What Is Conversion Lift?
by Andy Cooney
August 13, 2024
Conversion lift is a measurement technique used within digital advertising that determines the incremental impact of ads on conversions.
Why does incrementally matter? And why is it more powerful than digital attribution?
We will define incrementality before diving into conversion lift, which is often confused with digital attribution. Incremental conversions are directly attributed to the marketing or campaign being measured and would not have occurred if that campaign had not been live. This differs from digital attribution, which shares the total number of conversions across multiple touchpoints but does not consider that some of these may have happened without any marketing being live. Incrementality measures the true causal impact on the desired outcomes caused by your ads.
Incrementality is the most powerful tool in an advertiser's arsenal. It allows them to explain to their senior staff how the ads actually contributed to sales. This allows the advertiser to make informed budgeting decisions that maximise their ability to grow their business instead of risking overvaluing specific channels using attribution.
Both incrementality studies, measured through tests and digital attribution, have their place in any marketing strategy. However, all brands should strive to judge themselves on incremental returns.​​
How do you run conversion lift tests?
In a conversion lift study, you need to split your audience into two groups, with a test and control group, to show one group of ads and the other nothing. You must keep these groups as close as possible to one another when splitting them, minimising variables. You are trying to ensure that, as closely as possible, you can estimate the conversions that happened because the person saw the ad. It does not work if the groups are wildly different. To avoid contamination, you must ensure no overlap between the test and control groups.
In our experience, geographic splits are the most common and simple way for advertisers to create a test group.
​​
Best Practices for Running Geographic Conversion Lift Studies
The geographies should be very closely aligned, with a particular focus on:
Marketing behaviour
The geos should spend similar amounts on advertising across similar channels. If significant differences in channel spending impact the channel you are testing, it will be hard to get an accurate read on the conversions caused by the ad. For example, if you are trying to understand the incrementality of paid search, but one geo had significant TV spend and the other didn't, the TV ads would likely influence behaviour in paid search.
Sales Behaviour
During normal periods, one geo should not drive drastically more sales than the other. The geographies should have very close historical sales performance.
Demographics
From a consumer standpoint, the geos should ideally be as close to each other as possible. The test aims to expose similar groups to the ads or be in the control group, so very different audiences who react differently to the product or brand make significant results hard to reach.
The testing schedule
The testing schedule should be created for a period that will generate enough conversion data to be statistically significant whilst not running the study at a time when the periods would be incomparable.
Some examples of this might be:
1
Seasonal Periods
Running tests where consumer behaviour is likely to change significantly, such as Christmas or during gifting periods, can cloud results
2
New Product Launches
If a product is launched during a test, it will likely significantly alter conversion behaviour via paid ads.
3
Heavy Discounting
Similar to above, if a period of discounting happens during the test, it will muddy it severely
Minimum Detectable Effect
When a user clicks a hijacked ad on the search results page, they are silently redirected to an affiliate network or scam website. Because the click appears to come from the affiliate network, the visitor and transaction data is therefore misattributed: while the user came from a brand search keyword, their sale is attributed to affiliates.
​
​
Measuring the Test Impact
After you have run the test for a period long enough to gain statistically significant results, you take the test and control group and compare the conversions. The difference in conversions from people in the geography who saw the ads and those in the control group who didn't provide the conversion lift.
Lift will be defined by the percentage lift in absolute or relative terms. Absolute lift is the difference in conversions between the two groups. Relative lift is the percentage increase in the metric between the control and test groups.
Why Use Google Ads Conversion Lift Testing and How Does It Work
The Google Ads platform provides conversion lift measures directly within your ads account. This is significantly more convenient than running a study yourself. However, we must note it is a case of Google marking its own homework.
Google's conversion lift works in one of two ways: either via geography, as per the above, where Google will split ads between them, or via audience. This method randomises who sees your ads to provide a control group who would have been served your ad in the auction but are instead shown the next advertiser or don't see any ads.
Setting up conversion lift studies in Google
To set up a conversion lift study, go to your Google Ads account and click "tools and settings." Select "lift and measurement" in the measurement column and click the + button.
You can then choose the campaign, provide start and end dates and choose the conversion type you want to measure. This has to be a compatible conversion action, so not everything you might have in your account is possible.
Google provides you with a feasibility score, which, in effect, tells you how many conversions are needed to achieve a minimum detectable lift. You should aim for high scores here to maximise the chances of a successful study.
The study will automatically begin measuring lift on the selected start date. Depending on conversion volume, you'll be able to view metrics such as Incremental Conversions, Incremental Conversion Value, and Relative Conversion Lift.
Restrictions on Google Conversion Lift Studies
Not everyone can access conversion lift measurement from their Google Ads account. You should contact your account representative to see if they can activate it. For large advertisers with significant spending, we know their Google ads manager can provide support from teams that run the studies and help you measure incrementally. We'd definitely suggest seeing what Google can do for you.
Smaller accounts may not have the necessary site visits, ad spend, or conversion events to run a study that can detect the lift.
What is the difference between conversion lift and brand lift?
Brand lift is again a measure of incrementally but focused on measuring the incremental effect on brand metrics of people exposed to ads. This is available in Google via an account representative and is commonly used to measure the impact of video, display or Facebook ads rather than search.
The lift measurement is typically survey based on metrics such as:
-
Unaided brand awareness
-
Aided brand awareness
-
Consideration
-
Ad recall
-
Purchase intent
These metrics are typically associated with brand campaigns where the impact on sales would not be immediate, so conversion lift wouldn't be appropriate.
If you are going to run these tests, we suggest using purchase intent, consideration, and unaided brand awareness as primary brand metrics, as studies have proven that they impact sales.
Conclusion
Incrementality is a very powerful tool in the modern marketer's toolbox. It allows you to understand the direct impact your spend on advertising has had on sales via conversion lift or brand metrics via brand lift. Google Ads offers a simple way to help you understand how many incremental conversions are delivered by paid search.
