Modern interior featuring a symmetrical floating staircase with glass railings, large windows on both sides, and sunlight casting shadows on the floor.

Conversion lift studies

A conversion lift study measures a campaign's causal impact by comparing a group of people who saw an ad against a matched group who didn't, rather than assuming that anyone who converted was influenced by it. 

A publisher's exposure data is compared against an advertiser's or retailer's conversion data inside a data clean room, so the resulting figure reflects a causal test rather than a correlation between exposure and purchase.

Abstract design of radiating lines and geometric shapes on black background, suggesting secure data collaboration.Abstract geometric design with circles and arcs representing secure data collaboration by Decentriq.

How it works

Before the campaign starts, the publisher splits its audience into a test group eligible to see the ad and a control group held back from it, then provisions exposure data for both groups into the clean room, typically updated daily. The advertiser uploads its own conversion data separately, during or after the campaign, and matching runs on common identifiers inside a clean room built on confidential computing, so the operator has no access to either party's raw data during the match.
‍
The result includes a match rate, the lift itself as a percentage difference between test and control conversion rates, and often a breakdown by cohort such as product line or region. Because the analysis runs on a pre-built template rather than custom queries, marketing teams can run a study without needing a dedicated data science team. Any result built from too small a group is automatically withheld before it leaves the environment, and once the study ends, both parties' data is deleted. Read the full guide on conversion lift measurement.

See it in action: Conversion lift in a Decentriq data clean room

See conversion lift on the Decentriq platform in action, showing how a test and control group are measured inside the clean room.

Frequently asked questions

How does this differ from first-party conversion lift products offered by ad platforms?

Platform-native conversion lift tools, like those built into Google Ads, Meta, or Spotify, work by having the platform hold both sides of the equation itself: it serves the ad and also sees the conversion, typically through its own pixel or Conversion APIs. A clean room-based study is built for the opposite situation, where the advertiser's conversion data lives outside the platform (in its own CRM, point-of-sale system, or e-commerce backend) and can't or won't be shared with the ad platform directly, often for regulatory, competitive, or governance reasons. Exposure and conversion data come from two separate parties, matched inside a neutral and transparent environment, rather than the platform measuring and grading its own performance.

What are the eligibility criteria to run a conversion lift study?

A few practical requirements: enough audience scale for the expected effect size to reach statistical significance, smaller expected lifts need larger audiences; a way to match the advertiser's conversion data to the platform's exposure data, typically via a shared identifier like a hashed email; and a campaign that can be reasonably isolated from heavy overlapping exposure on other channels, since that overlap can blur the result.

What type of outcome can be measured in a conversion lift study?

Any conversion event the advertiser can supply into the clean room, not just an online purchase, in-store visits, offline sales, app installs, sign-ups, or other business outcomes. The result includes the lift percentage, the estimated number of incremental conversions, a statistical significance check, and often a breakdown by cohort such as product line, region, or customer segment.

What happens if the advertiser does not collect outcome data, e.g. for CPG brands?

CPG brands often don't hold their own conversion data since the sale happens through a retailer rather than directly with the brand. In that case, the conversion side of the study would need to come from a third party that does hold it, a retailer's loyalty or point-of-sale data, or a syndicated sales data provider. Decentriq has partnerships with sales data providers in selected markets to offer this out of the box.

View Other Measurement use-cases

Attribution

Connect campaign exposure to customer outcomes without exposing either party's raw data

explore attribution ->

Cross-media reach & frequency

See deduplicated reach and frequency across publishers and channels while protecting exposure logs

explore cross-media reach & frequency ->

Incremental reach

Identify the audience a channel (e.g. CTV) uniquely reaches beyond another channel (e.g. Linear TV)

explore incremental reach ->

Script lift (pharma)

Pharma-specific version of conversion lift, built for US brands.

explore script lift ->

Audience quality

Pharma-specific version of reach quality, built for US brands.

explore audience quality (pharma) ->
Person in low light with colorful screen reflection; conveys focus and modern technology in data collaboration.

Talk to us about conversion lift studies