The Truth About Testing Your Facebook Ads

👤 by The Bootstrap Marketer // 📅 April 29, 2022

Are you running Facebook or Instagram Ads? Do you test different audiences, offers, and creatives?

In this article, you’ll learn the biggest mistake business owners are making when testing their ads and how to ensure your get the best results from your ads.

Let’s face it; every Tom, Dick, and Harry tells you that you need to test your ads. And they are all 100% correct. How do you know what is and isn’t working unless you test?

You could be fishing in a lake thinking there is fish because you caught a few, not knowing that by changing the bait you’d catch more. Or maybe the lake down the road is full of fish and it doesn’t matter what bait you use!

So terrible fish analogies aside, it’s important you’re always testing.

Testing 101

An A/B split test is testing in its simplest. It’s variation A vs variation B, with the winner being decided by your objectives.

These could be messages or Add to Cart (ATC) on Facebook Ads or click-through rate (CTR) or Conversions with Google Ads. Every split test system starts with the measurable outcome you want to improve.

It’s important to remember the metric measured needs to be relevant. With any advertising, we want to ultimately get customers for the lowest price (CPA or Cost Per Acquisition). Measuring which ad has more clicks isn’t a reliable measure for business success unless it’s part of a bigger plan.

With every split test, a few things need to remain the same so you can identify a winner.

1. Time Frame. Both ads need to be run over the time frame and duration. I’m sure an ad that sells turkeys will be more successful at Christmas vs an ad selling turkeys at Easter. (I may be wrong but – that’s why you test!)
2. Audience / Environment. To determine which ad is more successful, it must be shown to the same people. If we’re selling reusable baby nappies, I can’t compare if Variation A is better than Variation B if one ad goes to new parents and the other to 18-year-old boys.
3. Amount Spent. On most advertising platforms, spending is proportional to reach. Both variations need to have the same amount of spending attributed to ensure it’s a fair contest. You must take control of this with platforms like Facebook, which may favor the more successful ad variation to be shown.

Once you have a fair split test happening, it’s important to analyze the outcome to see if there is statistical significance. You must have measurable outcomes to determine if one ad is better than the other.

For example, let’s say we are doing a Messenger ad where the outcome and call to action are to get the potential customer to send us a message.

Ad Variation 1 = 3 messages @ $12/message

Ad Variation 2 = 1 message @ $36/message

At first glance, it looks like Ad Variation 1 is the clear winner. It’s had more success, and with the same budget applied it’s worked out a third of the lead price than Ad Variation 2. However, we need to look at if there is statistical significance.

But what does that mean?

In a nutshell, statistical significance in an AB test is how likely the difference between the results in the test isn’t due to error or chance. Most calculations are done on a 95% significance level, which is simply the confidence you would have that the result would occur.

Whilst the formula is advanced maths, we don’t need to know that. Survey Monkey has a great significance calculator we used to test our results.

As shown above, the results are not significant. One of the obvious reasons, is we have a very low sample size of success (we only got 4 messages total). We would either need to keep the ad running to get more results to find significance, or we could change the objective.

If we had a higher-level objective with more success such as Add to Cart or Website Traffic, we might be able to determine a result.

A general rule of thumb is there is no winner until it’s generated 10+ results at least. BUT you mustn’t keep an ad going for the sake of getting 10+ results. You need to know your numbers, and if your CPA exceeds how much you’re willing to spend to get a client; it’s time to put a stop to the ads.

The Big Misconception with Split Testing

Here is the WORST piece of advice I hear a lot of gurus giving: split test everything.

Front fonts, headline tweaks, emojis used, and to what day of the week you run your facebook ad.

There is a saying somewhere about polishing a turd. No matter what font you use for a bad ad, it’s not going to make a difference. You’ll keep losing money hand over fist until you give up on paid ads and your business is broke.

Simple rules apply to successful split-testing ads.

If your ad is not working well, you need to split-test major changes. This should start with the biggest to smallest changes. Normally it would be the offer that changes, the audience, and then the actual creative (image or video). Unlike the bad advice, you don’t keep everything else the same. I can tell you comprehensively that you can’t just keep everything the same. You need to make wholesale changes to the offer or creative.

Now if your ad is working well – fantastic! Now is the time to split test minor changes. For example, does image A or image B of my product get more conversions? Is headline A better than headline B? All these are good microchanges so we can get the best out of what we already know to be a good-performing ad.

Remember, you are only finding the best ad for those set of circumstances. The best ad is for that audience, at that cost in that time frame. Some ads may perform better with different audiences, so an ad that fails in one setting may be a winner in others.

The 80/20 Rule of Split Testing

To go back to our poor choice of metaphor, even when you’re catching fish there may be a lake around the corner that has bigger fish that are easier to catch. Unless you’re constantly trying different audiences, you may not know if there is more success to be had.

A general rule can be to devote a portion of your budget to constantly testing new audiences and creatives. Many people we know use this 80/20 rule; devoting 80% of your budget to ads that are working well and minor changes, whilst devoting the smaller 20% to testing new audiences and ideas. It’s not always a new audience, but it is a major change.

Keep up to date with our latest content on Facebook – https://facebook.com/thebsmarketer

Are you relying on word of mouth and referral marketing for your business? Do you think your business could benefit from more referrals from existing customers? Check out our latest article on referral marketing.

Disclaimer: The information provided on this website is provided solely for educational purposes and does not constitute any advice, including but not limited to, investment advice, trading advice or financial advice, and you should not treat any of the website’s content as such. The Bootstrap Marketer recommends that you independently research any information contained on this Website. The Bootstrap Marketer cannot guarantee the accuracy of any information listed on the website and is not responsible for any missing or wrong information. All information is provided as-is and should be used at your own risk. The Bootstrap Marketer disclaims all responsibility and liability for your use of any information found on the website.