About Test and Learn: The "Which campaign causes the lowest cost conversions to occur?" Test

30/11/2019

"Which campaign causes the lowest cost conversions to occur?" is a question you can answer using Test and Learn to run a campaign comparison conversion lift test. The test allows you to compare two campaigns to see which has a lower cost per conversion lift. You'll be able to start seeing results after each campaign has received at least 100 of the conversion events you're testing. In your results, you'll find the following metrics:

  • Cost per conversion lift
  • Conversion lift percent
  • Conversion lift
  • Sales lift
  • ROAS lift

How does the test work?

Basics

We measure the conversion lift of two campaigns. Conversion lift refers to conversion events that wouldn't have happened without your ads. We call this type of conversion event an "incremental conversion". After determining the amount of conversion lift for each campaign, we calculate and compare their average costs. The campaign with the lower average cost per conversion lift is declared the winner.

Details

To determine these metrics with scientific rigour (which ensures the accuracy and meaningfulness of your results), we start by randomising everyone eligible to see Facebook ads into four groups. Randomisation ensures that each group is balanced. Balance matters because we don't want any group to be more likely to convert than another. There are two "test" groups and two "control" groups. A test group is made up of people who can see your ads. A control group is made up of people who can't see your ads. Each of the campaigns in the test gets a test group and a control group. Having two groups help us determine which conversions are caused by your ads and which would have happened without them.

Note: We still only show your ads to people in your target audience. Because we start by putting everyone eligible to see Facebook ads into a test or control group, a subset of each campaign's target audience will be in its test group.

Once the test is over, the test and control group distinction is removed and we go back to showing your ads to anyone in your target audience.

Can I run a campaign comparison conversions test?

You can run a test if you have:

  • A Facebook ad account with at least two active campaigns
  • Events that are being tracked by a Facebook pixel, SDK and/or an offline data source

Should I run a campaign comparison conversion lift test?

It depends on what you want to learn about your ads. If you're trying to figure out which of two advertising tactics is most effective, we recommend this test. Ideally, there will be just one major difference between your two campaigns (e.g. audience or optimisation event). This helps isolate what aspect of a campaign leads to better results. High-confidence results (90% or higher) may even be applicable to future campaigns that weren't part of the test.

If you're just trying to figure out which of two campaigns causes lower cost conversions, this test can tell you that too. You're not required to limit yourself to one difference between campaigns. However, we do recommend it to get the most out of a test.

How do I set up a campaign comparison conversions test?

To set up an campaign comparison conversion lift test:

  1. Go to Test and Learn.
  2. Find the "Which campaign causes the lowest cost conversions to occur?" box and click Set up test.
  3. Enter the two campaigns you want to compare.
  4. Click Add event to choose which conversion events you want us to measure during your test. You can choose up to ten events from across your Facebook pixel, SDK and/or offline conversion tracking data sources.
  5. Set start and end dates/times for your test.

    Tip: We recommend running your test for at least a month.
  6. Optional: Edit the title of your test.
  7. Click Review and make sure that everything is set up as desired.
  8. Click Create test.

For tips on setting up the most effective test possible, read our best practices guide.

What if I need to cancel my test?

If you need to cancel your test after you've created it:

  1. Go to Test and Learn
  2. Click Learn.
  3. Tick the box next to the test you want to cancel.
  4. Click Cancel test.

How do I see my results?

Learn when and where you can see your results, and how to interpret them, in our campaign comparison conversion lift test results guide.

* Nguồn: Facebook