Introduction
No single data source sees the whole consumer, and even government statisticians work around the gaps. The US Bureau of Labor Statistics notes that some food items have significant sales in outlets that do not scan purchases or do not supply data to scanner companies, with convenience stores as one example. Every source bankers use has a similar blind spot somewhere, which is why the strongest analyses never rest on one dataset.
Consumer bankers draw on four main families of data. Scanner data shows what sold in which stores. Household panels show who bought and whether they came back. Card transaction data tracks spending at retailers and restaurants almost in real time. And foot traffic data counts visits to physical locations. Each answers a different question, and knowing which one to trust for which question is a core skill in diligence, pitching and sale processes. How these sources fit into the job is introduced in the article on what makes consumer and retail different.
Scanner Data: What Sold, and Where
Retail point-of-sale data, collected from checkout systems and sold by Circana and NielsenIQ, is the backbone of consumer products analysis. It reports units, dollars and average prices by product, brand, category, retailer and week, and it shows how much of each sale happened on promotion.
Its strengths are breadth and timing. A banker can see a brand's market share by retailer and channel, split growth into price and volume, measure how quickly a new product is selling and compare a client with every competitor in its category, weeks before any company reports. That makes scanner data the standard way to test a seller's growth story or a buyer's synergy case. Its blind spots are just as real:
- Unmeasured retailers: some retailers share data only partly, through separate arrangements, or not at all, so a brand that sells heavily through them looks smaller than it is.
- Partial online coverage: e-commerce sales are tracked less completely than in-store sales, and coverage varies by platform.
- No shopper identity: scanner data shows what sold but not who bought it, how often or what else they bought.
A brand with large sales through untracked channels can therefore show a declining share in scanner data while its total sales grow. Bankers reconcile scanner data with company shipments before drawing conclusions, a step that matters most for brands with a large club, convenience or online business. How analysts use these figures day to day is described in the article on a day in the life of a consumer analyst.
Household Panels: Who Bought, and Did They Come Back
Panels follow the same households over time, which lets analysts see behavior that store data cannot.
- Household Panel Data
Purchase records collected from a fixed, demographically balanced group of households that report everything they buy, by scanning barcodes at home or uploading receipts. Panel data shows who buys a product, how often, how much and from which retailer, and it is projected to the whole population using the panel's demographic weights.
NielsenIQ's Homescan panel covers more than 250,000 households across 25 countries, and other providers such as Numerator build panels from shoppers' receipts. Panels answer questions that matter for valuing a brand:
- Household penetration: what share of households bought the brand in a year.
- Buying rate and frequency: how much and how often each buyer purchases.
- Repeat rate: what share of first-time buyers come back, the key test for a new brand.
- Switching: which brands buyers came from and went to.
- Demographics: which income, age and household groups drive growth.
The weaknesses are sample size and bias. A small brand may have only a few hundred panel buyers, too few to read reliably by retailer or region, and panels can under-represent some shoppers and channels. Panel data is best used for trends and behavior, with scanner or company data used for the absolute size of sales.
Card and Transaction Data
Card transaction data comes from anonymized credit and debit card purchases, aggregated by providers such as Bloomberg Second Measure and Consumer Edge, which acquired Earnest Analytics in 2025. It reports how much shoppers spent at each retailer, restaurant chain or online store, often with only a few days' delay, and it can show the number of transactions, average ticket and how customers split their spending across competitors.
Card data is especially useful for retailers and restaurants, where it gives a near-real-time read on sales before a company reports and helps measure cross-shopping between rivals. It is less useful for packaged goods brands, because it captures the total at the checkout rather than the items in the basket. Its blind spots include cash purchases, uneven coverage across card issuers and regions, and panel skew toward particular income groups. Public market investors have used it for years; bankers increasingly use it to test sales trends in sale processes for restaurant chains and retailers.
Foot Traffic and Location Data
The newest major source estimates visits to physical locations from mobile phone data.
- Foot Traffic Data
Estimates of the number of visits to stores, restaurants and other locations, built from the location signals of a sample of mobile devices and extrapolated to the whole population. Foot traffic data shows visit trends, visit frequency, dwell time and the areas visitors come from, but it measures visits rather than sales.
Placer.ai, one of the largest providers, describes data drawn from tens of millions of mobile devices through partner apps, covering about 7% of the US population, with a proprietary algorithm that extrapolates panel visits to total visits. Bankers use foot traffic to compare traffic trends across a retailer's stores, measure the trade area of a location, judge how new stores are ramping and check whether a chain is gaining visits from competitors. Its limits are that visits are not purchases: a busy store with small baskets and a quiet store with large ones look the opposite of their sales.
Foot traffic is most valuable for new store ramps and competitive questions that companies rarely disclose, such as whether a rival's opening nearby took visits from a client's store.
Using the Sources Together
Each source answers a different part of the question, so bankers combine them:
| Source | Best for | Main blind spots |
|---|---|---|
| Scanner data | Share, price and volume by retailer and week | Unmeasured retailers, partial online, no shopper identity |
| Household panel | Penetration, repeat, switching, demographics | Small samples for small brands, panel bias |
| Card data | Retailer and restaurant sales trends, cross-shopping | No item detail, cash excluded, panel skew |
| Foot traffic | Visit trends, trade areas, new store ramps | Visits are not sales, modeled estimates |
In diligence, the combination is what gives confidence. A restaurant chain claiming traffic gains can be checked against card transactions and foot traffic; a brand claiming share gains can be checked against scanner data by retailer and panel data on new households. When the sources agree, the story is credible; when they disagree, the disagreement usually points to the question worth asking. The diligence process that uses these tests is covered in the article on consumer due diligence, and the store metrics they feed in the article on comparable sales.
Data also has its own legal and practical limits. Syndicated data is licensed, so its use in a sale process depends on the terms the company or the bank holds, and personal data rules restrict how granular location and transaction data can be. Within those limits, the analyst who knows each source's strengths and gaps can build a clearer picture of a consumer business than any single report provides.
