| November 27, 2017
Click-through is the most ubiquitous digital metric, freely and immediately available for every digital execution. As a result click-through has become the dominant metric in digital campaign optimization irrespective of all the evidence that clicks have little relationship with brand performance and sales.
As Jane Ostler notes in her recent Admap article titled, ‘Join up the blurred lines between digital and analog media’ not all data is of equal value and if marketers optimize solely on clicks then they lose all sight of whether their campaign will build the brand and drive sales. Look, we all understand that click-through facilitates real-time optimization but what is that worth when the there is no relationship between the metric used and the outcome you want?
The fact that click-though does not correlate with brand outcomes is not new news. Years ago a study conducted by Millward Brown and Yahoo! demonstrated that there was no correlation between click-through and brand response. Since then we have seen that finding replicated by Nielsen and others for attitudes, online behavior and sales. Recently, Gonzalo Fuentes, CEO Global Media & Digital Practice, presented at the IAB conference in Australia titled ‘MEASURE UP!’ and told me about a presentation from Quantium, a company which databases loyalty and bank card purchase data from over 10 million Australians. Quantium’s analysis found no relationship between click-through and incremental sales per person.
If there is no correlation between click-through and sales it means that in 50 percent of cases optimization based on click-through will improve effectiveness and in the other 50 percent of cases it will undermine effectiveness. Marketers often complain that attitudinal data does not perfectly predict behavior; well, guess what, attitudinal data is way more predictive of behavior than click-through. Using click-through to guide decisions is a classic case of convenience being more important than relevance, it enables real-time decisions but whether those decisions are optimal is a whole different ball game.
Are there alternatives? Yes there are. First off, do not make assumptions. A systematic approach to measuring all aspects of a campaign, clicks, brand lift and sales, will allow you to figure out whether clicks are a good proxy for brand and sales success or not. If click-through is not a reliable predictor, you could run a test before committing the full media budget and check whether and how the campaign works, allowing you to adjust the creative mix and spend allocation for the main campaign. Or, as some of our bigger media spenders do, you could use brand lift monitoring response during the campaign and adjusting the mix as the campaign evolves, there might be more human intervention involved but the payback will be higher too.
Click-through is an easy and convenient way to enable real-time optimization but unless you have proof it relates to outcomes that build brand value you might as well use a random number generator to optimize your buys. How do you think that will go down with the CFO? Please share your thoughts.