The secret of data analytics projects that succeed

  • Richard Thelwell
  • July 15, 2015

data analytics project successIf you’re thinking of launching some kind of data analytics project, you’re far from alone. Call it what you will—Business Intelligence, data mining, data science or anything else—right now, data analytics is hot. Nor is it difficult to see why: some very, very compelling use cases have emerged.

Simply put, data analytics can unlock the door to a more profitable, nimbler, smarter business. A business better able to achieve higher revenues from its existing customers, better able to precisely target new customers, and better able to identify and bring to market compelling new customer propositions.

There’s only one problem. Data analytics projects aren’t risk-free. Or, put another way, it’s not difficult to discover businesses where an investment in a data analytics or Business Intelligence capability doesn’t seem to have delivered much that is of lasting tangible value.

And yet, even in retrospect, those same businesses seem to have followed a sensible project plan, completing it on-time and on-budget.

So how can your data analytics project avoid this fate?

It’s not just about execution

Stripped to the bones, most business projects follow a familiar pattern. Something like: Here’s what we’re going to do, or build, or implement—and here are the resources, and a project plan. So now let’s focus on the task, and complete the project.

And to be sure, you can indeed go about a Business Intelligence project in precisely that manner.

But if so, you’re at risk of winding up with exactly that: an implemented Business Intelligence system.

Which isn’t the same as having a compelling set of answers, insights, and actionable information.

Imagine the ‘To Be’

A better approach is to borrow a trick from the early days of Business Process Reengineering, and focus on some vision of how you’d like the business to be working. In Business Process Reengineering terms, in other words, the ‘To Be’ state, as opposed to the ‘As Is’ state.

data analytics projects to be
Data Analytics projects should focus on what is ‘to be’

And the beauty of doing it that way is that you’re more likely to frame your end state in terms of business outcomes, rather than boxes ticked.

A proven ability to mine customer data for hidden sales relationships, for instance. An ability to put computer tablets in the hands of managers on the factory floor, complete with ‘drill down’ dashboards. An ability to cut the end-of-month and end-of-quarter reporting cycle to (say) two days. Or an ability to pull together data from widely disparate systems, and see a complete picture of what is going on in the business.

Where’s the beef?

It’s not necessary to be a genius in order to see the difficulty with this approach, though.

However desirable such outcomes are—and for most businesses, they are highly desirable—they’re not particularly easy to sell to a board of directors. A board of directors that, needless to say, is accustomed to thinking in terms of cold, hard, ROI numbers.

The problem? Outcomes expressed as ‘end states’ aren’t the same as solid, hard results—from which those cold, hard ROI numbers can be derived.

And yet, who would commit to the financial return from a prospective data mining exercise, before that data mining exercise had been carried out, and before any inkling had been arrived at as to what nuggets it might uncover?

One step at a time

So how to square this particular circle? Actually, it’s not difficult.

Because unlike with vast, traditional on-premise Business Intelligence systems, modern solutions—delivered as a service from the Cloud—can enable end-state outcomes to be achieved in manageable, bite-sized chunks. Mini-projects, in short, that are unlikely to alarm any board of directors.

cloud data analytics projects steps
Cloud data analytics projects offer a more manageable, step-by-step approach

And which, usefully, can be built upon and extended, as confidence grows, and ROI numbers start to emerge.

With data analytics software that’s subscribed to on a monthly basis, for instance, there’s no upfront capital expenditure hit arising from an outright purchase of software licences. And with data analytics software that runs in the Cloud, there’s no need for pricey infrastructure and server upgrades.

The result: data analytics projects that deliver—and that deliver capabilities that are genuinely of lasting value to the business.

If you’re currently evaluating a data analytics project, you might find our bumper guide to Business Intelligence useful