Marketing platforms are good at telling you which campaigns get credit for a sale. What they can’t always tell you is whether that sale would have happened without the marketing.

Brands can spend significant portions of their budgets reaching customers who were already likely to buy, while missing opportunities to influence customers whose behavior could actually change. Without a way to measure incrementality, it’s difficult to know which investments are generating new revenue and which are simply capturing demand that already exists.

This was the challenge facing a national QSR brand. After unifying its customer data through the Customer Intelligence Hub, the brand had the foundation to answer a more valuable question: which marketing activities were actually changing customer behavior?

Actable, an Ekimetrics company, helped the brand put that question to the test using predictive models and controlled experiments across search, loyalty, and retention campaigns. The results challenged some conventional assumptions about marketing performance. Excluding 1.27 million high-propensity customers from branded search targeting produced no statistically significant difference in order rates, pointing to an opportunity to improve advertising efficiency. Targeted loyalty offers increased purchase rates by about 15%, while revenue was approximately 8% higher than the control-group-based estimate.

The results gave the brand a clearer understanding of where marketing was making a difference and where spending could be reconsidered, providing a stronger basis for future decisions about targeting, promotions, and marketing investment.

For more on how the brand established its data foundation, read How a National QSR Brand Built a Unified Customer View in Four Months.

The Challenge: Which Marketing Investments Actually Drive Sales?

The brand had customer data and multiple marketing channels, but lacked a consistent way to distinguish customers who responded because of marketing from those who would have purchased anyway. That uncertainty affected advertising efficiency, retention campaigns, and promotional decisions.

Several conditions made the question hard to answer. Campaign performance came mainly from each platform’s own reporting, which did not always match the orders the brand actually recorded. Paid search, for example, counted conversions over its own seven-day lookback window. A consistent approach to holdout groups had not yet been established, so there was rarely a clean comparison between customers who were marketed to and similar customers who were not. Customer and campaign data sat across loyalty, online ordering, and messaging systems, and inconsistent campaign naming made results difficult to roll up across channels.

The brand also sent a high volume of messages, and every text carried a cost. Without a reliable way to measure incremental impact, it was difficult to know which messages, offers, and ads were earning their place.

The Solution: Predictive Models and Marketing Incrementality Testing

Actable built a testing program on top of the brand’s customer data, starting with predictive models. Each customer was scored on how likely they were to buy soon, how likely they were to stop buying, and how much they were likely to be worth over time. Those scores guided each campaign decision: who to reach, who to leave out, and which offer to send.

To measure what each campaign changed, Actable paired targeting with control groups. In each test, a group of customers was held back from the marketing, and results were compared with similar customers who received it, within the same predicted-score band. The results helped isolate the incremental sales attributable to each campaign, rather than relying solely on platform-reported conversions. They then fed directly into the brand’s next decisions about audiences, offers, and channels.

Actable ran each test through the same six-phase playbook: recommendation, measurement plan, execution, launch, measurement, and readout. The playbook was applied across search, social, email, and SMS, with the measurement method adapted to each channel. That consistency turned individual campaigns into a repeatable way to learn what works.

Incrementality Testing Across Search, Loyalty, and Retention

Three tests show how the approach worked in practice, and how each result shaped the brand’s next decision.

Branded Search: Identifying Where Ads May Be Unnecessary

The brand randomly split approximately 2.54 million high-propensity customers into two equal groups. One group of approximately 1.27 million customers was excluded from branded search ad targeting, while the other 1.27 million remained eligible to receive ads. Actable compared actual orders between the two groups. Based on the results, the brand expanded the exclusion to most of that audience and began testing the exclusion of its least likely buyers as well.

Loyalty Offers: Finding Incentives That Generate Incremental Revenue

Loyalty members who were not among the brand’s most likely buyers received one of two offers by email and push notification: $5 off or a free side. A control group in each score band received neither. The brand chose to lead with the offer that produced more incremental revenue.

Retention: Understanding Which Customers Respond to Incentives

Customers were scored on their risk of leaving and sent retention offers by email and SMS, with a control group in each risk band. The results led the brand to standardize on a single offer and begin matching offers to customer segments.

