We use cookies. You have options. Cookies help us keep the site running smoothly and inform some of our advertising, but if you’d like to make adjustments, you can visit our Cookie Notice page for more information.
We’d like to use cookies on your device. Cookies help us keep the site running smoothly and inform some of our advertising, but how we use them is entirely up to you. Accept our recommended settings or customise them to your wishes.

Using unstructured data analytics to make better business decisions

There are two types of data – structured and unstructured, says Merkle. Structured data is any data that resides in a fixed field within a record or file. Think spreadsheets or database. Unstructured data is everything else. Which do you need for better business decision making?

The finance director loves structured data because it’s predictable and she doesn’t want any surprises. Whereas over in marketing, the data landscape, the information coming into that part of the business, is changing almost by the hour. New sources of data, new social media channels, spring up all the time. So the ideal solution for most businesses is to have a combination of structured and unstructured data.

What makes the structured world such a nice place?

The fact that you can run reports because everything is structured. You can build a report, and tell it to ‘get that data from over there’, and this data by ‘adding up these columns over here’, to show that view of the world. You can run that report every day and find out how your business is doing. The key thing is that you find out how it was doing an hour ago, yesterday or whenever.

This backwards-looking view is what we call descriptive analytics.

The future power of unstructured data

A data scientist only uses backward data if it’s going to help inform what’s happening right now and going forward. This person is forward-thinking, and is trying to understand the past in order to model the future; that’s where you get the predictive, forward-looking capability of analytics.

Therefore, structured data doesn’t necessarily work for data scientists or analysts. They need to be able to see all the data.

Historical data and the untold story

It’s a bit like looking at the history of the world – someone has already decided what’s important. They’ve thrown away everything else and only chosen to keep what they consider to be memorable or significant.

A data scientist can’t be sure that the stuff that is thrown away isn’t significant. In a marketing context, perhaps the discarded data is having some influence on how people are behaving.

Therefore you need data structures and data capabilities to allow data scientists and analysts to see everything, to process the data and create a model that describes the operation of the business.

Which could be something as simple as product sales – can we build a model, the shape of which when laid out over time, correctly maps to the sales we saw in a particular product line, for example.

But that’s another story, for another day.