How Marpipe Uses Its Own Creative Testing Ads

As head of Performance Marketing, I oversee every multivariate test Marpipe runs on our own ad creative. In a typical month, we run two to three tests with anywhere from eight to 50+ ad variants and we’ve learned some pretty important lessons along the way. In this blog post, I’ll share a few of my favorite lessons with you in hopes that you can use them in your own testing efforts.

Put some real thought into what your product solves vs. what it does. Test both. Then proceed with restraint.

Two things we like to pit against one another in our creative testing are problems solved vs. product benefits. We recently ran a test of the 12 ads below. The ads were split evenly across three headlines, two images, and two background colors. 

‍

But what we were really focused on were the headlines; which message about Marpipe would our audiences find most compelling? The first headline highlighted a major problem Marpipe solves: removing bias from ad creative. The second and third headlines were focused squarely on the benefits of Marpipe’s features: design efficiency and structured test budgeting, respectively.

What we found was that the first two headlines had significantly higher lead-generating power than the third.

‍

Perhaps the biggest takeaway was that we need to be careful about what we choose to highlight and test in our ads. For example, “We’re testing 12 versions of this ad” highlights the benefits of design efficiency, which is a great part of our app, yes — but not the part that makes us truly unique. That lands squarely on our platform’s ability to deliver micro-level creative intelligence.

That uniqueness is highlighted in the ad with the headline “Our CEO hates this copy, but it works” which addresses this very ethereal problem all marketers have: that preferences and opinions can get in the way of winning ads.

The big lesson here is to take a 50,000-foot-view approach to what you provide your customers. Then be super choosy about which messages you scale. If it tests well but doesn’t necessarily represent your brand in the right way, there’s no point in scaling it.

Start with messaging. Find a winner. Then play around with design tweaks in subsequent tests.

My testing ad creative strategy is centered around starting with big rocks and working my way to the little ones. I typically like to choose a larger concept to test, and I usually focus on messaging first — does a potential Marpipe customer like what this ad is saying about what Marpipe offers? I don’t worry about testing images and visuals early on. My goal is to first find a larger, strategic message that resonates with my audience.

Once I find a winner, I start to dig a little deeper into the nooks and crannies of the ad. Is there a different background color that would perform better than the previous winner? A different image? A different CTA? Often, small adjustments like this lead to incremental performance improvements. 

Let’s go back to the test we talk about above. I had already tested the premise of the messaging and found that our winning ad was the combination of the CEO headline, the gray background, and the fashion brand image. Next, I wanted to see what else I could do to make it work just a little harder for us.

‍

I added a background color variant in my next test. You can see the variants I tested above. Turns out, my audience definitely preferred bright green to the original grey. So long as this ad continues to perform, I’ll keep playing around with background color. You never know — maybe we can kick the performance up even one notch further.

You won’t always find a winner. Adjust your media budget accordingly.

Like most marketing orgs, our team has monthly cost-per-MQL goals. And hitting those goals has become an art in balancing strategic tests and evergreen ads we know work. 

If you were to take a look at our media budget, you’d be able to tell the weeks we’re running tests. The cost-per-MQL increases because we’re putting testing budget behind a bunch of ads, only a few of which are winners. Once we learn what we need to, we switch over to our evergreen ads to even things out.

It’s definitely a catch-22 of sorts: we can’t identify our evergreen ads without creative testing, but we can’t hit our MQL goals without our evergreen ads. To be successful, you have to temper your experimentation with your ongoing spend. This also sets you up for success for the next week, month, or quarter — you now know what ad concepts to abandon in your tests and can begin testing with fresh potential winners.

Even the pros never stop learning.

While we might be the team and the minds behind Marpipe, even we are perpetual students of multivariate testing. My team and I are always experimenting, learning, and finding new ways to apply creative intelligence to our brand.

For more testing tips, check out The definitive guide to modern creative testing — seven chapters chock-full of info for every step of the multivariate testing process. And, as always, you can reach out to our team to get help scaling up your ad creative testing.