Data driven decision making and the HiPPO in the room

Despite paying lip service to DDDM many executives would still rather rely on their own gut instinct

Numerology, entrails, tea leaves, astrology, Tarot cards and mystic seers are all used to support decision making. Do they work? Well, probably best not to tempt fate by casting aspersions, but when subjected to scientific analysis their record does tend to be a little shaky.

The information age has a new method - data-driven decision making (DBDM). In theory, this should be a lot more rigorous as it is based in the science of data modelling and statistics, but in practice, it can be just as unreliable as gazing into a crystal ball.

However, while the mechanics of a crystal ball divination will always be a mystery, at least something can be done about the data and its supporting infrastructure in order to improve its accuracy.

The HiPPO effect

Rather than theomancy or water witchery, the main rival to DDDM is intuition or gut instinct. The role of intuition in business decision making is worthy of debate. After all even good data has to be interpreted, and the quality of decisions is still ultimately dependent on the skills of the interpreter. However, it's easy to see how instinct can be given undue prominence in decision making.

It's much less arduous for a confident executive to make a gut decision without trawling through reams of data Even if they are inclined to take cues from the data they probably feel their own intuition and experience is just as valid, if not more so, as patterns revealed by analytics. Perhaps the data is incomplete, or maybe certain nuances haven't been captured. It's easy to find a reason to overrule what the data seems to be saying, indeed it's pretty much human instinct to do so, particularly if the picture is unfavourable or unclear.

The result known in management circles as the HiPPO effect for Highest Paid Person in the Organisation. The scenario will be familiar to all even if the acronym is not. If good data is not available to inform decision making, HiPPOs step in and because of their seniority, dissenting voices tend not be heard. The problem is that HiPPOs don't always make the right
decisions - and if they get it wrong the chances of the error being rolled back is small because those who lack their power and authority find it tricky to speak up. The problem is cultural but has a technical cause at its root - lack of access to data.

This problem has two strands: actual availability of relevant data and the timeliness of its delivery.

Timeliness and quality

Data needs to be of good quality and readily available. Crucially too it needs to be in the right format so that it is actually useful and accessible to its consumer. This can be a challenge when it is derived from multiple sources. As more data comes from more external sources such as open datasets and sensors, combining it with internal data to create a meaningful output is one of the great analytics challenges.

Ensuring that internally generated data is clean is a matter of good data governance which has been a real focus for many organisations in recent years. The veracity of external data is another matter entirely but eliminating this kind of data from analysis will mean that many decisions, if they involve external factors, will be less data-driven than they want them to be. Companies that are serious about DDDM need to focus on the integration side.

DDDM is the corporate equivalent of evidence-based policy making in government. Almost everyone agrees that it's a good idea but it takes a long time to actually bring about. Changing culture around decision making is a long game, but at least organisations can do their best to get their data house in order. In the long run it's either that or back to the tea leaves.

DDDM and challenging the HiPPO effect are covered in a recent Computing white paper which is now available for download.