Quick answer: Conversion lift is the number of extra conversions your advertising causes, measured by comparing a group exposed to your ads with a control group that was held back from seeing them.
Conversion lift is reported two ways. Incremental conversions is the raw count of conversions the ads caused. Relative lift is that gap expressed as a percentage of the control group’s conversion rate. Both come from a controlled experiment, not from attribution rules.
Meta and Google both run conversion lift studies inside their ad platforms. On Meta, users in the chosen audience are randomly assigned to a test group that can see your ads or a control group that cannot. Google defines Conversion Lift as measuring the incremental conversions, site visits and other actions directly driven by your ads.
Why conversion lift matters
Your ad dashboard counts every conversion it can connect to an ad. It cannot tell you how many of those buyers would have bought anyway. Conversion lift is the only standard platform measurement that answers whether your ad spend created a conversion or simply sat next to one.
This matters most for retargeting, branded search and campaigns aimed at people who already know you. These campaigns often show strong ROAS (return on ad spend) because they reach buyers who were already close to converting. A lift test shows how much of that revenue the ads really caused, so you can move budget to the campaigns that create new customers.
Honest caveat: lift tests cost you something. The control group does not see your ads for the whole study, and small accounts may not have enough conversion volume to get a clear answer. Google states that Conversion Lift is not available for all Google Ads accounts and directs advertisers to their Google account representative.
Conversion lift vs attributed conversions
| Attribute | Conversion lift | Attributed conversions |
|---|---|---|
| Question answered | Would this conversion have happened without the ad? | Which touchpoint gets credit for this conversion? |
| Method | Controlled experiment with a held-back control group | Credit rules or models applied to tracked journeys |
| What it counts | Only conversions caused by the ads | All conversions linked to an ad interaction |
| Availability | Needs eligibility, volume and a set study window | Always on, in every campaign report |
| Best used for | Deciding whether to scale, cut or restructure a channel | Daily optimisation and reporting |
Attribution tells you where credit went, and conversion lift tells you whether the credit was deserved. Use both: attribution to run campaigns daily, lift to audit them.
How to calculate conversion lift
The core formulas are simple:
Relative lift = (Test group conversion rate – Control group conversion rate) / Control group conversion rate
Incremental conversions = (Test group conversion rate – Control group conversion rate) x Test group size
Incremental CPA = Ad spend / Incremental conversions
Worked example: a Mumbai coaching institute runs a Meta lead campaign and holds back 10% of the audience. It spends ₹9,00,000. The platform dashboard reports 3,000 attributed leads.
- Test group: 90,000 people, 3,600 conversions, a 4.0% conversion rate
- Control group: 10,000 people, 250 conversions, a 2.5% conversion rate
- Relative lift: (4.0% – 2.5%) / 2.5% = 60%
- Incremental conversions: 1.5% x 90,000 = 1,350
- Conversions that would have happened anyway: 2.5% x 90,000 = 2,250
- Dashboard CPA: ₹9,00,000 / 3,000 = ₹300
- Incremental CPA: ₹9,00,000 / 1,350 = about ₹667
In this example, 1,350 of the 3,600 test group conversions (37.5%) were caused by the ads, and the true cost per extra lead is about ₹667, more than double the ₹300 the dashboard shows. The 10% holdout and every number here are illustrative, not a benchmark.
For revenue goals, incremental ROAS (iROAS) is incremental conversion value divided by ad spend. Google’s geography-based Conversion Lift reports iROAS as incremental conversion value divided by incremental cost, with a confidence interval around the estimate.
How to run a conversion lift test
- Pick one conversion that maps to revenue. For lead generation, use qualified leads or booked calls rather than page views.
- Check eligibility and volume. Google’s Display & Video 360 guidance says campaigns should generate at least 3,000 conversions during the study period for statistically significant results. Other tools and markets set their own thresholds.
- Set the holdout and duration. Google’s Display & Video 360 studies run for a minimum of 7 days and a maximum of 56 days. Choose a window that covers your normal purchase cycle and conversion lag.
- Keep the control clean. Do not target the control audience from campaigns outside the test. Smartly’s Meta lift guidance warns that doing so makes the measured effect look smaller.
- Do not change the campaign mid-test. Budget, creative and audience edits muddy the comparison.
- Read the result with its confidence level. A lift figure without a confidence interval or significance status is a guess.
- Act on it. Shift budget toward campaigns with high lift and re-test the ones with low lift before cutting them.
If you cannot run a platform study, a geo holdout is the minimum viable alternative: pause ads in one or two comparable markets for 2 to 4 weeks and compare conversion rates against markets where spend continued.
Common mistakes
- Contaminating the control group. If control users still see your ads through another campaign, the gap shrinks and lift is understated. Exclude the control audience everywhere.
- Running the test too short or too small. Sparse conversion data cannot detect real lift. Confirm volume and feasibility before you launch.
- Measuring only the easy action. Google notes that upper-funnel and mid-funnel actions such as page views or lead forms show higher lift than lower-funnel actions. Pick the action tied to revenue, not the one that flatters the result.
- Treating one result as permanent. Lift changes with creative, audience, season and spend. Re-test after major changes.
- Ignoring a low or neutral result. Neutral lift means the campaign is not adding conversions at current settings. Change the creative, offer or targeting, then test again.
Frequently asked questions
What is conversion lift in advertising?
Conversion lift is the number or percentage of extra conversions caused by your ads, measured by comparing an exposed group with a held-back control group. Platforms such as Meta and Google run it as a controlled experiment and report it as incremental conversions and relative lift.
How is conversion lift different from attribution?
Attribution assigns credit for conversions among the touchpoints a customer saw. Conversion lift tests whether those conversions would have happened without the ads. A campaign can show high attributed ROAS and low lift when it mostly reaches people who were already going to buy.
How long should a conversion lift study run?
Google’s Display & Video 360 documentation sets a minimum of 7 days and a maximum of 56 days. Run the study long enough to cover your purchase cycle and conversion lag so late conversions are not missed.
How much budget or volume does a conversion lift study need?
Google’s Display & Video 360 guidance says campaigns should generate at least 3,000 conversions during the study period, and it recommends a budget of at least $5,000 USD. Thresholds vary by platform, market and study type, so confirm them in the tool before you plan.
Can a small advertiser in India run a conversion lift test?
Possibly, but access is not guaranteed. Google states that Conversion Lift is not available for all Google Ads accounts and directs advertisers to their Google account representative. If your volume is too low, a geo holdout, where you pause ads in comparable markets for 2 to 4 weeks, is the minimum viable alternative.
How Uginix helps
Uginix designs and reads lift and holdout tests as part of its performance marketing services, so budget decisions rest on incremental results rather than dashboard credit. If you suspect your retargeting or branded spend is claiming conversions it did not cause, book a free growth audit: no lock-in, no pitch deck. For platform-specific context, read our posts on Meta Advantage+ and Google Ads in 2026.
Related terms
- Incrementality: the extra outcomes your marketing causes beyond what would have happened anyway.
- ROAS: revenue earned for every ₹1 of ad spend.
- Holdout test: an experiment that withholds ads from a group to measure their effect.
- Attribution window: the period after an ad interaction in which a conversion can be credited to it.
Sources: Google Ads Help, About Conversion Lift, Google Ads Help, Conversion Lift based on geography, Gordon, Moakler and Zettelmeyer, Marketing Science (2023), Blake, Nosko and Tadelis, Econometrica (2015).