When does data help?
There are situations where data is useful, and situations where it does not add much value to a decision making process (and may even get in the way).
Situations where data is helpful:
Concealed - We do not need data to tell us if we are cooking good scrambled eggs. We can just look at the eggs and see. We do need data when we are cooking a souffle. We cannot go in and look for ourselves (the souffle will collapse if we open the oven door!), so we must rely on data as a proxy (oven temperature, timing, etc). It is very useful to have data when the thing we are dealing with is concealed (like a souffle) - that is it cannot straightforwardly be directly observed (like scrambled eggs).
Counter-intuitive - When we see a traffic jam we assume the problem is cars moving too slowly. And yet, actually many traffic problems can be solved with lower speed limits, leading to greater throughput of traffic. This is counter-intuitive but we can run an experiment and measure the outcome to challenge our intuition. Data is useful when something counter-intuitive is going on.
In other situations, we live in a very data oriented culture but we don't often stop to think "what problem are we trying to solve by having data" - IE we do not have a clear hypothesis about how having "more data" on a particular issue will lead to an improvement in our ability to make a decision. In other words, we tend to look for 'more data' without defining a way that we will check if having more data helped us. Which means we do not collect data on whether 'more data' helped us or not. We just assume more data must be better without testing it. Which is not actually a very data-driven approach!
In fact, a lot of data gathering is wasteful. We put resources into building and maintaining databases against hypothetical future needs - instead of focusing on current, actually existing needs.
Often what we need in organisations is not an All Powerful Database that can answer any query on the basis of what happened in the past. Instead we need to take a more empirical orientation. In other words, to see a problem, and to find out what the solution is by trying something and then by gathering data about whether it worked or not, and then using this to course correct. (This is the PCDA cycle - plan, do, check, act).
