Experimental Design Details
The sample is comprised of 320 women-led microenterprises that are located across the Bangalore, Tumkur, Chikkabalapura, and Ramanagar districts on Karnataka. The design randomizes participants into three groups: one control and two treatment groups corresponding to the three experimental treatments described below. All three groups are offered access to the app. To study whether learning costs drive adoption decisions, I vary the extent to which subsidize learning costs across the three arms.
The control group (C) will be given minimal help in overcoming learning costs. As is common in traditional training sessions, entrepreneurs will be provided guidance on why the MeraBills app is helpful and how to use it. They will be provided assistance on how to download it. This is done to ensure that the control group has basic knowledge of how to enter business transactions on the app. However, the learning costs will not been subsidized for this group.
In addition to what is provided to the control group, treatment group T1 is provided additional hand holding in entering transactions on the app. For this treatment I subsidize the cost of learning by hand holding the entrepreneurs on using the app. As entrepreneurs learn by doing, this reduces the learning cost they will bear while using the app after the training sessions. I hypothesize that firms will be more likely to adopt the digital technology in group T1 as compared to group C, as T1 has been moved further along the learning curve.
For treatment group T2 I intensify the extent to which I subsidize learning costs. For this treatment arm, trained project staff will visit the entrepreneurs on a weekly basis for one month and hand hold them to enter all the transactions on the app. I hypothesize that this group is group is moved further along the learning curve as compared to groups T1 and C.
The sample is selected for this study based on three criteria - they are actively running a business, they are involved in decision making in the business and have regular access to a smartphone. From pilot work in September 2021, I note that it is key to check that the entrepreneurs are currently actively running a business. During the pandemic a lot of entrepreneurs stopped running their business (temporarily or permanently). Further, as I work with women entrepreneurs I check that they are involved in decision making. In LMIC settings it is quite for women to be involved in a family business, but not have much agency in the decision making, Since I want to study decisions around adoption of a new technology, it is important that the women are closely involved in running the business. As for the last criteria, I ensure that women have a smartphone they regularly use. Driven by the low cost of phone data and widespread use of apps like WhatsApp, I find that 25-40\% women in the region have a basic smartphone. However, there in our setting there is a distinction between ``owning" a phone and regularly using a phone. Quite often, smartphones are shared by members of the household. In the pilot, I find that 70-80\% of women who own a smart said that they share the phone with other members. To be able to adopt this digital technology, it is key that the participants have regular access to the phone. Therefore, instead of asking if participants own a phone, I ask whether they have a regular access to a phone.
Based on the criteria discussed above, I select sixteen field officers from Buzz Women who are tasked with selecting 20 entrepreneurs each from their respective panchayats. A panchayat is a geographical area within a district. I ask the field officers to choose the women eligible for this study as they have local knowledge. The survey team will then call the 320 participants and double-check that they meet the criteria for our study. After baseline surveys are conducted, there will be an induction session where entrepreneurs will be informed about the study and their relevant treatments. Following \cite{Drexler2014}, I begin the induction session a rule-of-thumb training that focuses on very simple heuristics or routines for business and financial practices. In line with COVID precautions entrepreneurs will be invited to the induction session in groups of 40.
The randomization will be done after stratifying on district, age, and frequency of business transactions. As I choose firms across four districts, I stratify on this as there is likely to be some regional variation in the firms. I stratify on age as digital literacy and experience using digital technologies varies by age. Younger entrepreneurs are more likely to have used similar technologies in the past. Lastly, firms in the sample vary in their frequency of transactions.