Is the measurement tail wagging the media dog?

It is a truth universally acknowledged that a brand investing money in advertising must be in want of a return (with apologies to Jane Austen!). No advertiser spends just for the sake of it, they do so to deliver business results.

That may be stating the obvious, and it would appear that in today’s media world, measurement of effect is apparently ubiquitous. Proving the link between advertising spend and return is, in theory, more straightforward than ever before, with the arrival of digital marketing opportunities seemingly offering an easy way to establish that link.

But is it really that simple? Have we been seduced by the vast quantities of digital data available that offer a supposed magic bullet when it comes to measurement of effect?

With the data generated from digital marketing, there are almost limitless, instantaneous reports available about the activity that is running, at a highly detailed, site by site and second by second level. However, the ‘outputs’ of these reports aren’t business relevant, only ‘response optimisation’ relevant.

And are we forgetting that there are both direct and indirect influences on sales? That it’s not just about the last action that people took before the purchase they made?

In my view, we’ve gone backwards when it comes to measurement. We seem to have forgotten the complexities of establishing the influence of all different forms of advertising on sales. 

The illusion of precision offered by digital has even led to a belief that media channels that don’t or can’t supply a definitive ‘what happened next’ number are somehow less effective. Yet we’ve been measuring the effectiveness of these channels for years.

With ‘traditional’ media we have years of experience in applying rigorous structured tests, or market mix models, to tease apart the media and marketing factors that drive success and return to a business from advertising investment. Moreover, we know how to do this at the highest levels of statistical confidence. We have spent a long time determining the complexity of the interplay between media channels to deliver overall campaign effectiveness, as well as how those change at different levels of investment.

I believe also that the problems with measuring effectiveness are exacerbated with the growing focus on how media is bought, and the move to increasing automation. Buying media programmatically has brought with it many efficiencies, a way of navigating a way through the myriad of choices available online. 

It is, however, not all good news.

We all instinctively know (and work we’ve done at Newsworks proves it) that not all digital is equal, that there will be a difference in effect depending on where your ads appear. Yet with programmatic it would seem that it is, with a focus on audience alone, regardless of context.

This often means that there is less transparency than might be expected. We’re in the middle of a research project on live campaigns measuring the impact of both print and digital newsbrands on the overall brand health metrics resulting from a campaign, which has meant recruiting brands in advance, and then setting up measurement when the campaigns are running. In quite a few instances, the people planning the campaigns have told us they won’t be able to determine whether they have appeared in digital newsbrands until after the campaign has run. With the increased focus on brand safety it seems extraordinary that this is possible. Never mind the not appearing in unsafe environments, what about the positive benefits of appearing in quality environment?

To my mind, lumping all digital together is like lumping TV, radio, newspapers and magazines into one, rather than using each medium for the different roles they play and the contribution to overall campaign performance.

The industry needs to stop conflating content with platform and start separating out the different types of online environment. And that’s not just about newsbrands, it’s about all quality environments online.

Our recent study, Planning for Profit, conducted for us by Benchmarketing, proved for the first time that the quality media environment online is profitable. Online display as a whole is dragged down by lower-quality sites, and therefore seems unprofitable. That is clearly not the case.

At our upcoming Effectiveness Summit, we will reveal the results of our long-running partnership with GroupM, which gets around the issue of all online inventory being lumped together by starting from scratch, and sets up processes which allow us to plan campaigns separating out the quality environment. And if we can plan it separately, then that is half the battle towards measuring its effectiveness.

This kind of work is vital if we are to stop taking an incomplete approach to measurement. If we are to plan effectively, then we need to plan with as much complete information as possible, not just plan for what we know. It’s time to stop the measurement tail wagging the media dog.

Denise Turner is insight director at Newsworks