PREMISE.pdf

Premise

MCC16RFA0001

Millennium Challenge Corporation: Data for Local Impact Innovation Challenge Project

Organization Name: Premise Contact: Adrienne Villani, Director of Business Development Contact Details: adrienne@premise.com | +1.914.419.2632

Premise is a data and analytics platform measuring economic, political and social developments in realtime. Using mobile software, Premise generates both macro and micro insights derived from groundsourced data collected by our global network of 30,000 contributors in over 35 countries. The platform enables global organizations to test hypotheses and monitor global trends, empowering decisionmakers to move faster and make more impactful policy, programmatic and investment decisions.

Premise has built a network of data contributors in both Tanzania and Kenya at the local, subnational and national level.

Collectively, the network has collected nearly 300,000 discrete data points in Kenya and Tanzaniaon behalf of clients including World Bank, The Rockefeller Foundation, Stanford University, Standard Chartered Bank, and Evidence Action(see adjacent figure of Premise’s analytics dashboard depicting observations and geographic coverage in Kenya and Tanzania). Our data is collected and analyzed in realtime; each data point possesses an associated image and is timestamped and geolocated.

We actively employ over five hundred local Tanzanians and Kenyans.

Each person works an average of 10 hours per week, earning approximately $50 per week. Our contributor recruitment and onboarding process includes outreach and lead generation (via online and offline channels); onboarding and training; engagement and incentivization; and communication and feedback.

Our technology and existing network of contributors in Tanzania and Kenya have already uncovered previously invisible data gaps. As such, we seek to partner on the Data for Local Impact Innovation Challenge Project for two primary reasons:

1. To bridge the data gap and inform core MCC/PEPFAR programming, as well as support country capacity with highlylocalized, realtime, transparent data

2. To provide economic livelihood opportunities for local Tanzanians and Kenyans Premise’s platform fills a gap in the data toolbox, serving as a nimble, localized, lowcost and easily scalable data collection and measurement strategy. Premise measures:

mailto:adrienne@premise.com

Premise

MCC16RFA0001

Millennium Challenge Corporation: Data for Local Impact Innovation Challenge Project

In our view, data is critical to decision making at three key stages of a successful intervention:

1. Scoping and landscape assessment of an intervention’s feasibility and expected impact

2. Developing operational models as projects scale; data needed to assess ongoing impact and quickly detect anomalies

3. Monitoring and evaluating service delivery as an intervention is in steady state at scale; data to assess usage and ensure quality program implementation.

Premise’s platform is designed to work through these stages, informing research, enabling smarter programmatic decisionmaking and costeffectiveness, and improving accountability.

Our iterative process by which we translate customer queries into tasks, allocate these tasks to our contributors, gather observations, and transform those observations into usable outputs for our customers is as follows:

Survey design: translates highlevel questions into precise survey design and discrete task inventories. As contributors are typically not professional survey enumerators, surveys are designed to pool information across contributors, reducing sampling error and bias.

Task allocation: distributes survey tasks to Premise contributors via our Android app to discover the right set of products, services and venues for data collection

Contributor management:

involves recruitment and maintenance of a sufficient base of contributors, and encompasses Premise’s app, through which contributors receive instructions, complete tasks, communicate with us, and are compensated

Quality control: combines both machine intelligence and human annotators to verify observations as they are submitted, identifying erroneous or fraudulent observations, and surfacing deficiencies in the survey design.

Analytics: transforms clean observations from quality control into usable data streams, derived indicators, alerts and predictive outputs.

Our approach features multiple feedback loops that adapt our survey approach over time to enhance accuracy and explore new hypotheses as they present themselves. This confers key advantages versus traditional survey methodologies:

Roll out survey changes in near realtime (vs. weeks/months for traditional surveys) Ensure consistent coverage across locations and observation types Take feedback from local contributors into account in realtime Task contributors to verify each other’s observations Incentivize contributors as needed with our costefficient dynamic task pricing

As part of the Data for Local Impact Innovation Challenge Project, we will enable organizations to monitor critical issues, providing a crucial advantage in a rapidly changing world: the power to move with greater speed and accuracy and make sounder policy and programmatic decisions.