Test and Learn is an interface for creating advertising tests and analysing their results. You can use it to learn which strategies are effective and how much they're helping you. The tests take the form of questions that you want answered. Based on the question you pick, we provide a template for you to fill in with the information that we need to try to get the answer.

All tests are run with scientific rigour to measure "lift". Lift refers to results that wouldn't have happened without your ads. The more lift your ads are causing, the more effective they are. This method of measurement helps you understand the true value of your ads. It also ensures that the results you get are accurate and meaningful – including when they don't reflect positively on the money that you spend with Facebook. We want you to be able to use the answers you get to make improvements to your advertising strategy that grow your business.

Our tests can measure conversion lift (how many conversion events your ads are causing) and brand lift (the effect that your ads are having on brand perception). We offer the following tests:

• Which campaign causes the lowest-cost conversions to occur? This test compares two campaigns to see which one gets a lower cost per conversion. You can learn what factors improve performance and where your money will be spent most effectively.
• How much impact are all my Facebook ads having on my business? This test measures how many conversions are due to your ads. You can learn the overall impact of your advertising across the Facebook family of apps and services, and see what demographic groups you're reaching most efficiently.
• How much impact is my campaign having on brand perception? This test runs a poll to measure how much a campaign is influencing people's perceptions of your brand. You can learn how effective your campaign is at achieving your brand goals, and how that compares to the standards for your region and business type.
• Will setting a campaign budget improve performance? The test allows you to see if campaign budget optimisation leads to a lower overall cost per result compared to ad set budgets.

Can I run a test?

Anyone can run a test as long as they meet its prerequisites. For conversion lift tests, you need active campaigns and event tracking set up. For brand tests, you need an active campaign and a budget of at least $30,000. There are additional recommendations that we make to optimise your tests for meaningful results (you can learn about them in the articles on each test linked above), but they're not required to run one.

Should I run a test?

In general, we recommend running tests. Adopting a data-driven approach to advertising is key to growing your business. However, when trying to decide to run any individual test, you should make sure that you know what information you want and what you'll do once you have it. In other words, you want to test with specific intentions – not just for the sake of running a test. This helps guide your test setup and ensure that your results are meaningful and actionable.

Which test should I run?

It depends on what information you want to learn. If you want to learn what aspect of a campaign is the cause of more conversion lift, or which of two campaigns you should spend more on, run the "Which campaign causes the lowest-cost conversions to occur?" test. If you want to get a sense of the conversion lift caused by all your ads, run the "How many conversions are all my Facebook ads causing?" test. If you want to get a sense of the brand lift caused by a campaign, run the "How much effect is my campaign having on brand perception?" test.

Does it cost money to run a test?

No. Tests themselves don't cost money. However, you need to be spending money on the ads that you want to test for us to test them. The amount that you need to spend to get meaningful results varies from test to test (all brand lift tests have a $30,000 minimum; conversion lift tests have no required minimum), but be aware that spending more creates more test data. The more data we have, the more accurate your results will be.

Note: During a test, the cost per impression (CPM) of campaigns that are part of the test and use the lowest-cost bid strategy may increase. This is because, to generate meaningful results about the effect of your ads, we have to use hold-out groups during tests. A hold-out group is a randomised subset of people in your target audience who, during the test, won't see ads from the campaign that you're testing. Because the group is randomised, it'll contain lower-cost and higher-cost impressions, just like your target audience does. As not everyone in the hold-out group (including the people who cost less to reach) can be shown your campaign's ads during the test, we may have to go after higher-cost people outside the hold-out group to spend your budget. Bear in mind that we keep hold-out groups as small as possible to mitigate this potential effect.

* Nguồn: Facebook