So far we've seen how a conversion lift study can help to evaluate the effectiveness of an advertising campaign. In this video, we'll look at how a conversion lift study works in practice for your Facebook campaigns. Of course, you could run conversion lift studies for campaigns on other platforms as well, and it works in a very similar way. Let's go back to the example we talked about in the previous video. Irma at Calla&Ivy ran a campaign on Facebook focused on increasing sales for the fall bouquets. Irma wants to know what the incremental impact is of her campaign on sales. In other words, how much lift in sales will there be as a result of the advertising campaign? It's Irma's hypothesis that the advertising campaign will cause significant lift. This campaign aligned with her SMART goal of increasing the online sales of flower bouquets during the month of November with 15 percentage points compared to October sales. To be able to evaluate whether her campaign helped her achieve this goal, Irma decides to run a conversion lift test. While Irma and her team can set up and manage the advertising campaign themselves in Ads Manager, to run a conversion lift test they need the help of an account representative at Facebook. Who can connect them with the researchers at Facebook to set up an experiment. This enables the team at Facebook to make sure all the conversion API connections work well, and that people in the target audience can be randomly assigned to the test and control groups. There used to be a time where advertisers could set up a test like this without the help of the Facebook team. But with increased complexities in monitoring conversions due to privacy expectations, careful monitoring of the test is necessary to get unbiased results. Conversion lift studies need a bit of preparation, and you have to start the test when your campaign starts running, that's the nature of experiments. They're carefully planned ahead of time. As you prepare to work with Facebook team to get conversion lift studies started, there are a few things to keep in mind to make your test successful. First, you have to decide on the duration of your test. To decide how long the test should run for, it's important to ask yourself how long you believe it typically takes people to convert, or, in other words, what's the conversion cycle? If you want to see the influence of your ads on conversions, you have to make sure you have given people enough chances to convert. Irma believes that for her flower bouquets it typically takes people about a week on average to decide to buy. Facebook recommends that you run the test for one to two conversion cycles or at least two weeks. Irma decides on three weeks, just to be sure she's captured all the conversions. Then you have to make sure that sufficient budget is allocated to your campaign. If you want to see your ads have a significant influence on purchases, it's important that your ads reach your audience and are seen by that audience more than once. If you spend too little, chances are much lower that your campaign can make a real difference. Irma decides to set aside $50,000 for her campaign. Of course you also want the creative of your ad to be good, so it can have a real impact. It's important that your ad is clearly branded and that you include your call to action. For Irma, she added a Shop Now button, specifying the action she would like people to take when seeing her ad. To have a clean experiment, you want to make sure your audience doesn't see any other ads you may be running. If you're running another campaign targeted at the same audience, chances are that your control group gets exposed to an ad of yours. While it may be a different ad from the one you're testing for, it could still inflate the baseline. In other words, it may increase the chances that your control group takes the action you're testing for based on the other ad you were running. In that case, it will be harder to understand the effect of the campaign you're testing. For the same reason, Facebook advisors to consider pre and post study dark periods, where you don't run ads to avoid contamination. It's best to make that dark period be as long as you believe the conversion cycle to be. That way, if there's still an effect of a previous campaign you were running, it wouldn't contaminate the results for this test. Irma stopped the awareness campaign she was running three weeks before starting this shopping campaign. It goes without saying that the conversion event you choose should reflect your primary business goal. That's just good practice for any campaign. But you want to make sure that the conversion event for the campaign you're testing, aligns with the stage of the conversion funnel you believe your audience is at. It's hard to push your audience towards an action they are not ready for, like making a very expensive purchase if they've never heard of your brand for instance. In order for your advertising to be effective, you should use a call to action and optimize for a conversion event for which you can expect success. Without that, it will be hard to prove that you're advertising had any effect. Irma feels confident that her audience is ready to shop for her bouquets on her website, so she's aiming at the right conversion event. One final recommendation, don't edit your campaigns mid test or stop your campaigns from running all together before the test is over. It goes without saying that any changes you make while the test is running will contaminate the results of your experiment. Keeping all these recommendations in mind will help make your test successful. And of course, when I say make your test successful, I mean that your test will help you to evaluate whether your campaign was effective. It doesn't mean that the results will always be positive. Let's look at what the results of conversion lift tests look like in our next video.
