Brand lift study and conversion lift study at a glance
A brand lift study measures whether a campaign changed what people remember or think. The brand awareness and lift guide describes how the platforms run them as their help centres document them: Google Ads Help describes Brand Lift for eligible video campaigns, in which short surveys are shown to people who were served the ads and to a control group that was held back, and Meta's Business Help Centre describes brand lift studies on the same principle. The questions cover ad recall, awareness, consideration, favourability or purchase intent, and the result is the difference in answers between the two groups. The guide covers the levels of awareness, tracking surveys and how to read a lift figure; this page is only about when a survey outcome is the right one to measure.
A conversion lift study measures whether a campaign changed what people did. The platform withholds the ads from a random share of the audience and compares conversions, such as purchases, leads or sign-ups, between the exposed and held-back groups. The incrementality testing guide places this design among the other ways to measure IncrementalityConversions caused by marketing. and covers sizing and reading; it is the study most performance teams think of when they hear the word lift.
The shared design is a strength. Both studies are randomised, both have a control, and both produce a difference with an interval. The difference in outcome is what decides between them. Memory can move quickly and can be asked about directly, but it is several steps away from revenue. Actions are revenue, or close to it, but they arrive slowly for anything that is bought rarely, and a short study window will miss most of them.
Compare table
Brand lift study vs Conversion lift study
| Row label | Brand lift study | Conversion lift study |
|---|---|---|
| Outcome measured | Survey answers: ad recall, awareness, consideration, favourability, intent | Actions the platform observes or receives: purchases, leads, sign-ups |
| Data collection | Short polls served to samples of the exposed and held-back groups | Conversion tracking already in place for the campaign |
| Sample needs | Enough survey responses per group to separate the answer from noise | Enough conversions per group, which is hard when the action is rare |
| Time to read | During or soon after the flight, because memory is asked about directly | After the conversion window closes, often weeks after the flight |
| Objectives it suits | Awareness, recall and consideration, usually for video and reach campaigns | Sales, leads and other actions inside a measurable window |
| Link to sales | Indirect: memory has to turn into purchases later | Direct for the actions counted, within the window measured |
| Platform availability | Offered for eligible campaigns, mostly video and awareness formats | Offered for eligible campaigns with working conversion tracking |
| Main bias | Stated answers from people willing to respond, read on the platform's own surfaces | Only conversions the platform can see, inside a window too short for slow effects |
Two rows usually settle the choice. The objective row says which outcome the campaign was bought to move, and the time-to-read row says whether that outcome can arrive inside the study at all. A campaign for a product people buy once every few years cannot show its full effect in conversions within a few weeks, however well it worked.
When to choose a brand lift study
Choose brand lift when the campaign's job is to be remembered. Video and reach campaigns that introduce a brand, reposition it or keep it front of mind are bought to change memory, and the honest test is whether memory changed. The brand versus performance guide explains why that kind of work pays back over a longer horizon than any conversion window, which is exactly why a survey outcome is the right one to read in the short run.
Choose it when the purchase is rare or slow. If most of the audience will not be in the market during the flight, a conversion study will find little whatever the campaign did. Recall and consideration can still be measured in everyone who was reached, so the study has an outcome to read.
Choose it to check creative and delivery early. A brand lift study shows within the flight whether people remember the ad, which separates a creative that did not register from one that registered with people who were not yet ready to buy. Read lift alongside the ReachUnique people who saw an ad. the study actually achieved, because a study run at low reach describes what the ads did to a few people rather than what the plan would do.
The limits are the limits of asking. People answer what they believe or what they are willing to say, the sample is drawn from those who respond to polls on the platform's surfaces, and a movement in intent is not a movement in sales. A brand lift result is evidence that the first step happened, not proof of the last.
When to choose a conversion lift study
Choose conversion lift when the campaign's job is a trackable action. Prospecting, retargeting and lead generation campaigns are bought to produce purchases or leads within a window, and the conversion study measures exactly that. Its result can be turned straight into an incremental cost per action or an incremental return, which is what a budget decision needs.
Choose it when the action is frequent enough to count. A store with many orders a week or a lead form that fills daily gives both groups enough conversions to compare; a business that closes a handful of deals a month does not, and should measure an earlier, more frequent action as a proxy or move to a brand study.
Choose it when the question is whether the platform's reported conversions are real. Conversion lift compares caused actions with attributed ones, which is the most direct check a performance team has on the numbers it optimises against.
The limit is the window. A conversion study sees only the actions that happen inside it and inside the platform's view. Anything that works by building memory, and pays back after the study closes, will look weaker than it is.
A mattress brand measures both sides of one campaign
Ambermoor Sleep, an online mattress brand, runs a six-week online video campaign at £60,000 with a brand lift study and, separately, a conversion lift study at £25,000 on its prospecting campaigns. The brand study reports an absolute lift in ad recall of 11 points and in consideration of 3 points among people reached. The video campaign's own conversion lift, read over the same six weeks, is too small to separate from zero, which the team expected: people buy a mattress every several years. The prospecting study shows 140 incremental orders at an average order value of £550, or £77,000 of incremental revenue, an incremental return of about 3.1. Ambermoor keeps both lines, judges the video on recall and consideration each quarter, and tracks whether the prospecting campaign's incremental return rises in regions where the video ran.
Decision table
| Situation | Choose | Why |
|---|---|---|
| Video or reach campaign meant to introduce or reposition the brand | Brand lift study | The campaign was bought to change memory, so measure memory |
| Prospecting or retargeting campaign with frequent purchases | Conversion lift study | The outcome is an action the window can contain |
| Product bought rarely, with a long consideration cycle | Brand lift study, plus a longer-run sales measure | Few buyers will act inside a short study |
| Lead generation with daily form fills | Conversion lift study | Enough actions to compare, and the result prices each lead |
| New creative whose impact is uncertain | Brand lift study first | It reads early and shows whether the ad registers at all |
| Board asks whether brand spend drives sales | Both, over different horizons | Brand lift shows the first step; sales evidence needs a longer window |
Sequencing the two across a campaign
The two studies fit different moments in a campaign's life. Run brand lift early on the upper-funnel activity, because it reads during the flight and tells you whether the creative and the delivery are doing their first job. If recall does not move, there is no point in waiting for sales; fix the creative or the reach first. Run conversion lift on the activity that is bought to produce actions, and read it after the conversion window has closed.
Connect the two through measures that sit between memory and action. Branded search volume and the tracking survey described in the brand awareness guide move after recall and before purchase, and they show whether the memory a brand study found is turning into demand that performance campaigns then capture. When awareness rises and sales do not, the troubleshooting page for awareness that is not driving sales works through the causes in order. For a cost view of reach-buying against action-buying, the comparison of CPC and CPM explains how the two are priced.
Report each study against its own objective. A brand study judged on sales, or a conversion study judged on recall, tells the reader nothing useful about either campaign.
Common mistakes
- Judging an awareness campaign on conversion lift over a short window, then cutting it before its effect can arrive.
- Reporting a modelled ad recall estimate as a brand lift result. One is an estimate inside the platform; the other is an experiment.
- Reading relative lift from a small base as a large effect, instead of comparing absolute lift across studies.
- Running a conversion study on a rare action and reading an underpowered result as proof of no effect.
- Ignoring reach in a brand study. A study run at low reach measures a small campaign, not the plan.
- Treating intent as purchase. A rise in stated intent is the first step, not the sale.
- Changing survey questions between studies, which removes the ability to compare one campaign with the next.
