As I write this blog post about the continued but inappropriate use of the click for evaluating the effectiveness of display ads, I’m reminded of Henry II’s famous quote when he obtusely referred to Thomas Beckett, the Archbishop of Canterbury: “Who will rid me of this meddlesome priest?” Back in 1170, some Knights in armor were only too happy to help Henry rid himself of the Archbishop. If only it were so easy to rid ourselves of the meddlesome display ad click today. Where’s a good Knight when you really need one?
I thought it had become evident to everyone that the click is an inappropriate metric to use to gauge the effectiveness of display ads, since the evidence of the irrelevance of the click is overwhelming. Consider the following:
In light of this extensive evidence, it’s puzzling that some in the online advertising industry still stubbornly cling to the use of the click. A recent study of the online advertising ecosystem conducted by Advertiser Perceptions and sponsored by Collective found that about two-thirds (64 percent) of advertiser and agency executives continue to use click-thru rates (CTRs) to evaluate ad network performance. The study also found a disparity between senior- and lower-level agency decision makers, with the latter relying heavily on CTRs and the former leaning on other metrics. This would appear to suggest that the industry needs to do a better job of education.
In a recent blog post, Forrester analyst Emily Riley wrote:
Marketers must move to the "equal credit" measurement model and abandon "last click." I can't tell you how many marketers and agencies I know that still run ads across 10 or more ad networks or portals and don't measure frequency, exposure or share of voice across the entire media buy. Most importantly, they aren’t distributing credit across all ad exposures, but are still tied to the last click model, ensuring that they capture no early funnel value at all.
Why the fascination with the click? I suspect one reason is that computing clicks on display ads is fast, cheap and easy. In a world where time and money are of the essence, these factors may well be sufficient for some to put aside concerns about the click’s relevance. If correct, that’s certainly not a positive commentary on the state of the online ad business.
During a recent client meeting, I was asked how click rates on display ads compare to those on search ads and how the two approaches compare in terms of behavioral impact. I thought you’d find the answers to be interesting and, hopefully, the following discussion will help move more executives to use non-click metrics when evaluating the effectiveness of display ad campaigns.
First, we can see that searchers (who represent fully 90% of monthly Internet users) are far more likely to click on an organic link (97%) or a paid search ad (63%) than Internet users are to click on a display ad (16%). This isn’t surprising. Searchers are, by definition, much more likely to be “in the market” for a product or service and based on the fact that they conduct a search are very likely to be looking for information. So clicking on a link (paid or organic) in the search results is not surprising.
The average click rate (defined as the percent of paid ads that were clicked on) for paid search campaigns (3.5%) is massively (35X) higher than for display ad campaigns, reflecting the fact that the click is a far more relevant metric for search as opposed to display advertising. Still, the vast majority of even paid search ads (about 96%) are not clicked on by searchers.
The advantage of search over display as a direct response tactic is not just reflected in its far higher click rates; we can see from the data above that clickers on search ads are twice as likely as clickers on display ads to complete a purchase during the Internet session where they clicked. So, as a direct response tactic, it’s clear that search works better than display because it elicits a far, far higher click rate and a higher conversion level among its clickers.
However, this is not to say that display advertising doesn’t have value as a branding strategy or as a supplement to a paid search campaign – because, as we’ll see, display ads can substantially increase the number of trademark search queries while also lifting brand sales. This is where it’s important to go beyond the click when evaluating the effectiveness of display campaigns and measure consumers’ response to advertising over time.
Here’s what Comscore has found. We used the Comscore panel of 1 million U.S. Internet users whose behavior we track. For a variety of retailer ad campaigns, we analyzed the offline buying behavior of three groups of consumers:
We measured the panelists’ offline buying by linking them to their retail store loyalty card data which gives us a direct measurement of their actual in-store buying activity. As a control, we also looked at the buying behavior of a balanced control group of panelists who did not receive either a search or display ad. The following chart shows the lift in retail sales within the exposed groups relative to the control group over a four week period following initial exposure to the ads:
Among those who were exposed only to display ads, the lift in sales over a four week period following exposure to the ads was 16% (even though the average click rate on the ads was only about 0.1%). By using a four-week period subsequent to ad exposure, we’re able to tease out the latent impact of advertising and move beyond a measurement of just the immediate impact. In essence, we’ve proved that online display ads work in the same way as traditional media by building sales beyond the initial exposure – and that an immediate click isn’t a relevant metric. To an experienced advertising practitioner this is probably obvious, but to those in the digital ad world without prior advertising experience, this is likely to be valuable new insight.
As might be expected, the lift in sales among those exposed only to search ads (+82%) was greater than the lift among those exposed only to the display ads (+16%). This reflects the fact that searchers are much more likely to be “in the market” (i.e. way down the purchase funnel). However, there is synergy when a display campaign is overlaid on a search campaign, with the combined impact (+119% lift) being greater than the sum of the parts.
There’s more to the story however, because valuing the sales impact of display versus search advertising isn’t complete without addressing the higher reach that’s possible with display advertising. There is a natural limit to the number of people that can be reached with a search ad, which is a direct function of how many people search using the keywords purchased by the advertiser. On the other hand, display ads can theoretically be delivered to all Internet users, whether they are searching or not. When one factors in the typical reach that Comscore has observed for display ad campaigns relative to search, the total sales lift from a display campaign can be substantially higher than for a search campaign, as can be seen below.
Search is a very powerful tactic to use when you’re intent on reaching consumers far down the purchase funnel, but smart marketers would be advised to always consider the use of a display campaign as an overlay to a search campaign. With this approach, they can reap the benefits of the synergy that exists between search and display advertising and have the best of both worlds – latent branding impact and immediate short term response.
But when evaluating the importance of your display ad campaigns, steer well clear of the meddlesome click. Henry certainly wouldn’t approve.