Brand Lift Study: What It Is, How to Run One and How to Read It

Brand Lift Study: What It Is, How to Run One and How to Read It

Brand Lift Study: What It Is, How to Run One and How to Read It

Written by

Written by

Patricio Luna

Patricio Luna

Published on

Published on

What Is Brand Lift?

Brand lift is the change a campaign makes in what people think of a brand: more of them remember the ads, know the brand, consider it, like it or plan to buy it. A brand lift study measures that change with a survey, by comparing people who saw the campaign with similar people who did not, or the same audience before and after the campaign.

Want to check this for your own idea?

Free to start. No card needed.

Clicks and views tell you that people saw an ad. They do not tell you whether the ad changed anything in their heads. Most brand campaigns are not meant to get a click today; they are meant to make people remember the brand and choose it later. Brand lift is how you measure that.

This guide covers the metrics a brand lift study measures, the three ways to run one, the questions to ask, how to calculate lift, and how many respondents you need before a lift is real and not noise.

Brand Lift Metrics

Metric

What it tells you

Typical question

Ad recall

Whether people remember seeing advertising for the brand

"Have you seen this advertisement for [brand] before today?"

Aided awareness

Whether people recognize the brand in a list

"Which of these [category] brands are you aware of? Select all that apply."

Consideration

Whether the brand made their shortlist

"Which of these [category] brands have you seriously considered purchasing in the last 3 months?"

Favorability

What they think of the brand overall

"Overall, how would you rate [brand]?"

Purchase intent

Whether they plan to buy

"How likely are you to purchase from [brand] in the next 3 months?"

Message association

Whether the campaign's idea is linked to the brand

"Which brand do you associate with [the campaign's main claim]?"

Choose two or three metrics that match the campaign's goal, and decide before launch which one counts as success. A campaign for a new brand should move awareness and ad recall; a campaign for a known brand should move consideration or intent. Measuring ten metrics and reporting the one that moved is how studies fool their authors.

Three Ways to Run a Brand Lift Study

1. Exposed Versus Control

People who saw the campaign (exposed) answer the same questions as similar people who did not (control). The difference between the two groups is the lift. When the two groups are chosen at random before the campaign runs, this is the most reliable design, because the only difference between them is the campaign.

The ad platforms offer this for their own ads: they hold back a random group from seeing the ads and survey both groups. These studies are well designed, but they measure only the ads on that platform, usually need a minimum spend, and leave the data with the platform. They cannot tell you what a TV, radio or billboard campaign did, or what all your channels did together.

2. Before and After

You run the same survey with the same audience definition twice: once just before the campaign starts and once just after it ends. The difference between the two waves is the lift. It works for any channel, including TV and outdoor, and it needs nothing more than a survey you can repeat exactly.

Its weakness is that other things change between the waves too: the season, a competitor's campaign, news about your category. Keep the time between the waves short, and if you track competitors in the same survey, check that their scores did not move in the same way as yours.

3. Split by Ad Recognition

You run one survey after the campaign and split the answers by whether people recognize your ad. Show the ad at the very end of the survey, after every brand question, and ask "Have you seen this advertisement for [brand] before today?" Then compare the brand scores of people who say yes with those who say no.

This is the cheapest design, and it works for any channel. It is also the easiest to fool: people who already like a brand notice its ads more, so part of the gap would be there without the campaign. Treat it as an estimate, and read it next to a before-and-after comparison when you can.

Which to choose: use the platform's own study for a large campaign on one platform, before and after for a campaign across several channels, and the recognition split as a check on either. For a campaign that runs only in some regions, you can also compare the regions that saw it with similar regions that did not.

Brand Lift Survey Questions, in Order

The order matters as much as the wording. Ask about the category and the brand before anything mentions the campaign, and show the ad, if at all, only at the end.

1. How often do you purchase or use [category] products? (Weekly or more / Monthly / A few times a year / Rarely / Never)

2. When you think of [category], which brand comes to mind first? ([brands] / A different brand / No specific brand)

3. Which of these [category] brands are you aware of? Select all that apply.

4. Which of these [category] brands have you seriously considered purchasing in the last 3 months?

5. Overall, how would you rate [brand]? (1 Very poor to 5 Excellent)

6. How likely are you to purchase from [brand] in the next 3 months? (1 Very unlikely to 5 Very likely)

7. Where have you seen [brand]'s advertising most recently? (TV / Streaming, Social media, Online / digital banner, Out-of-home (billboard, transit), Print / magazine, I haven't seen it before)

8. Have you seen this advertisement for [brand] before today? (Shown with the ad, last)

Question 1 is the screener: people who never buy in the category leave the survey. Questions 2 to 6 are the brand measures, worded exactly as in the SegmentOS brand tracking survey, so a lift study can sit inside your regular tracker. Questions 7 and 8 come from the ad testing template and measure exposure, so they go last.

How to Calculate Brand Lift

Brand lift is reported two ways:

  • Absolute lift: the exposed score minus the control score, in percentage points.