The Results: Measuring What Marketing Actually Changes

High-Propensity Customers Excluded From Search Ad Targeting

Customers in the excluded group ordered at 23.23%, compared with 23.13% for customers who remained in ad targeting (May 8 to June 21, 2026). The test found no statistically significant difference in order rates between the two groups for the high propensity group, saving hundreds of thousands of dollars in ad spend for higher incrementality tactics. A later review, after the exclusion was expanded, noted no material impact on conversion metrics. The results identified an opportunity to reduce unnecessary advertising spend by excluding customers whose purchasing behavior showed no measurable benefit from branded search ads.

 

Targeted Loyalty Offers Increased Purchase Rates by 15% and Revenue by 8%

In an April 2026 test, the $5-off and free-side offers together increased purchase rates by about 15% compared with a control group. Offer recipients purchased at 4.15%, versus 3.62% for the control group, a statistically significant difference. Revenue was approximately 8% higher than the control-group-based estimate. Revenue grew less than the purchase rate because buyers who received an offer spent slightly less per order.

Retention Offers Generated Stronger Results Among Lower-Risk Customers

The retention tests suggested stronger incremental results among customers at lower risk of leaving than among those at the highest risk. The most at-risk customers responded the least, which raised a new question: how to reach customers before they reach that point.

Not every test showed a lift. Results that came back flat were just as useful, because they helped the brand identify tactics and segments that warranted reconsideration before further investment. The results give the brand a measured view of incremental impact to weigh alongside platform reporting as it makes future marketing decisions.

What’s Next: From Individual Tests to an Ongoing Practice

In its second year, the brand is building these lessons into how it markets every day. Some of that work is already in place; the rest is planned.

Expanding Model-Driven Campaigns. The brand plans to double model-driven email, SMS, and loyalty campaigns over 12 months.  It also plans to move from one-off tests toward automated, always-on customer journeys that focus on driving topline business impact.

Connecting Campaign Results to Revenue. A campaign performance dashboard that reports email and SMS response is already live, and a version that adds revenue attribution has been built and is in review. Next, campaign results will feed back into the predictive models, and new reporting will show which coupons and promotions create incremental sales.

Testing New Capabilities Before They Scale. Planned pilots, including weather-based targeting and continuously optimized campaigns, will each be introduced as a controlled test before they scale.

Measure What Your Marketing Actually Changes

Most brands know what their campaigns report. Fewer know what their campaigns cause. Actable helps marketing teams measure the difference and invest where it counts.

Schedule a call to talk about bringing incrementality testing to your marketing.

 

Frequently Asked Questions About Incrementality Testing

What is incrementality testing in marketing?

Incrementality testing measures the additional sales, conversions, or revenue generated by a marketing activity that would not have occurred without it. By comparing customers exposed to marketing with a control group that was not, brands can better understand the actual impact of their campaigns rather than relying solely on platform-reported attribution.

How is incrementality testing different from marketing attribution?

Marketing attribution assigns credit for conversions to specific channels or customer interactions. Incrementality testing goes a step further by measuring whether those conversions would have happened without the marketing activity. Both approaches provide useful information, but incrementality testing helps brands understand which investments are actually changing customer behavior.

How can incrementality testing reduce wasted advertising spend?

Incrementality testing can identify customer segments that are likely to purchase regardless of advertising exposure. For example, a national QSR brand excluded 1.27 million high-propensity customers from branded search targeting and found no statistically significant difference in order rates compared with customers who remained eligible for ads. The result identified an opportunity to reduce unnecessary advertising exposure without a measurable decline in orders.

How do predictive models improve incrementality testing?

Predictive models help brands group customers based on factors such as purchase propensity, churn risk, and potential lifetime value. By testing marketing activities within these customer segments, brands can identify where offers, messaging, and advertising are most likely to generate incremental results. This allows marketing teams to make more informed decisions about targeting and budget allocation.

What marketing channels can benefit from incrementality testing?

Incrementality testing can be applied across paid search, paid social, email, SMS, and loyalty campaigns. The testing approach may differ by channel, but the objective remains the same: determine which marketing activities generate additional business value and use those findings to improve future investment decisions.