The cash-register receipt was supposed to be a disposable object: a curling strip of proof stuffed into a pocket, abandoned in a bag or offered immediately to the nearest trash can. Numerator sees something else. Each receipt is a compact account of demand - a timestamped list of products, prices and places. When enough shoppers voluntarily share those lists, the little scraps begin to answer questions that keep brand managers awake. Who tried the new flavor? Did a promotion create an extra sale or merely discount one that would have happened anyway? Where did a loyal household go after it stopped buying?
That is the practical idea behind the Chicago company. Consumers use Receipt Hog, a gamified rewards app, to upload paper receipts and connect other purchase sources. Different groups of those consumers form Numerator's panels. The company then layers purchase histories with demographics, psychographics, media habits and surveys. Its clients - packaged-goods makers, retailers, restaurants, consultancies, investors and public agencies - buy access through software, data feeds and research projects.
The useful reversal
Most consumer research begins with a question. Have you bought toothpaste recently? Which brand? How much did you pay? Human memory, always eager to help, supplies an answer that may be vague, polished or wrong. Numerator can reverse the sequence: establish what a panelist bought, then ask why. Its Verified Voices surveys target people through known purchase behavior rather than relying only on screeners and recollection. A study can reach current brand buyers, a competitor's buyers, people who lapsed or households that tried a product once and never returned.
This helps with the industry's familiar "say-do gap." A shopper may describe herself as health-conscious while her receipts show a weekly cookie habit. Neither signal is useless; together they are more revealing. The purchase says what happened. The survey can expose the compromise, occasion or aspiration around it. Numerator says its survey pool includes more than 600,000 verified buyers and can be targeted using more than 2,500 consumer attributes.
A product stack built on the same household
Numerator is best understood as several businesses sharing one underlying signal. Numerator Insights is the longitudinal workspace: teams study buyer profiles, loyalty, leakage, cross-shopping and brand performance with data updated daily. TruView approaches market share from the household outward. Instead of depending solely on sales reported by a subset of stores, it uses a measurement panel of more than one million active households to follow shopping across physical stores, ecommerce and less visible channels.
The distinction matters. Store sales can tell a cereal maker how many boxes moved through a participating chain. People-based measurement can show the types of households buying those boxes, the other stores they use and whether a share increase came from new buyers or heavier existing ones. It does not make point-of-sale data obsolete. It answers a different set of questions and can expose channels that a conventional store view misses.
Promotions Intel supplies another angle. Numerator tracks print, digital and temporary price-reduction activity, coding more than 70 promotional attributes across hundreds of retailers. Promo Insights connects that activity to later purchases. Data feeds put panel, market-share and promotion metrics into a client's own dashboards. Research consultants handle segmentation, brand health, concept testing and longitudinal studies, while NuQual adds video, photos, moderated discussions and interviews with verified buyers.
The grocery aisle is only the beginning
Consumer packaged goods remain the natural center of gravity. A shampoo, snack or detergent is bought often enough to create rich patterns, and its manager must continually negotiate price, placement, promotion and loyalty. Numerator says 80 of the top 100 U.S. CPG manufacturers are clients. Public examples across its materials include Unilever, Kraft Heinz, General Mills, L'Oreal, Diageo, King Arthur Baking and Albertsons. But the same panel can illuminate restaurant visits, consumer electronics, ecommerce, media preferences and the economic pressure on households.
That breadth makes the company useful outside marketing departments. A retailer can inspect which rival captures the trips it loses. An innovation team can recruit verified category buyers for a concept test. A media planner can identify the channels used by a growth segment. An investor can watch household penetration before a company's quarterly report catches up. Public researchers can examine how inflation or benefit-policy changes affect different groups.
A useful way to steal the model: recruit once, observe repeatedly, and let every new research product reuse the same consented relationship. The compounding asset is not a single survey. It is the history attached to a willing participant.
