Data-Driven Learning and Firm Performance: Experimental Evidence from SMEs in Kenya

Entrepreneurs benefit from information about the demand for their products, customers, and their business performance. However, many entrepreneurs in developing countries under-invest in collecting and learning from such information even when this information is readily available. Several studies have documented that businesses in developing countries do not keep records of sales and inventories even though doing so would help them better understand the relationship between their activities and performance, and design better business strategies. Moreover, standard training interventions that nudge businesses to keep these records have typically yielded limited success, especially in the long term. This project seeks to answer i) whether reducing the cost of identifying and digesting patterns in sales and inventory data improves learning and performance among SMEs in LICs, and ii) whether this increases record-keeping among SMEs in LICs.

The researchers will run a field experiment among retail small and medium enterprises (SMEs) in Nairobi and Mombasa, Kenya. They will roll out a point-of-sales (POS) app developed by the partner Technoserve that lets a user enter information about their inventory, (e.g., the level of stock, purchase price, and selling price), tracks changes in the inventory as the user updates the list with sales data, alerts the user of low inventories, and enables direct ordering of new stock. In the basic version of the app, the user can view reports of the transactions in a spreadsheet format. In the experiment, the researchers will vary 1) whether entrepreneurs receive the basic version with no automated analyses, 2) whether they receive additional simple-to-understand analyses of the data they enter, or 3) whether they receive simple-to-understand analyses of the data they enter plus information about comparison firm activities and performance. The RCT will investigate the impacts of providing firms with summarised information about relationships between their recorded activities and their performance.

The findings from this study can inform the private sector policies of Kenya and many other low- and middle-income countries, especially East African countries with a similar selection of entrepreneurs (e.g., Tanzania). In particular, the lessons from the study will be directly relevant for SMEs, which employ 81-91% of labour in developing countries. Moreover, the extent to which digital technology can reduce barriers to learning and foster SME growth has direct implications for private sector policy and for firm strategy in developing countries.


Anik Ashraf

University of Munich

Elizabeth Lyons

University of California San Diego