November 3, 2025
Ben Hamilton
Author
CPG brands need precise tools to measure marketing effectiveness and sales impact on an ongoing basis. Ibotta's innovative measurement solutions do just that, eliminating guesswork and helping brands maximize ROI with data-driven decisions that leverage the latest developments in machine learning (ML).
Challenges in the industry
Traditional methods of CPG ad measurement suffer from several challenges. First, cookies, pixels, or other digital indicators of conversion are insufficient for determining performance, as most sales occur in-store. Second, employing multiple tactics simultaneously – including media, trade spend, and promotions – makes it difficult to determine the degree to which each tactic led to an incremental sale. Finally, brands generally lack access to data with the granularity and availability required to leverage ML models for optimizing campaigns while they are still in flight.
Faced with these challenges, media agencies and brand teams traditionally rely on high-level econometric models, such as media mix models (MMM), to assign relative causation to different tactics based on a combination of prior assumptions and empirical data. These models serve an important purpose, but they are highly sensitive to their key assumptions, which are unique to their proprietary methodologies. MMMs often take months to prepare, leaving CPG brands waiting for measurement to inform their strategic planning. Responding quickly to changing market dynamics, therefore, becomes a challenge when relying primarily on MMMs.
Another common technique for ad measurement is the lift study. This methodology isolates the causal impact of ads by looking at longitudinal purchase data for a group of people exposed to the ad versus another statistically similar group that was not exposed, where all other variables are held constant. Third-party measurement companies run and provide these studies to advertisers once or twice a year, as a way of sense-checking the ROI of the ad spend. For example, Google, Facebook, Twitter, and other digital marketing platforms rely on lift studies to establish the incremental lift they generate with their ads. These studies, however, generally provide a point-in-time measure of performance vs. a rolling view of incremental sales while the campaign is live.
Ibotta has developed a new and sophisticated measurement approach that addresses these gaps by providing in-flight sales lift measurement, such that CPG brands can optimize performance while the campaign is running. By partnering with a network of trusted, industry-leading measurement providers, we've ensured our approach complements the tried-and-true methodologies outlined above.
Ibotta’s measurement breakthrough
Ibotta (NYSE: IBTA) operates the largest digital promotions network in the U.S., reaching more than 200 million consumers. The Ibotta Performance Network seamlessly distributes thousands of brand offers across a network of publishers, including leading retailers and delivery service providers such as Walmart, Dollar General, Instacart, Family Dollar, DoorDash, Schnucks and AppCard, as well as on the Ibotta app, which is one of the most widely used savings apps in the U.S. Because Ibotta has access to item-level data for a panel of millions of U.S. consumers, it has created a breakthrough system for measuring incremental sales on a rolling basis and delivering a Cost Per Incremental Dollar, or “CPID”. CPID provides visibility into how much was spent to drive one incremental dollar in sales, and the equivalized nature of the metric enables universal application across brand, price point, and spending tactic.
Using the well-established lift study methodology, Ibotta isolates a consumer’s exposure to the promotional campaign. This is done by taking statistically similar populations and dividing them into two groups: one that has been exposed to the campaign and one that has not, holding all other variables constant. Because both groups have an equal opportunity to be exposed to all other tactics live in the market, including media and trade investments, any statistically significant differences in purchase behavior over time are attributable to exposure to Ibotta content. In this way, Ibotta measures the incremental dollars directly attributable to the campaign, accounts for the fully-loaded cost of execution (including any fees, consumer rewards, or setup costs), and computes the cost per incremental dollar.
Rolling cost per incremental dollar measurement represents a breakthrough in the industry for three reasons.
CPG leaders and media agencies can have greater confidence in the accuracy of their measurement.
The effectiveness of tactics assumed to drive incremental sales is directly observed through scientific experimentation and then factored into future projections. Brands and agencies can track that performance on a per-campaign basis and decide whether to continue to invest, pause, or increase their budget commitments. As brands run more campaigns over time, Ibotta’s models will incorporate learnings to become even smarter, further enhancing brands’ ability to predict cost per incremental dollar and incremental sales volumes for new campaigns.
Brand managers and sales leaders can allocate resources in a more agile way, ensuring their dollars flow to the highest-performing channels.
In place of a rigid annual budget, CPG brands can gain a competitive advantage by becoming much more nimble and operating like true performance marketers. They will be able to set cost per incremental dollar targets that are consistent with profitable revenue growth, keeping campaigns on as long as they are delivering positive contribution margin. In this way, brands can better monitor their businesses day to day, gain valuable insights, and drive both top-line and bottom-line growth.
