Who is responsible for success of decision making with data

Human decisions are emotional. Thresholds soft. (Dare I say…) Gut feelings key.

In this insightful post by Cassie Kozyrkov, Chief Decision Intelligence Engineer, Google, she outlines some very common realities of using data in decision-making.

“For a decision to be data-driven, it has to be the data — as opposed to something else entirely — that drive it. Seems so straightforward, and yet it’s so rare in practice because decision-makers lack a key psychological habit.”

Due to confirmation bias, she feels decisions are often simply “data-inspired” rather than data-driven.

With subsiquent data analysis being interpreted to previously hoped for outcomes.

“Many people only use data to feel better about decisions they’ve already made”

Her reasoning for this is (essentially) lack of process.

Peoples decision making processes often don’t start with a proper hypothesis.

Say, something along the lines of “This has happened. We suspect that was the cause. To decide we’ll look at x,y,z data. If the data says this is then we’ll do that, if it doesn’t we won’t do that.”

“The antidote is setting your decision criteria in advance.”

So this is setting up an experiment, if you like, for every decision you make.

The hypothesis approach is a good one. It is standard scientific practise. Proven. It isn’t well used and it is definitely a good place to start.

There are more issues than simply process though.

It is the willingness of individuals to change behaviour, to have the confidence to re-evaluate their organisational contribution, and to start to (want to, even consider) understanding the issues around analysing data … these are the challenges I see everyday.

The data science/analytics/engineering community take comfort in the numbers.

We didn’t drop Maths/Stats/Mechanics at our first opportunity at school.

It comes easy.

We (largely) understand our data limitations but what this article says to me is that many are also really frustrated by the fact their data models aren’t more influential.

My view is that, yes, decision makers need to be “trained” but it is our role to do that.

As a technical community we have to take more responsibility and do a (much) better (and more sensitive) sales job on all this.

Especially, in industries where knowing about numbers, maths and “data” hasn’t traditionally been a core competence.

It is a thought provoking article that highlights many explicit and implied realities of trying to effectively integrate data into decision making.

Technical people, if you don’t want to be frustrated, you need to get better at selling your ideas and potential, rather than blaming other side.

It has been well proven people are not rational decision-makers.

Don’t expect otherwise.

People make emotional decisions.

It is not about the data, its about people.