Action! Why big data analytics projects fail

By Martin Duffy, SAS Ireland

Most people would agree that technology in itself is just a tool; it’s what you do with it that counts. So you’d be amazed by how many companies invest in data analytics ‘solutions’ without first asking themselves what business problems they want to solve, or which actions they want to take.

It’s easy to see how it happens. While the whole company may be using the ‘big data’ buzzword, it’s often left to IT to make sense of what it actually means. So investment is made in big data analytics, analysis is presented to the business, decisions are made and, finally, strategic action is taken.

But this is the wrong approach. As ‘action’ is the only part that makes or saves money, wouldn’t it make sense put that first, not last? Data analytics can only drive business results if the organisation knows what actions it wants to take, and what decisions it will need to make to drive those actions.

For example, let’s suppose your company wants to reduce churn. Instead of analysing the data blindly, you could ask yourself or the relevant person in the business:

Q: What action do I want to take?
A: Incentivise customers to stay.

Q: What decisions will I need to make to ensure that action succeeds?
A: Who needs to be incentivised? What will influence them? How much should it cost? What channel will be most effective? How long do we expect them to stay?

At this stage, you should have the foundations of a successful, business-centric analytics project. And if you haven't, well it’s only cost you a couple of days’ thought – better to drop the project now before it costs serious time and money.

This sounds obvious, right? You’d be surprised. Let me give you a typical example. Not so long ago, a leading internet provider launched a month-long free trial offer. We built them a model to accurately predict, within the first week of the trial, which users were most likely to convert and which needed a push, so they could follow up with further incentives. Unfortunately, they didn’t foresee that it would take them more than 5 weeks to prepare the second wave of the campaign. So all those valuable customers had finished their trials, would be harder to convert, and all the work was for nothing.

On the other hand, we work with a Belgian bank that uses big data analytics to help them execute around 3,000 tailored campaigns a year; that’s a new campaign every 40 minutes of the working day. And help a Danish bank execute event-triggered campaigns within 3 minutes of a customer’s transaction. Both can only do this because they built their data analytics tools with actions, not just data, in mind.

So, over the next few blogs, I’ll be sharing some insight and tips on how you can:

• Prioritise desired actions and required decisions when building your data analytics framework
• Ensure you’re gathering the right data to inform those decisions

Follow SAS Ireland on Twitter to get the next instalment. In the meantime, find out what IDC says are the 4 myths that must be understood before you can truly grasp the big data analytics opportunity.