In other situations, sales leaders may prefer to use Ibotta’s platform to close anticipated sales gaps before a quarter’s end or ensure they win volume during important seasonal windows, accepting a higher cost per incremental dollar to maximize incremental sales. In the future, these “gap close” tactics may be triggered automatically, meaning sales leaders can begin closing sales gaps as soon as sales data shows a gap is emerging.
Machine learning models can be built on top of daily sales lift signals.
Much the same way that Google and Facebook have used ML to dramatically improve their delivery of search and social ads over time, CPG brands can now optimize each parameter of their promotional campaigns, including the targeting criteria, offer value, offer thresholds, UPC composition, and distribution logic. Ibotta is among the first to offer this degree of both automated and customizable fine-tuning to the realm of CPG pricing and promotions, ensuring that brands use every dollar in the smartest way possible. In the future, the Ibotta Performance Network Portal will allow brands to see and act on recommendations based on their specific goals and prior campaign results, while at the same time factoring in the latest market dynamics and trends in consumer behavior.
Methodology
Figure 1 below illustrates key elements of Ibotta’s cost per incremental dollar measurement approach. The yellow line represents the level of baseline sales that would likely have occurred if the campaign had not run. The curved line represents the increase in sales both during and after the promotional window. The area between the curves represents Incremental Dollars captured, both during and immediately after the campaign. Specifically, Ibotta takes into account:
- In-period sales. Incremental sales while the campaign is live and available to consumers.
- Follow-on sales. Incremental sales of the featured items in the 60 days after the campaign, where exposure to the campaign and/or trial of the product led to an ongoing statistically significant delta compared to the unexposed group.
- Halo sales. Sales of products not in the campaign, but that are in the same brand and category, and are driven by a statistically significant change in sales due to campaign exposure.
Measuring incremental sales
Purchase data
Ibotta has unique access to first-party, cross-retailer, item-level purchase data at scale collected from our direct-to-consumer and Ibotta Performance Network properties.
Ibotta receives purchase data from our popular direct-to-consumer mobile app.
Over the course of 12 years, Ibotta has built purchase data integrations with more than 80 different retailers, giving the company access to full basket, item-level data over time. This generally includes the Universal Product Code (UPC), quantity, and price of all the items purchased, along with the total amount paid, date, time, and store location. Whenever an Ibotta app user connects a loyalty account within the app, Ibotta begins receiving their purchase data, either via an API or a transaction log file. Ibotta’s data panel additionally includes purchase data extracted using optical character recognition from the millions of receipts that are uploaded each year by Ibotta’s app users. Ibotta tracked over 500 million unique transactions in 2024.
The publishers on the Ibotta Performance Network send purchase data to Ibotta regularly.
As part of providing white label promotional capabilities for retailers on the Ibotta Performance Network, Ibotta receives a large volume of item-level purchase data in order to verify purchases and keep track of campaign redemptions. This includes online and in-store shopping data from Ibotta Performance Network publishers, which Ibotta uses to inform its measurement models.
Tailored measurement for every Ibotta Performance Network partner
In order to comply with each publisher’s requirements regarding data sharing, Ibotta uses a distinct approach to measure incremental sales: Matched Audiences.
Matched Audience models
Full basket data provides Ibotta with the single best representation of consumer behavior and has driven our migration to Matched Audiences for calculating incremental purchases. This model removes the need for pre-defined “holdout” groups. One hundred percent of the target audience is eligible to view the offer. With Matched Audiences, the control group is created post-hoc by identifying which users were not exposed to the offer. Once that control group is defined, they are reweighted so that, on average, they are statistically identical to the group that was exposed to the offer (i.e., the treatment group). Therefore, the two groups become equal on important characteristics like shopping habits and past behavior, allowing Ibotta to accurately measure the true effect of the offer. The dynamic nature of these control groups enables faster reporting and statistically significant results, even for targeted campaigns.
Figure 2 below helps visualize how Matched Audiences leverage deep consumer data across multiple years of full basket purchases to ensure that the control group (Group A) equals the treatment group (Group B) in every way except being exposed to an offer.
Fig 2: For statistical balance, Ibotta controls for many attributes:
Summary
Ibotta’s rolling measurement of incremental in-store and online sales represents a significant milestone that Ibotta believes will rapidly change the way CPG brands allocate resources and maximize the efficiency of every dollar they spend. CPG leaders crave solutions that can deliver incremental sales at a scale large enough to move markets, even for the world’s largest brands. Ibotta delivers reliable sales lift measurement so our clients can begin to run their businesses like true performance marketers. Ibotta believes that the scale of the Ibotta Performance Network, combined with the new cost per incremental dollar measurement framework will allow these leaders to achieve results like never before.
Ready to drive incremental sales at scale?
Unlock the power of LiveLift today.