Accurate measurement of stock levels, turnover, and profitability in microenterprises in developing countries is difficult due to the fact that the majority of these firms do not keep detailed records. We test the use of radio frequency identification (RFID) tags as a means of objectively measuring stock levels and stock flow in small retail firms in Sri Lanka. In principle this offers the potential to track stock movements accurately. We compare the stock counts obtained from RFID reads to physical stock counts and to survey responses. We have three main findings. First, current RFID-technology is more difficult to use, and more time-consuming to employ, than we envisaged. Second, the technology works reasonably well for paper products, but very poorly for most products sold by microenterprises: on average we were able to read only about one-quarter of the products tagged, and there was considerable day-to-day variation in read-efficiency. This results from technical issues arising from read efficiency being comprised by liquids, metal, and product stacking. Third, a comparison of survey responses and physical stock-takes shows much higher accuracy for survey measures than RFID. As a result, we conclude that this technology is currently unsuitable for improving stock measurement in microenterprises, except perhaps for a few products.