Experimental Design
We will conduct a randomized controlled trial with 1,024 women entrepreneurs in Nairobi. Eligible participants will be self-identified women, adults aged 18 and above, who own and operate a small or medium enterprise and own a smartphone.
We will employ a street-level recruitment strategy to directly engage entrepreneurs. Trained enumerators will utilize standardized scripts to assess interest in information regarding loan access and application support while simultaneously collecting contact information from potential participants. This approach is designed to capture a diverse sample, including entrepreneurs who may not currently engage with formal loan systems, thereby enhancing the study's external validity. Digital invitations for study participation will subsequently be sent to interested entrepreneurs.
Following baseline data collection, participants will be randomly assigned to one of three treatment arms or a control group, with 256 participants in each arm. The interventions will be digitally administered for scalability via WhatsApp, with minimal or no in-person component depending on findings from pilot testing.
Treatment Arm 1 (T1): Personalized Loan Recommendation Tool (N=256)
Women entrepreneurs in this group will receive personalized loan recommendations through a gender-neutral algorithm trained on data from both male and female entrepreneurs. The research team has already collected preliminary data to build an initial version of this recommendation tool. However, additional data will be collected to further train and refine the algorithm, with support from local financial institutions to ensure comprehensive coverage of borrower profiles and lending patterns.
The digital interface will allow entrepreneurs to input their key business data—including industry type, business type, size, sales and revenue history, personal credit score, inventory value, and working capital. The tool will then generate recommendations showing what an entrepreneur with similar characteristics typically borrows and the associated objective risk of default. By providing this data-driven borrowing guidance calibrated to the entrepreneur's cashflow and business profile, this approach will serve as an "anchor," potentially encouraging women to consider loan amounts that align with their business needs and creditworthiness.
The recommendation tool will be developed as a web application by a local developer, who will be hired and managed by the Principal Investigator. The research team will be responsible for training the machine learning model on the collected data, while the developer will program the tool's interface and functionality based on the team's specifications.
Treatment Arm 2 (T2): Interactive Digital Business Planning (N=256)
Women entrepreneurs in this treatment group will be introduced to an interactive business planning tool designed to assist them in strategically thinking about expanding their business. The tool will offer a structured series of prompts and questions to encourage critical thinking about optimal fund utilization for growth, including investments in inventory, marketing, and operational efficiencies. Through active engagement with this tool, entrepreneurs will gain clarity about their potential business strategies and opportunities, reducing uncertainty about how to effectively use loan funds.
Treatment Arm 3 (T3): Digital Training on Consumer Rights and Redress Mechanisms (N=256)
Women entrepreneurs in this treatment group will be introduced to an interactive digital module on redress mechanisms that addresses: (i) illegality of seizure of property in the case of unsecured loans; (ii) informing consumers of exactly what property can be collected in the case of secured loans; (iii) informing consumers of what constitutes harassment in the repayment process; (iv) informing consumers of complaint procedures in case they face any of the above issues. This intervention directly addresses fears about post-contract exploitation and harassment that our exploratory work identified as significant barriers.
Control Group (N=256)
Women entrepreneurs in the control group will receive business-as-usual marketing and outreach materials from partner financial institutions, with no additional interventions.
Baseline Survey
We will conduct baseline data collection through in-person firm surveys with all 1,024 participants prior to the interventions. Each survey will last approximately 45 minutes and collect quantitative data on demographics, business characteristics (including sales, profits, and products), risk preferences, baseline credit demand, incidents of post-contract harassment, and trust in financial service providers. Respondents will be compensated with a gift worth $5 for their time after completing the baseline survey.
Endline Survey
Endline data collection will occur after the interventions through in-person firm surveys with the same respondents from baseline. Each survey will last approximately 1.5 hours and collect quantitative data on business characteristics (including sales, profits, and products), credit demand—specifically willingness to apply and application amount measured through incentive-compatible methods—incidents of post-contract harassment, trust in financial service providers, business performance, and uses of credit. The endline survey will serve as the key data collection mechanism for measuring outcome variables. Respondents will be compensated with a gift worth $5 for their time after completing the endline survey.