Photo: Jeff Walker/CIFOR
Adaptive management of humanitarian responses is so important when we work in today’s complex situations. This month I wanted to pick out some examples that give us inspiration how to do that.
Our monitoring in humanitarian response is a bit stuck. We are under pressure to demonstrate reach, and at some point in the process we need to provide summative evidence for accountability. For this we really need consistent indicators (and as an evaluator I always struggled when indicators had been changed several times during a response). But this way we often miss the changing realities facing the response, and we end up not adapting. The recent ALNAP study “Shifting Mindsets” makes the case for flexible monitoring systems that go beyond accountability and capture quality and shifting response realities in much more compelling ways.
A case in point for more adaptive management is the use of unstructured community feedback in a structured system to inform a better quality Ebola response in the DRC, a collaboration between the IFRC and the Centre for Humanitarian Data. This per se is not traditional monitoring, but it allows responders to feel the pulse of what works and what doesn’t and adjust. Making such complex and diverse data available in a systematic way is a major step forward in making the response more relevant and appropriate.
On the other side of the data spectrum there is now a lot of interest in predictive models for humanitarian response. Algorithms of course need much more structured data – very different from what the work on perceptions is using. It is fascinating though, and the predictive analytics work stream at the Centre for Humanitarian Data is a prominent example (that also featured in a side event at last month’s UNGA). There surely is lots more to come from that work.
The Shifting Mindsets report with a chapter on flexible monitoring can be found here:
The work of the IFRC and the Centre for Humanitarian Data with perception information in the DRC is featured in this article.
This summary of a panel discussion on predictive analysis nicely summarizes the current state of play