Practical Ecommerce

How to Test and Change Facebook Ads, to Drive Performance

Advertisers should frequently change and test ad copy to drive results. Facebook's ad platform makes testing challenging. But it's still doable.

Advertisers should change and test ad copy to drive results. Facebook’s ad platform makes testing challenging. But it’s still doable.

When was the last time you updated your advertising copy, such as on AdWords or Facebook ads? If you’re not updating your ads, your frequency — the number of times one person sees a specific ad — is likely going through the roof. A high frequency is typically not good in terms of return on overall ad spend.

To change ad copy, do not just pick a new headline and image out of thin air. Instead, take the data from the ads you have been running to make a better ad, or at least test versions of your current ads to improve performance.

In this article, I’ll describe a mechanism for writing better ads. I’ll address when to change ads and when to stop them. I’ll use Facebook ads as examples, as Facebook provides positive and negative feedback.

Ad Frequency

What is an acceptable ad frequency? This will vary slightly depending on your shoppers’ place in the funnel — i.e., first time vs. remarketing.

When it comes to people that haven’t visited your site before, aim for a frequency of no more than two. Some advertisers might call this too low and too difficult to achieve. But think about it this way: Say you have a 1 percent Facebook lookalike audience in the U.S. That’s 1.9 million people. It’s going to take a while to reach them all. Avoiding a frequency of two isn’t hard.

Now, for remarketing ads. This frequency can be higher, as these ads serve more of a reminder. I wouldn’t go above five, however. That’s because, for a remarketing audience, it’s less about the wow and more about the reminder. Additional reasons for a higher frequency for remarketing ads are as follows.

  • Smaller audiences. Remarketing audiences are typically smaller, especially if the audience is heavily segmented. Less people means more chances they’ll see an ad.
  • Lower cost. Remarketing ads are often cheaper. Depending on how much budget you allot to your remarketing campaign, you could overspend, and thus drive up the frequency.

Refreshing Ads

So how to refresh ads, to keep the frequencies low and the performance high? Unfortunately, with the Facebook ads platform, advertisers cannot test rotated ads. The system will automatically show the ads that preform the best.

Unlike AdWords, there is no “Rotate Evenly” option on Facebook, much to many an advertisers’ dismay. There is a workaround, however, but it doesn’t always work. You can launch two identical ad sets, and have a different ad in each. This means if you want to launch a new ad, you’ll likely have to pause and re-launch the existing ads you want to submit against.

Facebook’s Power Editor makes this easy, but it’s still a laborious process. Until Facebook offers greater control over ad rotation, however, this is how it’s done. As is, when you launch a set of ads, you’ll have to keep a close eye on what works, because the system will quickly choose. Then it’s on you to innovate and improve.

Testing Ad Copy

So how do you take your learnings from Facebook and improve? It can be tricky if winning ads get automatically chosen in just a few days or even hours, depending on your budget. It’s mostly a game of statistical significance.

For example, say we have two ads: one has spent $1,000 and has a cost per acquisition of $100. The other has spent $60, and has a CPA of $20. Which one do we pick? The ad with the CPA of $20, right? Wrong. Even though the CPA is $80 lower, the scale is drastically smaller, so those numbers can’t be trusted.

You’ll likely run into scenarios like this frequently, waiting for the results to be statistically significant. (I use a free tool from Visual Website Optimizer to check if the results are statistically adequate.) Once you get to significance, choose your winning ad. From there, change it, and run the changed version against the original. Do this repeatedly until you’re satisfied with the results — knowing you should always strive for improvements.

Is ad frequency a concern for you? If not, are you testing your ads? Please let me know in the comments.

Michael Stancil

Michael Stancil

Bio   •   RSS Feed


email-news-env

Sign up for our email newsletter