A / B split testing is a great way to determine your best performing ad without delivering different versions of your advert to the same people within your audience, essentially forcing you to bid against yourself for advertising space.
Facebooks Split Testing does not work with all campaign objectives; you will need to select one of the following:
When creating the campaign, you will see a checkbox that you can select to ‘Create split test’, then, as normal, name your campaign and continue on to the ad set.
At ad set level your audience is automatically divided into unique, non-overlapping groups. Each ad set has one distinct variable. Your test variables can either be different audience types or different ad delivery optimisation methods. You can also choose to split the budget and reach evenly across ad sets, or weight one more heavily than the others.
Performances are measured and compared over a series of days. You can create up to five ad variations. On the left, you’ll see Ad A through E (if applicable). The success of each ad set is measured by your campaign objective, then recorded and compared. The best-performing ad set wins.
Until now, you could split test delivery optimization, for example Conversions vs. Link Clicks; Audience, split testing Web Traffic vs. Page engagement, being sure to exclude one audience. The latest feature to be added to Facebook split testing is the option to split test ad creative.
This is a major change to Facebook’s built-in split testing tool, the process of ensuring that audiences don’t overlap is now automated. Facebook randomly determines who is tested for each variation, without the need for exclusions.
Your split test will need to run for between three and 14 days. This is required so that Facebook can get the sample size necessary to determine a winner. There is, however, an option to end the split test early once a winner has been found…
After the test is complete, you’ll receive a notification email containing the results. Use these insights to fuel your advertising strategy. The email will include; the chosen winner, a small analysis of the results and a percentage chance that you would get the same winner again should you choose to repeat the test.
The only down side we found to this testing method is, that once the winner is chosen, you are unable to just turn the advert back on, or even duplicate the ad, instead you will need to duplicate the campaign, but there is a button at the bottom of the results email to speed the process up.
Overall the new dynamic split testing is much more streamlined and user friendly than previous methods. So, next time you are creating an ad, why not get creative split test multiple ideas to get the best out of your adverts.