Our team here at Algorhythm has collectively struggled to help nonprofit organizations, philanthropies and for-profit social impact companies effectively use data to strengthen their social change strategies. We are a group of researchers, evaluators, data scientists and impact business leaders who have spent more than a combined 100 years studying and evaluating social issues and programs. We have seen how effective data help leaders learn about if and how they have achieved success. We know that evaluation, when done well, empowers leaders to plan better for the future, and funders and impact investors to make better decisions that lead to positive change.

But, here’s where our struggle comes in. Many of us are also social workers, clinicians and practitioners, with “front line,” hands-on experience. It is here where we have long wondered if we are really maximizing the difference we can make with what we know how to do. We know evaluation is an effective tool for planning, program design and accountability. But, could we use our tools to take it even further – to the front line, where day-by-day, case-by-case decisions actually get made, and where social impact really begins?

We believe that the answer is a resounding, YES, and here’s how. Over the past few years we have raised the bar by honing one of the most underutilized, yet most critical, tools in our tool box – analytics. And, by adding next-generation data analytics – like machine learning and artificial intelligence algorithms – to our existing social science statistics toolbox, we have found ways that we can make a real difference for the direct decision-maker, while still meeting the evaluative learning needs of the big picture leaders, designers and investors in social change.

What we have discovered is that we can add much more to the needed “rearview” descriptive analytics, where findings summarize and conclude what happened in the past. We have added the use of next generation analytics to guide decisions and solutions, at the moment when change agents need to act. Specifically, we are augmenting the social scientists’ tools of descriptive analytics with the data scientists’ tools of predictive and prescriptive analytics. The result: through combining these three phases of analytics, we are finally closing the learning loop, where the lessons from evaluation data get to be applied, in real time and prospectively, by the hands-on change agents, which in turn leads to the measurable advancement and expansion of impact.

We are calling these next-generation analytics, in service to social change, “impact analytics.” Impact analytics are ground up rather than top down. The analytics work begins in service to the hands-on change agent, using algorithmic predictive analytics to build models that determine the likelihood of an outcome, followed by prescriptive analytics to build models that identify all of the solution options to maximize success. The descriptive analytics work is in service to leaders and managers, creating a birds-eye view of what happened, what significant results or outcomes were achieved, for whom, and how well the whole group of change agents did to help people find their pathways to success. Lastly, we believe, that as data grow, impact analytics should be re-applied order to evolve existing predictive and prescriptive models to insure that solution options keep up with the very real changes that are happening in the lives and environments of those being served.

Impact analytics could be the key to bringing social science and evaluation research to the front end, significantly improving the likelihood of success, one situation at a time.