About

The client (retail) has fewer operators in their call center on weekends, so they don’t need as many leads as they do on business days. Plus, when you contact a lead from Saturday on Monday, they are much less likely to convert (some of them don’t even remember they’ve submitted their info).

Problem

Facebook ads don’t allow you to set up a schedule using daily budgets. And lifetime budget isn’t a suitable option for many reasons.

Solution

So, the only way is to pause and resume ads via automated rules. But in this case, AI learning/efficiency may be skewed and we could get a higher CPA because of that. If you would simply try this without prior testing or running both campaigns in parallel, you might get some results. But you won’t know: was it because of your changes or because of some other factors that influence CPA?

Will the pausing ads on Fridays and resuming on Mondays lead to a higher CPA during weekdays? I decided to run an A/B test to find this out.

Duration: 28 days (4 full weeks)

Results

Notice:

• how different CPA can be even during the first and the second weeks’ business days while it’s basically an A/A test. And it’s the Facebook feature when they serve impressions as evenly as possible, so the influence of other factors is minimal.
• the difference in CPA weekends vs weekdays isn’t statistically significant. A bit lower CR, a bit lower CPC.

So, why spend on lead generation on weekends, when we could get them when the probability to convert them into sales is higher? AI does its job well even if it’s restricted by time slots.

We decided to increase the daily budget and run ads on weekdays only.