowed for the fact that not everything we observed fit cleanly into one of these four categories. Through our discussions on the research team (Nagula, Schiff, and Donner), we identified two main themes of “blurring” which required further discussion: “tech and touch” and “the digital side hustle”. We wrote an essay on each of these blurring themes.  

The final set of essays, “takeaways”, merges the results/observations uncovered in the deductive and inductive essays to translate these observations into potential actions linked to specific stakeholders in the broader community: policymakers, technologists/product planners, and researchers.


  1. Dorothy Horsburgh, “Evaluation of Qualitative Research,” Journal of Clinical Nursing 12, no. 2 (March 1, 2003): 307–12.

  2. Kenya National Bureau of Statistics , “2016 Micro, Small and Medium Enterprises (MSME) Survey Basic Report” (KNBS, March 2017).

  3. Savita Bailur et al., “Digital Lives in Ghana, Kenya and Uganda” (Caribou Digital, December 2015).

  4. Marissa Dean and Jonathan Donner, “How Can Platforms Improve Financial Inclusion in Africa?,” FiDA Medium Channel, September 10, 2018.