  • Relative lift: the absolute lift divided by the control score.

Take an example with illustrative numbers. Of 400 people in the control group, 30% consider your brand. Of 400 people who saw the campaign, 36% do. The absolute lift is 6 points; the relative lift is 6 ÷ 30 = 20%.

Report both, and say which one you are quoting. "A 20% lift" sounds larger than "6 points", and both describe the same result.

How Many Respondents a Brand Lift Study Needs

A lift is only real if it is larger than the noise in the survey. In the example above, a 6-point gap with 400 people in each group is not statistically significant at 95% confidence (a two-proportion test gives z = 1.8, below the 1.96 needed). The campaign may have worked, but this sample cannot show it.

Small lifts need large samples. As a guide, for a metric around 30%, at 95% confidence and 80% power:

Lift you want to detect

Respondents per group

10 points (30% to 40%)

About 360

6 points (30% to 36%)

About 960

3 points (30% to 33%)

About 3,800

Most single campaigns move brand metrics by a few points, not ten. Decide the smallest lift worth knowing about before you set the sample, and use the sample size calculator and confidence interval calculator to check your own numbers. If you split the results by age or region, each group needs its own sample.

When to Field a Brand Lift Study

  • Before and after: field the first wave in the days just before launch and the second as soon as the campaign ends, while people still remember it.

  • Exposed and control: field both groups at the same time, so the only difference is exposure.

  • Keep fieldwork short. A survey that stays open for weeks mixes people who saw the campaign recently with people who saw it long ago.

  • Keep everything else the same: questions, order, audience definition and sample source. A change in any of them can look like lift.

Common Mistakes in Brand Lift Studies

  • Asking about the ad before the brand. Once people have seen the ad in the survey, their brand answers are no longer clean.

  • No pre-set success metric. Choose it before launch, or every study finds something that moved.

  • Samples too small for the lift. A 3-point lift needs thousands of respondents per group.

  • Trusting a recognition split alone. Fans of the brand notice its ads more, which overstates the lift.

  • Changing the survey between waves. New wording or a new competitor list breaks the comparison.

Measure Brand Lift in SegmentOS

To run a before-and-after study, use the SegmentOS brand tracking study template for both waves: it asks questions 1 to 6 above in this order, with an attention check, and the same template keeps the wording identical. Add questions 7 and 8 from the ad testing template to the end of the second wave to split the results by ad recognition, using cross-tab analysis on the Premium plan. Both templates are Premium templates. To reach category buyers rather than your own customers, the Pro plan adds a verified panel of more than 30 million respondents in 127 countries, with the cost of each wave shown before you launch.

To test the creative before the campaign runs, rather than after, see our concept and ad testing survey template.

Key Takeaways

  • What it is: brand lift is the change a campaign makes in ad recall, awareness, consideration, favorability or purchase intent.

  • Three designs: exposed versus control (most reliable), before and after (works for any channel) and split by ad recognition (cheapest, least reliable).

  • The order matters: brand questions first, campaign and ad questions last.

  • Report both numbers: absolute lift in points and relative lift in percent.

  • Size the sample for the lift: a 6-point lift needs about 960 respondents per group; a 3-point lift needs thousands.

What do you want to learn?

Say it in a sentence. SegmentOS writes the survey, and real people answer it.

Free to start. No card needed.

In this article

No headings found on page

Stop guessing. Ask your actual buyers.

Stop guessing. Ask your actual buyers.

Build a study, reach verified respondents in your market, and get answers back in days.

Free plan available · No card required

Frequently Asked Questions (FAQ)

Frequently Asked Questions (FAQ)

What Is a Brand Lift Study?

It is a survey that measures how much an ad campaign changed what people think of a brand, such as ad recall, awareness, consideration or purchase intent, by comparing people who saw the campaign with people who did not, or the same audience before and after.

How Is Brand Lift Calculated?

Subtract the control group's score from the exposed group's score to get the absolute lift in percentage points. Divide that by the control score to get the relative lift. If 30% of the control group and 36% of the exposed group consider the brand, the lift is 6 points, or 20%.

What Is a Good Brand Lift?

Any lift that is statistically significant and moves the metric the campaign was meant to move. Most single campaigns move brand metrics by a few points, so the sample has to be large enough to detect a few points.

How Many Respondents Do You Need for a Brand Lift Study?

It depends on the lift you want to detect. For a metric around 30%, a 10-point lift needs about 360 people per group, a 6-point lift about 960, and a 3-point lift about 3,800, at 95% confidence and 80% power.

Can You Measure Brand Lift Without an Ad Platform's Study?

Yes. Run the same survey just before and just after the campaign, and compare the results. You can also show the ad at the end of the second survey and compare people who recognize it with people who do not. This works for any channel, including TV and outdoor.

Insights & Updates

Explore articles, resources, and ideas where we share updates about the product, thoughts on technology, and lessons learned while building along the way.

Insights & Updates

Explore articles, resources, and ideas where we share updates about the product.