Les Binet and Sarah Carter of DDB get a little bit angry about some of the nonsense they hear around them... like the idea that ‘accountability’ and ‘effectiveness’ are valued.

Listen to any speech at any marketing conference these days, and, before long, the brave new world of evaluation in a digital age is celebrated. This new world means no adspend is wasted. Uncertainty is eliminated. And we have full accountability, as the evaluation game is raised to hitherto unknown heights.

Grrr... Dig a little deeper and you'll find that much of this promise is a false dawn.

Yes, we have more data than ever before. Yes, technologically, online measurement is very sophisticated. But analytically, we are not so impressed. A lot of online evaluation is actually pretty crude. We saw an example of this last week. A brand with large media spend over many channels had done analysis of response by channel and concluded conventional advertising did almost nothing for its business - because all the responses came via other channels.

One of our common industry faults is our tendency to see the world in discrete camps: buyer/non buyer; Generation X/Generation Y; heavy user/light user. Not surprising then that we tend to view the online and offline world quite separately. This can lead to some dangerously oversimplified conclusions about the effectiveness of digital spend, which gets nowhere near measuring its true effectiveness. Most digital evaluation is based on the direct response paradigm - really just a variation on the kind of evaluation that mail order companies were doing 100 years ago. So we run a banner ad, measure response and work out ROI.

But this gets the numbers wrong. Why? Well it becomes more obvious when we put aside our separate boxes and consider digital in a wider context.

First, there is the ‘last click attribution’ problem. Suppose a thousand people clicked on our banner ad. Simplistic analysis attributes all of these clicks to the last click - the banner - but in reality, those clicks will have been partly due to other marketing activity. To measure the contribution of the banner properly, we need to account for these influences, including offline marketing like TV ads, and posters.

Second, there is often a cannibalisation issue. Not all those clicks represent incremental business. Some responses would have happened anyway without the banner. And some might have happened offline -people might have phoned a call centre or walked into a shop. To measure the effect of the banner properly, we need to understand how it affects these other channels.

Third, there may be ‘halo effects’ not being measured this way. Some people who see the banner may go straight to the website, pick up the phone, or go into a shop, rather than clicking through.

All three effects show the importance of understanding inter-relationships between online and offline worlds.

As an industry, we like to focus on change at the expense of the unchanging. But contrary to what most people think, most of our life still takes place offline, not online. IPA Touchpoints data shows around 90% of all media consumption takes place offline. In the same way, while social media and online spending are growing strongly, still over 90% of communication and 90% of spending takes place offline.

So, sophisticated digital evaluation giving true digital ROIs needs to measure how both online and offline media drive sales, both online and offline.

Fortunately, we are now beginning to learn how to do this. Increasingly, analysts are using econometrics to measure these linkages, rather than relying on simple direct response techniques now 100 years out-of-date. Google are fans of this new approach.

So, as far as evaluation is concerned, the real excitement lies not in what is happening now. The real digital revolution will occur when we stop thinking and talking about ‘online’ and ‘offline’ as two discretely different worlds. Only then will we be able to measure its true potential.

This article originally appeared in the November 2011 issue of Admap. Click here for subscription information.