.png){kind=icon}
Which is your most effective campaign on Facebook? Not only do some marketers and businesses not know the answer to this question, but they also don’t even know how to find it. In a world where data is ubiquitous and numbers cause confusion, businesses struggle to identify their most effective campaigns.
Thankfully, Facebook has made this process easier for everyone with the addition of the Test and Learn tool. Today, we’re going to learn all about the tool and how you can implement it in your business. After this, we’ll also explore other options you have within the Facebook Experiments umbrella!
Head into the Ads Manager and open the side menu; in the Measure and Report column, you’ll see ‘Test and Learn’. As soon as you click on the link, Facebook will ask two questions, and these form the basis for the two different tests. While one looks at the conversions caused by Facebook ads in general, the other assesses the performance of individual campaigns. The two questions are:
The first question essentially looks at the conversions caused by all your ads across Facebook, Audience Network, and Instagram. As a business, this is designed for a big-picture review. For example, you might want to learn the impact that ad campaigns have on purchases or lead generation. As an overall assessment, you’ll quickly learn how much Facebook contributes to the wider marketing strategy. If the impact is negligible, it’s time to reevaluate how you use the platform.
Alternatively, the second question digs deeper into the performance of individual campaigns. Focusing on purchase events, for example, you might compare two different campaigns to measure their efficacy in this regard. Elsewhere, some companies use ads as a lead magnet and this feature helps to compare two campaigns and how cheaply they generate these leads.
For those who use conversion reporting, you might think that Facebook is just introducing unnecessary features. However, the difference is that this new feature uses the conversion lift measurement.
For example, the process starts by splitting a target market into two groups for ad account testing and four groups for campaign testing. At an ad account level, you’ll have a test group and a control group. Alternatively, campaign testing has two test groups and two control groups.
Either way, you choose a conversion event and Facebook measures between the test and control groups. From here, it’s easier to see the impact of your campaigns or your account. When using Ads Manager metrics, you need to base it all around an attribution window. Test and Learn is different because it focuses solely on the number of people who saw the ad and the conversions that came as a result.
In this section, we’re going to show those who want to learn the impact of their Facebook strategy how to create an ad account test. We mentioned the two questions popping up after clicking on Test and Learn, and the next step is to click ‘Set Up Test’ on the ad account question. At this stage, you give the campaign a name and choose specific conversion actions you wish to track during the test.
On this page, you’ll also choose a schedule; this includes a start and an end time for the test. Be careful - schedule the test to finish too early and you won’t generate the required 100 conversion events for the test to work. By default, the test will schedule for 28 days but you can change this depending on your ad spend. For smaller businesses, you might need a longer test to achieve the 100 conversion events.
When you’re ready, you’ll have a chance to review the test and start.
This time, we’re helping those who want to compare the performance of multiple campaigns rather than the Facebook ad account itself. Once again, choose Set Up Test next to the appropriate question and choose a name. At this stage, the process differs from the ad account test because you’ll choose two campaigns to compare. Facebook has clearly thought about the user here because you only need to click a drop-down menu and find the right campaign.
Ultimately, this is your test, and you decide which two campaigns to compare. However, one tip is to keep the campaigns quite similar. Otherwise, it’s hard to draw insights at the end. For instance, one common example is to compare campaigns that only differ in creative or objective.
As we saw with ad account testing, you’ll then choose a conversion event. While some choose offline events, others choose specific app events or something from the Pixel. Either way, keep clicking the Add Event button until you’re happy (the limit is 10 events).
It’s one thing performing a test, but this effort goes to waste if you ignore the next stage. When first opening the Test and Learn feature, you’ll notice that the page is actually split into two. At the top of the page, there are two tabs labeled ‘Test’ and ‘Learn’. After creating the test, the Learn tab is blank (you can’t expect something to happen immediately!).
As soon as 100 conversion events occur, Facebook splits all results into two sections:
Hopefully, you’ll notice a difference between the campaigns and have some momentum to make changes and improve specific campaigns. If you tested the whole ad account, you’ll finally learn how much Facebook contributes towards marketing goals. This is a chance to put a tangible number on the effect of Facebook.
Over time, Facebook continually expands the Experiments section of the platform, and this is great news for businesses (hooray!). For example, all businesses now have access to the following experiments:
Have two creatives and don’t know which is best for your campaign? Equally, you might want to compare the performance of campaign objectives, copy, audiences, or another feature of the campaign. As long as the two campaigns have the same budget, A/B testing means you can continually test ad features and identify the best combination for the audience.
Depending on your goals, you need to choose the most pertinent metric for success. Do you care more about the cost per conversion lift or cost per result? Follow all the instructions Facebook has to offer and ensure the test power is above 80% before starting - if lower, change the length of the test or budget.
For this feature, you’ll need to have spent over $10,000 in the last three months. If so, surveys are an effective way of learning whether people that have seen your ad remember it and know your company. Once again, Facebook is user-friendly in this area and guides businesses through the process. You can choose poll questions from drop-down menus and preview everything before sending the survey out.
Finally, this is an opportunity to see how your advertising efforts compare against each other. Essentially, the test keeps your ads from 10% of the target market. Then, it compares those who see your ads to those who don’t. This test is split into three variations:
As always, you’ll need to define success by choosing a conversion event. As soon as you have 100 conversions, you’ll see data explaining the difference between the impact of your ad and those who don’t see your ad in the audience.
If you aren’t already using Facebook Experiments and Test and Learn, now is the time to start. Before that, however, make sure that you encourage normal advertising conditions when testing. If you don’t keep things as normal as possible, you won’t get accurate results in return. With this in mind, the time of year is critical - don’t run a test during a major holiday because this isn’t representative of the rest of the year.
Also, focus all your efforts on one test at a time because this prevents overlaps in data and testing. At the same time, remember your goals and create tests that help your business to move forward rather than just copying tests online for the sake of it.
