So often when I work on development projects data is in the centre of the discussions. ‘We do not really know this’, ‘We have no real evidence for this or that’. Most of the time, however, I have difficulties imagining what data looks like and what it actually can do. Millennia ago Star Trek found a solution. Here he is:
He is quite smart. Dressed impeccably. Most of all, he has this understanding look – almost human. According to Wikipedia he even became a sex symbol.
Data is good, solves any problem, draws logical conclusions and he’s infinitely faster than present day computers, seems almost right from the beginning, until Captain Kirk decides otherwise….
Although we often feel we do not have the data we need, we want a lot of it. It helps us come to terms with our actions, logically reason how our actions will lead to results, and with hindsight explain why things worked, or not.
Until not so long ago, data needed to be collected, analysed and subsequently translated into policy recommendations – the traditional time consuming and often tedious process.
We design our logical framework, identify the SMART indicators – thus measurable – and we compile the data. But we ran (and run) into many difficulties. Often data is old, it takes time and often serious amounts of money to compile, validate, analyse and draw conclusions. So much time that, by the time conclusions are drawn, development has moved on to something different then what we were analysing, or the project for which we needed the data has already closed.
More worryingly, as complexity theory gains increasing recognition in the development sector we start questioning whether knowledge about the past (what data essentially is, but not only….) is really useful to manage our actions for future results (hindsight does not necessarily lead to foresight). This is no minor issue if we want to make forward looking policies that are evidence based.
Really? Not so fast.
Past data are still relevant in a complex world
First of all, not all development challenges are complex. Some are even simple, others complicated (in terms of David Snowden’s Cynefin model). But if simple and complicated development challenges exist, then linear development also exists, and knowledge about the past does say something about the future.
In these situations, data making sense of the past is of course useful, and we will not be able to run away from the need for it, and we will not be able to run away from the need to build capacities of those who collect and use data.
The issue becomes a bit different in ‘complex’ development situations. Here, extrapolating trends from the past to the future is hazardous. But that does not mean it is useless. I still fail to understand how one could act without having an idea how it will produce results that contribute to a certain objective. In whatever way we formulate this, it is and remains a linear thinking process: A leads to B leads to C; also in complex development situations.
So here, data in the traditional sense are needed to make projections for the impact of the action to be undertaken. I cannot see how anybody would buy into an action without a reasonably logical development story. And I cannot understand how one can make a scenario (or development story) without a good deal of historic information.
What also does not change, is that we need to account for the way in which we act (and use money). Complex systems are difficult to understand and future developments cannot really be predicted reliably. But by studying what happened, an understanding can be built about just that; what happened.
That means that we can report about the past, and account for our actions. Also in complex systems we will continue to need data in the traditional sense.
Making sense of changes and adapting for the present, real-time
What does change is that we are much more aware that our development story is a theory that may or may not come true.
We are also aware that changes in the development context can happen very quickly. And here a new concept of data comes in: real-time data. In an increasingly complex world, we need ever faster ways of getting to terms with that world to determine our line of action. We need to be able to see in ‘real-time’, whether or not our action has the desired result. And if not, we need to be able to find those actions that will have the desired result, and we need to find them fast. OECD makes a commendable effort.
These kind of quickly updated statistics are a very good step in the right direction, but are not enough. We need to be able to select the data we need.
However, complexity theory also means that small changes in the system can bring huge impact. How then do you find these small changes, if you don’t really know what you are looking for? What we need is a system scan that would identify small changes in the making early enough to look closer at them and decide whether or not they are likely to have big impact and respond accordingly.
We need a methodology that helps us to make choices using data. Two different approaches can help us in doing so: collecting narratives and big data analysis.
Dave Snowden’s micronarratives approach is designed to detect patterns from hundreds, even thousands of stories. We are about to start testing it on the ground so we will be able to report on this soon.
Big data is also an area we are looking into with great interest, together with many other development organizations. (See: Big Data for Development: Opportunities and Challenges)
These technological solutions do help us to get the things we need, but we still need to identify what we need and how we are going to use it. Data, in the end, is just Data. Smart, well dressed. Sexy maybe for some. Simply one source of information to help Captain Kirk’s decision-making.