Making Data Driven Decisions

Data Driven. A term we hear so often in both our professional and our personal lives. But, how can you be sure you are truly making effective data driven decisions?

Data Driven. A term we hear so often in both our professional and our personal lives. Throughout the pandemic we’ve all heard politicians, friends and colleagues proclaim we should ‘follow the science’ and ‘trust the data’ — but what does that actually mean? And, how can you be sure you are truly making effective data driven decisions?

Gartner predicted by the end of 2022 90% of companies will cite information as a key resource, and data literacy an explicit and necessary driver of business value — with that, this post can get you started!

Let’s get started with some of the foundations:

• What is a data driven decision, and what are the benefits?
• Influence, and its impact
• Begin with a hypothesis
• Defining your actions

What does it mean, and why is it important?

Let’s start with what a Data Driven Decision is not — making such a decision is not based on a technology, it’s not a few reports or charts showing last week’s performance, and it’s not something that you can just proclaim you want to do — and expect it to happen without a second thought.

A data driven decision is an objective, reasoned and explainable decision. It is protected against your natural bias and ultimately leads to higher quality decisions and insights.

Determined by or dependent on the collection or analysis of data.

– Oxford Languages

Or: a data inspired decision is much quicker and easier, but your natural bias runs the show meaning the value and quality of the decision is diminished.

To completely let data drive your decisions and actions, it has to be interpreted free of any influence, bias or predetermination. But why is that important?

The impact of influence

Influence is an incredible thing. By our very nature we always look for confirmation of our own beliefs, opinions and suspicions. A common pitfall we see in businesses is the desire to use data to support decisions already made, or to use more data to support a decision or insight that we hold in esteem from years gone past.

Being Data Driven means that you have to commit to an opinion, belief or suspicion before you have seen the data and in the absence of influence — you need to write down your hypothesis.

Writing a strong hypothesis

A hypothesis, and your ability to test it, is fundamental to a data driven decision. There is plenty of information available online on writing effective hypotheses, so I’m not going to labour the point too much here — but as a quick guide, I use the following:

Describe the result (effect), and the variables (or causes)

A strong hypothesis should include the result you are trying to explain, and the variables or causes of that effect you are wanting to understand.

For example:

• Bad weather (cause), reduces footfall (effect) in our stores.

Make sure it’s testable

It’s no good penning the most grandiose of hypotheses to find that the data you need to test it isn’t available.

The whole purpose of writing a strong hypothesis as a prelude to a data driven decision is that you want to be able to trust the result, and take action.

Keep it simple!

Sometimes it’s fine, even encouraged, to start small. Taking my eight-year-old son as inspiration — there is nothing wrong with asking ‘why?’ again…and again…and again. Build the complexity, make sure you can be completely confident in the result.

Actions, Boundaries and Confidence

Defining your actions is a key step in the process. The commitment to the decision you will take, before seeing the results, allows you to think objectively and pragmatically. If done after, it’s far too easy to allow yourself to become data inspired, when you want to be data driven.

The Default Action

You should always start by defining a Default Action. This is the action taken (or not taken) in the absence of data, or enough confidence in the results. It’s the time to be really open with yourself and stakeholders about the sometimes disappointing reality — a hypothesis is often wrong, and that is not a bad thing. You have committed to letting data drive your decision, don’t falter now, make the effort to truly ask yourself what will do by default!

Rejecting the hypothesis — because the data didn’t support it — is not a wasted effort. It’s an opportunity to take what you have learned, and go back round the data driven merry-go-round. What is your next hypothesis?

Decision Boundaries

Secondly, and perhaps the more exciting step, would be to establish the boundaries and action matrices — in other words, what would it take for you to move past the default action, and take a different approach.

Again, this doesn’t have to be insanely complicated to start with (we’ll make it more scientific later), it just needs to be honest. Treat this matrix as a commitment, write it in stone, and be sure that the decisions you have committed to can be taken if the data supports it.

Confidence

The last piece of the puzzle we’ll look at in this post is setting your evidential and confidence requirements.

At this point, it’s a good time to start asking yourself some questions:

• How much data do I need? Is volume or completeness more important?
• How would we handle outliers? What is our definition of an outlier?
• Can I document and reproduce the results if needed?

The aim of this process is not only to provide the safety, comfort and a great night’s sleep with knowing that you made a great decision today — it is also a fundamental step that means you can continue to assess and measure the actions again and again.

Wrap Up…

To wrap things up, I hope after reading this quick blog you have a good foundational understanding of what a data driven decision is, and some of the first steps you need to undertake before you even begin to look at the data itself.