Some thoughts on the relationship between CTR and fraud.

By November 20, 2014Uncategorized

The Folly Of The Click-Through Rate And Simple Math

As Advertising Week drew to a close, a colleague of mine who is relatively new to ad tech voiced disbelief that many people still believe in the click-through rate (CTR) as a performance measurement.

“Is it because the math in more accurate methods is too complicated?” he asked.

Unfortunately, CTR is not going away anytime soon, according to AOL’s Tim Armstrong. The click was the metric of choice when he started in the industry 19 years ago, he said in a recent IAB MIXX speech. Although there was skepticism about the click from the onset, here we are with the same metric nearly two decades later.

That’s a tough one to swallow. I tend to agree with Tim that the media industry moves at a glacial pace in this arena – look at our continued reliance on coarse demographic TV ratings, for example – but the notion that we are stuck with simplistic and misleading attribution for the foreseeable future defies the potential and power of digital media.

Do CTR and click attribution models remain because we don’t want to acknowledge how these metrics support a fraud economy that occurs across digital channels?

The relationship between CTR and fraud is symbiotic; fake sites and spoofed browsers clicking on ads lead to seemingly prodigious CTRs that exceed anything produced by human audiences. Yet, optimization of CTR remains the performance criterion of choice for many agencies and marketers.

As a result, agencies often drop vendors that may be providing the best audience for the brand. Without a movement to multitouch attribution or a more holistic cross-channel measurement, many vendors that are rewarded by agencies are gaming the CTR metric, often by buying bushels of fraudulent impressions.

Even with the current focus on fraud, the reliance on CTR is still with us, along with its enabling impact on fraud perpetrators. If you believe you can achieve triple-digit click growth at cheap CPMs, you’re probably fooling yourself with simple math. You’re certainly not practicing good science.

With the recent announcement by the IAB, ANA and 4As of their joint effort to fight fraud, good science may yet win the day. If we start to tackle the issue as an industry and begin re-evaluating CTR as a metric, we may yet achieve the promise of the medium.

Let’s not forget that mobile complicates matters even further, as small screens and minuscule ad sizes combine to render click-based metrics virtually meaningless. Further, CTR is going to pollute the data we use for analytics and measurement. The metrics that should matter are those that capture engagement or purchase as a result of marketing. Examples include real site visits and purchase, conversions, downloads, brand impact, in-store visits and purchase, all of which are metrics that indicate interaction beyond a click.

While last year’s Adobe ad poking fun at the folly of using CTR as the basis for strategic decisions gave us a good chuckle, it is scarily real, even now. We need to retire this flawed metric at long last in favor of more accurate and nuanced measures that reflect the gains we have made with data and science.

Yes, legacy and simple math make it tempting to stick with the old standard, and while there is cost involved, using a combination of cross-channel metrics and multitouch attribution will provide marketers with much stronger information on a brand, and ultimately, a better experience with digital media.

by Lauren Moores, vice president of analytics at Dstillery via Ad Exhanger http://www.adexchanger.com/data-driven-thinking/the-folly-of-the-click-through-rate-and-simple-math/