Assessing Cost Effectiveness of Digital Development Projects

There is scant data on the effectiveness of different digital development projects that seek to connect and serve underserved communities. This stymies data-driven decision-making: absence of comparative cost-effectiveness analysis hampers key funding decisions by grant-making organizations and local governments in developing country environments, and can impact right-sizing of funding opportunities to project needs. This remains as true of demand-side initiatives that seek to deploy digital skills training, or supply-side initiatives using a myriad of technologies to connect the unconnected. This paper makes headway on this problem by providing costs, reach, and simple metrics of cost-effectiveness in the form of cost/beneficiary/year incurred by different digital development projects. Using a first-of-its-kind dataset hand collected through in-depth interviews conducted with project practitioners, we report on the costs and reach across different demand- and supply-side initiatives seeking to connect the unconnected. Our contribution is unique, in that it seeks to collect and compare different connectivity initiatives across a range of domains: education, health, and agriculture; as well as standalone grassroots connectivity initiatives in the form of community networks and small scale rural ISP deployments. The contribution of this paper is two-fold: first, we report on original costs and reach data on several projects through in-depth interviews, providing costs for a field where even order of magnitude estimates are hard to procure. Second, as changes to underlying assumptions can widely vary cost-effectiveness estimates, and these assumptions are rarely discussed in the literature for their relative strengths and weaknesses, this paper seeks to also report on a meta- analysis of cost effectiveness measures as used in various literatures in development and healthcare particularly, and use those learnings for developing cost effectiveness measures for digital development initiatives. We report costs of projects in terms of fixed and operational costs, as well as development, management, and implementation costs in different development domains. We then provide purchasing-power-parity and real-dollar estimates of completed and ongoing projects by per beneficiary and per year costs. In instances where revenue estimates are available, we compute return on investment and time to break even using different discounting assumptions. We additionally conduct sensitivity analysis on our assumptions for discount rates for ongoing projects. A key finding in collection of this dataset has been that 2/3rd of the projects that we study do not have a revenue model, when both supply-side and demand-side projects are taken together for analysis.