By Matthew Robinson

I've always believed that the fundamental task of analytics is to understand the consumer experience such that it can be improved. No counting things, measuring stuff and reporting ad nauseum for the sake of it. Analytics is an enterprise that should be charged with determining the effectiveness of ones digital work, and signposting what can and ought to be changed to make something work, or indeed to make something work even better.

So it was heartening yesterday to see the BBC selling in the latest version of their VOD iPlayer service not only on the basis of its snazzy new look, but in response to how they know people are actually now using iPlayer.

Turns out when iPlayer first launched (back in 2007), people were in the habit of simply finding and playing a specific programme that they wanted to catch up on. In and out, job done. Now though, analytics has revealed that 42% of users are going to iPlayer with nothing much in mind about what they want to watch, and are browsing far more extensively in order to find something that they might be up for viewing.

All of which has seen the BBC move to better surface and visualise what's available, to improve the navigation of all of that choice, to enhance search utility, and to supply more advanced recommendations for what you might want to watch.

In short, that's changes to design and UX not informed on a whim. Not even informed on the basis of just a respected set of opinions. But changes rooted in a precise measure and understanding of what users are wanting and needing from iPlayer.

Now I suppose you could argue that all of the aforementioned changes make sense and sound compelling irrespective of the numbers used to make the case for them. But in a digital world that permits us to understand, in great detail, so much about the user experience, I think the BBC's data-centric approach is one that we'd be advised to adopt as much as we possibly can. Not least because opinion-led changes aren't always a no brainer, and don't always see guesswork rewarded.

And so, what's the best way to become more properly data-centric? To not just pay lip service to the numbers in the reports that your analytics spits out? Well, I’d say the following three actions are crucial…

1 - Setting objectives: all work really ought to be understood on the basis of the digital strategy that’s guiding it and the business goals that it’s trying to meet. If no-one’s really very clear on precisely what it’s hoped will be achieved, it’s virtually impossible to assess effectiveness and to consider potential improvements. Indeed, in last year’s very first online Analytics Academy, Google were at pains to say that defining your overall business objective was always the fundamental first step to take.

2 - Setting KPIs: knowing what you’re shooting for is one thing, but recognising up front what success will look like is an essential next step. Writing a measurement framework that aligns your KPIs to your business goals lies at the very heart of being properly data-centric. Certainly it’s the difference between knowing that a modest response to assorted online tools might be far less of a concern compared with the lack of interest in your lifestyle content.

3 - Seeking out insights: analysis and reporting will always be a much more meaningful pursuit once results data can be set alongside business goals and specific KPIs. Insights don’t magically reveal themselves to those simply staring at the data for long enough. Rather, they’re born of understanding what results will be significant and why. And once identified they can then sit at the heart of ongoing plans for testing and optimisation.

No idea precisely what online measurement process the BBC submit themselves to. But I’d say it’s a good bet the launch of the new iPlayer was born out of a very systematic approach to identifying what’s working. And it’s an approach that each and every business really ought to take on board.