Optimize your targeting methods and improve your ROAS.
Create a testing framework and list the testing scenarios to give your testing efforts a better structure.
Prepare a list of tests that you want to run. Test different audience types, sizes, and targeting criteria.
- Size of the lookalike audiences. For example, 1% lookalikes vs 3% lookalikes vs 5% lookalikes vs 10% lookalikes.
- Demographics. For example, Target males vs females, and Users aged 25-34 vs Users aged 35-44.
- Locations. For example, Females in Chicago vs Females in New York.
- Interests. For example, users with interest in luxury products vs. users interested in specific premium brands like Louis Vuitton, Cartier, Ferrari, and Prada.
- Audience types. For example, lookalike audience vs interest-based audience.
Go to Redirecting... and create the target audiences that you want to use in your experiments.
Name them so that you can easily recognize them in the campaign set up process.
Create a control ad set, targeting one of the target audiences that you would like to test.
- Go to Redirecting...
- Click on the green +Create button.
- Choose the campaign objective. If you want to drive traffic to your website choose Website traffic. If you track conversions with your Facebook Pixel choose Conversions.
- Name your campaign accordingly, for example, Lookalike audience test 1% vs 3%, and click next.
- On the Ad set settings page define the daily budget.
- In the Audience section click on Saved audiences and choose one of the audiences that you want to test.
- Click on Next and create your ads on the next page.
- Click Publish.
Create a variation of your control ad set, targeting another audience that you would like to test.
- In the Facebook Ads Manager, overview page click on the Ad sets tab.
- Tick the box next to your control ad set.
- Click Duplicate, select New A/B test.
- Under Variable select Saved audiences from the drop-down menu.
- Click the Continue to experiments button.
- Click on Create a Variant.
- Under Edit variable navigate to Version B column.
- Click Saved audience and select your test audience from the drop-down menu.
- Under Pick metric choose how you want to determine a winner. Cost per result or Cost per purchase are the recommended metrics.
- Click on Review test details.
- Click on Create test.
View the results of your tests and pick the winner.
Facebook will automatically send you an email with the results of the test that you conducted, once conclusive results are available. Go to Redirecting... and click on Results to monitor your tests on an ongoing basis and see interim results.
Follow up on your experiments and use the learnings to improve your targeting strategy.
Update your testing framework table with results to maintain a good overview of what is working well. Invest in audiences that generated better results in experiments. Create similar variations of the winning variations. For example, if your lookalike audience performs better than the interest-based audience, test a different lookalike audience size. If your interest-based audience performs better, test new audiences based on additional interests.