Einstein and Online Advertising Measurement
"Not everything that counts can be counted, and not everything that can be counted counts." – Albert Einstein
Einstein kept that statement over his desk in his office at Princeton and the relevance to today’s situation in online brand advertising is incredible.
At iMedia's Brand Summit in early February, the Brand representatives were abuzz with the results of a comScore, Starcom and Tacoda sponsored study about the characteristics of consumers who click on advertisements that showed that only 6% of the online population accounts for 50% of the display ad clicks. And more impactful, those users are mostly lower household income (less than $40,000 annually) and spend up to four times more time online than non-clickers – often frequenting auctions, gambling and career services sites. comScore points out in their release that this is a very different traffic pattern than non-clickers.
The attendees at iMedia were clear that this presented a significant issue for measuring interactive media campaigns – the click through rate may represent a completely off-target audience for brand campaigns. In fact, Starcom commented that there is no correlation between clicks and brand metrics – and there was no connection between measured attitude towards a brand and volume of clicks.
So, what does a brand measure? Often the answer is engagement – with no industry standard measurement. Some people argue that engagement is a factor of time – how long a consumer saw the ad. Some recommend that panel based brand impact surveys are the thing. Others recommend in-campaign surveys. Perhaps the best answer is - "I know engagement when I have engagement."
Prior studies by MediaVest show that consumers are more favorable towards brands they encounter on smaller, niche sites closely aligned with their interests. It appears that engagement requires reach, frequency and placement online not unlike the offline media environment – but online there are many more choices and smaller segments to reach.
As for engagement – consumers want to be heard and brands want to build loyalty. Is this one of those intangibles that counts but cannot be counted? How do you want to measure engagement?
Comments (1)
Speaking of quotes, I think it was H. L. Mencken or Albert Einstein (a quick search showed me that they are both cited as the author) who said: “For every complex problem, there is one simple – but wrong -- solution.”
For me, that sums up the challenge of measuring the effectiveness of display advertising. While it’s alluring to believe that there is one simple, easily-obtainable metric that will accurately and reliably predict advertising success, I believe this is a siren’s song. And, I suspect that most experienced researchers who have spent decades searching unsuccessfully for advertising’s Simple Holy Grail have also come to the conclusion that, while there certainly are simple metrics that can give you some insight, they’re far, far from foolproof as a measure of advertising’s impact on sales. And sales, I would argue, is the one undeniably relevant metric for evaluating ad effectiveness. Unfortunately, however, measuring advertising’s sales impact is something that’s often difficult to do – especially since it’s often vital to measure advertising’s cumulative impact on sales across time and channels and to cleanly separate this from the impact of everything else that’s going on in a brand’s marketing mix.
This brings me to the validity of the click as a measure of advertising effectiveness. For many years, the click was used as a supposedly accurate measure of the effectiveness of display advertising. Now, I would be the first to agree that -- for some direct response ad campaigns -- the click remains a relevant metric. However, when it comes to measuring the impact of brand building advertising, the idea that consumers’ immediate response via a click is hard proof of the effectiveness of display advertising is just plain wrong. Perhaps, in the early days of online advertising when click through rates were running at levels of 5% or so, it was easy to believe otherwise. But, as click rates have dropped to a fraction of one percent it has become clear that some other metric is urgently needed. To believe otherwise today would be to acknowledge that display advertising has no impact at all. Perish that thought!
comScore’s objective in conducting the click study with Tacoda and Starcom was to prove -- once and for all -- the limitations of the click as a relevant metric to use to measure display advertising effectiveness. I believe this is a critical step in the evolution of online advertising because if our industry is to continue its torrid growth, we have to look beyond direct response advertising dollars. We have to convince the brand-building advertisers that they should move more of their ad dollars from traditional media to the Internet. Rest assured, we’re not going to be able to do that using clicks as the metric of choice. Instead, we have to be able to show that display advertising increases brand sales over time and across both online and offline sales channels. I think that Einstein would agree.