Can Display Ads Predict the Presidency?

By: Caitlin Baker  | 11/23/2016

Like most of the country, well, actually more like half of the country we now know, we were totally wrapped up in the election frenzy. It served as water cooler fodder and client small talk; we wrote a blog post predicting the outcome based on digital marketing success measurements, and then wrote another blog post when things seemed to hit a fever pitch, re-predicting the outcome (spoiler alert: we couldn't).

But that wasn't enough for us.  In our marketing minds, we knew we needed to find a way to capitalize on all the fervor.

A lightbulb went off in our heads after hearing a few coworkers constantly repeating, "make marketing great again" – a nod to Donald Trump's campaign slogan. We needed to use the phrase, especially since Donald Trump and the RNC would be gracing Cleveland with their presence in the very near future.

Our remarketing ads were ready for a refresh, so we put the design team to work with very few guidelines – except to make the ads as "American as possible." And they were:

We ran the Trump-inspired ads for a couple of weeks around the RNC and leading up to the key weeks surrounding the election. That's when we had another idea: we could have our own election, of sorts.

We'd been running the Trump-inspired ads, what if we also created Hillary Clinton-inspired ads? What would the results be? Would one set of ads outperform the other?

With our interest piqued, we put the design team to work again. This time, we told them to model the ads after Hillary's yard signs, buttons, and stickers, with a play on her "Stronger Together" campaign slogan. The end result was really convincing:

We ran the Hillary-inspired ads from October through the election, along with the Trump-inspired ads.  Now it's time to dig into the analytics to see which performed better.

Did our banner ads see something the pollsters didn't?


First, let's look at the ad metrics. The Trump ads saw significantly more impressions, which led to more clicks. But, the telling metric here is the click-through rate (CTR), or ratio of users who clicked the ad to the users who saw the ad.

The Trump ads had nearly double the CTR of the Clinton ads. This reveals that the Trump ads were much more likely to be engaged, compared to the Clinton ads. 

If we dig into the landing page metrics, we can see users who clicked from the Clinton ads were less likely to bounce from the page than a user clicking through from a Trump ad. But users from a Trump ad were more likely to spend a little bit of time on the page, as well as explore additional pages on our site.

So, overall, it looks like our ads did, in fact, predict the outcome of the election! The Donald Trump-inspired ads performed at a higher level than the Hillary Clinton-inspired ads. Not by much, but enough for us to call the race.

Just like the United States.


Share this

Blog post currently doesn't have any comments.
 Security code