For almost as long as digital marketing has existed, those involved in the space have debated how to qualify online conversions. The core of the issue is this: “If someone sees an online ad, but she doesn’t click on it, are there other actions she takes later as a result of seeing that ad? Or is that impression wasted entirely?”
We have long argued that the value of an online ad has almost nothing to do with clicks. In fact, we shouldn’t expect anyone other than the very bored, very clumsy or extremely interested to click on an online ad. Our argument has always been that the impact of an online ad is much like the impact of a TV ad, a print ad or an outdoor ad. That is to say, the impact is more latent in nature — or that it yields results through other channels (like search, direct URL entry in the browser, store visit, decision at shelf, etc.). Those in the industry who agree with our theory refer to these actions as “view-through conversions.”
Testing Our Hypothesis
We had the good fortune to put this hypothesis to the test with two very different clients. In both cases, a test was designed to measure the real view-through business impact of online ads. Each client’s internal analytics team established representative test and control markets. The test markets received the advertiser’s creative. The control markets received public service announcement (PSA) creative, graciously donated to national charities as free media weight in order to enhance these clients’ digital understanding.
In both cases, we tracked view-through activity on business metrics for the test and the control groups. There was no possible causal or even loosely correlated relationship between running a PSA (or clicking — since the clicks drove to the charity’s site) and performing a relevant business activity with our clients. Therefore, if there was a statistically significant increase in the test market view-through business lift versus the control view-through business lift, it was agreed that the view-through results were real.
Reviewing Our Results
- Client A saw business metrics (online sales) that were 50 percent higher in test markets than in control markets. Which means that half of the people buying things online would not have done so had they not seen the brand ad. This far exceeded thresholds of statistical significance in spite of the noise of media running in those markets through other channels at the same time.
- Client B saw business metrics (online leads) that were 300 percent higher in test markets than in control markets — again, statistically significant. In this case, there was no noise to note, as no other marketing activities were running during the time of the test.
In each of these tests, the clients donated tens of thousands of dollars in media impressions to charities. As advertisers, our hope would be that you learn from the curiosity and generosity of these two case studies. View-through conversions are indeed real. The magnitude of that realness is likely dependent on the noise you have coming from other marketing activities running at the same time. So what should you do given that the business reality dictates you will be running online ads at the same time you’re doing deploying several other tactics?
What Marketers Need to Do
- Ignore clicks (they do not consistently correlate to any other actions, including sales) and last touch attribution sources like ad servers and analytics packages.
- Deploy a statistical approach to attribution (that can track online and offline channels).
- Use lifetime value and the full understanding of statistical attribution to assign credit to channels and optimize within and between them.
Easier said than done — all of the above suggestions take time and resources. In the meantime, at least make sure you’re giving credit to your view-through conversions.