Send me a message

Daniel Hulme
Daniel Hulme

Data-driven decisions: First steps toward business intelligence

Posted on 3 June 2016

In a previous article, I introduced the hierarchy that exists within data science, illustrated through the DIKUW pyramid. For the purpose of the digestibility of this article, we are going to summarise this into Data > Insight > Action > Adapt, let’s label it the ‘decision cycle’. Now let’s explore how this really looks for businesses and begin to inspire thought about the steps towards a true data-driven culture.

Humans alone aren’t always right

The likes of Nobel laureate Daniel Kahneman, psychologist Dan Ariely and behavioural economist Richard Thaler, to name a few, have been conducting and publishing research regarding shortcomings of human decision-making collectively for decades. It is undeniable, even highly skilled and intelligent humans suffer huge cognitive bias and thus make irrational decisions in many business situations, even critical ones. Research has shown, that a criminal at a parole hearing is up to 6 times more likely to be released if said convict is one of the first three prisoners considered versus the last three prisoners considered. Admittedly, I am fascinated by the irrationalities of human decision making and I intend to discuss more of these in subsequent articles.

Gather the right business data and insights

To be clear, I am definitely not making a case for taking all decision-making away from humans, merely suggesting some prime first steps toward better business decision making, better KPI performance and happier customers. Back to our decision cycle, all businesses will in some way be mining data, sometimes unknowingly, from a vast number of touchpoints, both internal and external. Have another go at really identifying and analysing your existing business data, there is likely be some untapped value.

It is important to make sure how best to visualise any insights

Most businesses will also be attempting to create quality insights from analysis on which decision makers may, or may not be, basing their decisions. Note, there is a distinct difference between reporting and analysis. The quality of these insights, in various forms, depend on the availability, quality and relevance of the data being stored so it is first vital to decide which metrics really matter most to your objectives and the bottom line, rather than just mining any data that you can get your hands on. Note that this data is not just internal. Secondary data sources are key to providing context to the data you hold yourself. It is then important to make sure how best to visualise any insights – learn more here.

Take your analytics further

The next step for most businesses would be to utilise real-time data, stepping away from static analyses and utilising powerful predictive analytics methods. Cloud-based solutions and distributed computing techniques are allowing most businesses to query datasets of all sizes at a relatively low cost. This should allow proactive actions to be taken on multi-dimensional analyses and predictions. Businesses utilising this level of analytics will have their decision-making processes supplemented by notifications backed up by machine learning techniques in order to provide solid predictions of the future.

Rounding off the decision cycle, it is crucial for decision makers to adapt to the outcome of actions. Humans can find adapting particularly difficult where there is a lack of fast accurate feedback that’s readily available, so expect to see the most innovative businesses creating competitive advantages through optimisation and artificial intelligence techniques which can react to live outcomes and adapt future actions accordingly.

Suggested initial steps

Pick data that is quick and easy to collect and metrics important to your business objectives.
Identify existing external and internal data streams relevant to your business.
Connect critical data sources and share insights effectively with quality visuals where possible to allow for effective human action.
Explore predictive analytics tools to allow proactive adjustments and business decisions.
Remove potentially dangerous biases in decision-making through having a diverse decision-making team and focusing on the facts, evidence and real-time insights (if available).

Related topics: business intelligence

Something to add?

Let's take the discussion offline

Ed Riddle

Ed Riddle

Start a conversation

Send me a message

Ed Riddle
Ed Riddle

Message sent

Thank you for contacting us, we will come back to you as soon as possible