Where Numerator fits
The market around Numerator is crowded and old. NielsenIQ and Circana are major alternatives in retail measurement and consumer intelligence. YouGov, Ipsos and other panel companies compete for survey and audience work. Retailers increasingly sell access to their own first-party data. Each source has a different strength: census-like sales coverage at a chain, deep attitudinal research, fast polling or closed-loop media measurement.
Numerator's pitch is the connection across silos. Its panel watches households across retailers. Its research tools return to those same verified buyers. Its promotions data can be placed beside the transaction. The company offers software for analysts who want to work alone, feeds for data teams that prefer their own tools and consultants for questions that do not fit a template. That creates a subscription-and-services model rather than a single-product SaaS business. Enterprise pricing is private and the selling motion is built around demonstrations and longer contracts.
There are limits worth remembering even when the charts look precise. A panel is a sample, not the whole economy. It must be recruited, weighted and maintained. Receipt capture can miss purchases or misclassify obscure items. The people willing to share data may differ from those who are not. Numerator's 2026 Real-Time Attribution product uses AI to assign incoming items to brands and categories more quickly, but speed does not remove the need for quality controls. The company's advantage depends as much on panel stewardship as on software.
The work behind the clean chart
A photographed receipt is not analysis-ready data. It may be faded, folded, abbreviated or printed with a retailer's private shorthand. The useful work begins after capture: identify the merchant, resolve the line items, map each one to a product, brand and category, and decide how the household's activity should contribute to a weighted view of the market. Ecommerce records and loyalty connections arrive in different formats. Promotions have their own taxonomy. Survey answers must stay attached to the right participant without turning the analytical product into a directory of personal identities.
That operational middle is part of Numerator's expertise. The company employs engineers, data scientists, research consultants and a large data-operations organization. Its public technology recruiting materials describe a stack that includes Python, C#, Java, Airflow, Kubernetes and Django. The exact tools matter less than the recurring job: convert messy evidence from daily life into a stable series that a category manager can compare week after week. One incorrectly coded snack is trivial. Classification drift across millions of trips is a business problem.
Clients then need interpretation. A falling repeat rate might signal disappointment, but it could also reflect seasonality, distribution loss, a price increase or a competitor's promotion. Numerator's software allows an analyst to explore those branches; its consulting teams design custom work when a menu of reports is insufficient. This blend of data production, analytical software and human research is why the company resembles both a technology vendor and a modern descendant of the market-research agency.
From Chicago to a larger map
Numerator's corporate history reads like consolidation viewed through a kaleidoscope. InfoScout, founded in 2011 by Jared Schrieber and Jon Brelig, pioneered smartphone receipt capture. Market Track brought decades of promotion and advertising intelligence. The businesses joined under Vista Equity Partners and emerged publicly under the Numerator name in 2019. Kantar acquired Numerator in 2021 for roughly $1.47 billion in total consideration.
Then the structure changed again. In 2025, Numerator combined with Kantar's Worldpanel operation. Numerator now says it is independently run with Bain Capital backing, while Worldpanel has become Worldpanel by Numerator. The enlarged company operates in more than 50 countries and reports about 5,400 employees. It launched Receipt Hog in Germany in May 2025; by June 2026 the German survey panel had reached 50,000 monthly active users.
Global scale expands the questions Numerator can answer, but it also raises the stakes. Consumer consent, data sovereignty and privacy practice vary by country. A permissioned data model is not a permanent permission slip. The exchange has to stay legible and worthwhile to the household supplying the signal. Receipt Hog's rewards make that exchange visible in a way the research industry sometimes does not.
AI on top, evidence underneath
Numerator has added AI-generated survey questions, summaries of open-ended responses, report narratives, recommended next analyses and faster product attribution. Those features lower the cost of navigating a large analytical system. A marketing manager who does not speak fluent crosstab can get to a plausible starting point sooner.
But AI is not the most defensible part of the company. Competitors can build summarizers. The harder asset is the opted-in history underneath - years of purchases tied to stable households, followed by the ability to ask those households a new question tomorrow. The receipt remains the quiet protagonist. It is mundane, specific and stubbornly factual. In a research market worried about bots, synthetic respondents and confident nonsense, that is a useful personality.