The recent publication of the IAB UK Measurement Tool Kit was accompanied by the provocative suggestion that some advertisers were ‘clickheads’, relying too heavily on click-through rate as a measure of digital ad effectiveness. Dunnhumby’s Jake Ferguson explores why this may be the case, revealing results of research into the correlation between CTR and business results.
Back in August 2017, dunnhumby ran a study which showed there was no correlation between clicks and sales impact.
With the digital media landscape constantly evolving, we decided to revisit this 18 months later to understand if it has changed. During the intervening period we’ve seen GDPR come into play, the rise of brand safety and audiences becoming even more fragmented, whilst at the same time digital media spend in the UK has increased by 15% YoY (H1 2018).
Click-Through Rate (CTR) hovers around 0.10% for standard display which equals to 10 in every 10,000 clicking on the ad – and this figure doesn't even consider bots or accidental clicks. Whilst clicks are gradually becoming a less used part of the marketer’s toolkit, a lot of marketers still use them as a key metric related to sales.
When you think about your own behaviour – when have you ever seen an ad, clicked on it and then gone on to purchase immediately? This is not the typical customer journey for most brands or retailers. If, like many brands and retailers, the majority of your sales are transacted offline, to truly determine the success of your digital advertising campaigns, you need to be able to attribute this activity to real sales.
Rather than clicks, the true measure should be the impact on purchase behaviour after exposure to the campaign. While this might seem like an unattainable holy grail of measurement, it is possible. By linking display campaign exposure data directly to customer-level store behaviour, you can understand the effect of digital advertising on actual sales.
Our previous study looked at data from over 60 FMCG and General Merchandise digital advertising campaigns and found no pattern between sales uplift and click through rate. In February 2019 we used the same methodology again, looking at over 40 campaigns from the last 18 months and the results spoke for themselves.
Once again, can you see a pattern? We certainly can’t. Our second study shows that click-through is in no way related to sales performance for display campaigns. This again reflects the exact conclusion we came to in 2017. Whilst data isn’t yet available for other channels, it’s a safe bet that the story is the same.
In fact, the correlation coefficient in the graph above is equal to -0.28. A correlation coefficient has a value between +1 and −1, with 1 being a total positive linear correlation, 0 no linear correlation, and −1 being a total negative linear correlation, or inverse correlation – where one variable rises, the other decreases. The closer a correlation coefficient is to zero, as is the case here, the more likely there is no relationship between the two variables. With a coefficient of -0.28, our research suggests the correlation between CTR and sales uplift is non-existent.
What’s also interesting about these results is that the campaigns do, in general, work. In fact, we’ve found plenty of predictors of campaign performance; relevant target audiences; tailored, simple creative messaging; and the right frequency of exposure. These are the key drivers of performance.
In other words, good marketing practices lead to high-performing campaigns – and the right approach to measurement brings this to light. Concentrating on clicks does not. At best, you’re having no tangible impact on your digital media performance, and at worst you could be promoting bad campaigns at the expense of good ones due to a flawed measure.
Click-through rate is well past its sell by date, and may be hazardous to your brand’s health.
For further insight into the most appropriate and effective measurement strategies for your campaign, download the IAB Measurement Toolkit.