Beta

Learn About How to Make Winner Ads

Discover how you can create winning ads that will increase your conversions and drive more sales. Learn the tips and tricks from our experts!
Caroline Lowery

So you just ran a multivariate test on your ad creative. You’ve identified the winning ad elements. You’re creating a plan to apply all this new creative intelligence across your brand’s touchpoints. And you’re about to ban the losing elements from being associated with your brand ever again. *dusts off hands*

But I’m here to tell you that it’s actually incredibly smart to take a second look at your losers. Because, while it’s easy to glean actionable insights and learnings from your top-performing ads, the assets and ads that didn’t fare so well have a lot to tell you, too, if you just dig a little deeper.

Here are four simple tips for finding insights among the lowest-performing ads in your test.

Identify the major differences between the best- and worst-performing ads

Take a look at your top few ads and your bottom few ads. Can you categorize the differences between them? Categories can be as broad as “humans present” and as granular as “women wearing a blue shirt and holding a phone” for images, or “UGC generated quotes” and “Spring campaign TikTok testimonials” for copy.

(Pro tip: the more you group like assets into collections before you run your multivariate test, the easier it is to determine what separates your winning ads from your losing ads.)

Below is an example from our customer, Credit Report Canada. Here we can see that the majority of our bottom four ads feature people using technology, while the top four ads feature people looking directly at the camera. We can also see that our bottom four ads mostly feature families. While three of our top four ads feature lone women.

This would tell us that we should explore the “lone women” category a bit more, as well as “people looking directly into camera.” 

Top and bottom four ads for CPA are shown.

On the flip side, look to see if your losing ads contain any winning elements. We can see that two of our bottom four ads contain the orange and green graphic, while the other two contain the dark image overlay and centered headline, all of which are also featured in our top four ads. This would lead us to understand that the ad’s image is the key creative element driving performance and increasing — or decreasing — conversion.

Pinpoint what all of the lowest-performing assets have in common

Uncovering what’s similar across your low-performing assets can help you understand what to test next. 

Here are a few examples:

  • If all of the low-performing images include people looking at a computer or phone, then explore similar style images without a computer or phone.
  • If all your low-performing copy lines were related to a single value prop, then explore other value props and shift your tone of voice to that of the winning copy lines.
  • Are all of your low-performing product shots photographed at a certain angle? Explore other product shot angles.

Take another look using a different KPI

It is important to note that your losing ad from a CPA perspective could be a middle-of-the-road performer from an engagement perspective. Is it really a losing ad then? Nope! 

Here’s another example from Credit Report Canada. The highlighted ad was a low performer for CPA and lead generation, but landed in the top four ads driving CTR. 

Top four ads for CTR and bottom four ads for CPA. One ad is highlighted as it is in both groups.

The lesson here? Be sure to test assets and ads against different goals and take a look at the full spectrum of results before deciding which ads to scale. 

The best evergreen campaigns contain ads that drive conversions and ads that drive engagement. Market-leading brands test ads to find top performers that work together to drive both KPIs.

Retest losing elements down the road

Just because a certain image, headline, or color was a loser today doesn’t mean it will be in the future. There are a multitude of variables that can impact ad and creative element performance, including audience, trends, and seasonality. 

Maybe your black and white background didn’t perform very well in June, but will in January. Perhaps new customers are more likely to click on ads promoting discounts, while returning customers prefer ads promoting a new product line. 

The moral here is: don’t be afraid to bring a low performer back for another, slightly different test. It could become a winner after all.

Learn to love your losers

Ultimately, insights from losing ads are useful, even if they don’t all make it into the final campaign. They can point you in directions that were never considered or provide counter arguments to your original hypothesis. And in some cases, they can even be used to inform how you deploy similar ad elements and assets across touchpoints throughout the rest of your brand’s journey.

So the next time you’re looking at low-performing ads and thinking, "How can these even be worth digging into?" remember: there just might be some incredible learnings hiding in these losers — if you look at them in the right light.

Boost ad performance in days with a 7 day free trial.
Claim Trial

How to Run a Multivariate Test

The Beginner's Guide

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Tiffany Johnson Headshot

How to Run a Multivariate Test
The Beginner's Guide

Plus, Get our Weekly
Experimentation newsletter!

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

We run DPA for the world’s biggest brands… and then we share the important things with you.
Join 10k other marketers and subscribe to the Catalog Cult - the world’s best newsletter about catalog ads.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Are you crazy...

about catalog ads? You’re not alone. Join over 8,000 other marketers in The Catalog Cult - the world’s best newsletter about catalog ads.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.