Why financial media networks struggle with measurement, and what the public results show

A card issuer sees what its customers spend on its cards. It does not see the same customers' purchases on a rival's card, in cash or by bank transfer. The Federal Reserve's diary survey found that credit and debit cards together carry about two-thirds of consumer payments (Federal Reserve Financial Services, 11 May 2026). The rest never reaches a card network at all.

Even inside card spending, one issuer holds a slice. Chase's $1.94 trillion of annual card sales, set against $11.50 trillion of US card payments in 2024, is roughly 17% (Chase; Federal Reserve). That is our calculation. The two figures use different years and definitions, so treat it as an order of magnitude. A shopper who buys groceries on a Chase card and fuel on another bank's card looks like a lost fuel customer to Chase.

Coverage matters for two reasons. First, sales reported by a network are a floor, not the brand's total effect. Second, coverage is not stable. When a bank leaves a multi-bank platform, its customers disappear from the data, as Cardlytics' 43.7 million drop in monthly qualified users shows (10-Q, Q3 2026). Year-on-year comparisons on such a platform need the same panel of banks in both years.

A card transaction names a merchant and an amount. It does not list the items in the basket. That suits a restaurant chain or a retailer measuring store visits, but a brand that sells through retailers sees only "purchase at a grocery chain". It cannot tell its cereal from another brand's.

Matching the merchant is harder than it sounds. Each store location carries its own merchant identifier on each card network. Card-linking provider Fidel API, for example, tracks each onboarded identifier with network-specific IDs and onboarding dates (Fidel API docs, accessed 9 November 2026). A missing location means missing sales. Franchises, marketplaces and wallet payments add more ways for a purchase to be misfiled.

Product data must come from somewhere else. Cardlytics built a second platform, Bridg, on point-of-sale data that includes product-level purchases. In January 2026 it agreed to sell substantially all of Bridg's assets to an affiliate of PAR Technology (Cardlytics 10-K). For consumer packaged goods brands, that leaves card data as a store-level signal unless a retailer adds item data.

Card-linked offers make attribution easy: the purchase that earns the reward is the record the report counts. Card-linked offers are explained step by step in card-linked offers explained. The pricing often follows the same record. Cardlytics' main model charges a percentage of all purchases by customers who were served an offer during the campaign, "regardless of whether" they activate it (10-K). Chase charges brands "only when a customer views a campaign and makes a purchase" (Chase).

Neither rule asks whether the advertising caused the purchase. The IAB and IAB Europe guidelines put it plainly: attribution and ROAS "show what happened, not whether marketing caused the result" (IAB and IAB Europe, 3 November 2025). The same guidelines rate new-to-brand share, exposed-versus-baseline comparisons and platform-reported incrementality as "hybrid proxies" with weak causal strength. Incrementality needs a credible counterfactual, such as a randomised holdout.

A worked illustration. The numbers below are invented to show the arithmetic; they describe no real network.

The campaign is still profitable at 1.25x if margins allow, but it is not a 10x channel. The gap is what the IAB means by the difference between attribution and incremental measurement. Offers to existing customers are where the gap is widest, because many of them were going to buy anyway.

Public claims rarely say which number they report. Mastercard claims its network produces up to 22 times return on ad spend in verticals such as retail and travel (company claim, as reported by Marketing Dive, 1 October 2025). PayPal's June 2026 release is more explicit, labelling one result "incremental" and another simply "ROAS" (PayPal, 22 June 2026). Ask for the label before you compare.

A card payment happens in two steps. Authorization approves the purchase in seconds. Clearing, which moves the money, typically starts when the merchant's bank collects the day's transactions, then follows that bank's processing schedule (Fidel API docs, accessed 9 November 2026). Returns and chargebacks can reverse a sale weeks later. Rewards trail further: Cardlytics says it generally pays consumer incentives one to four months after redemption (10-K).

So a mid-campaign report is preliminary. A network that optimises on authorizations will count some sales that are later returned; one that waits for clearing reports later. Neither is wrong, but the report should say which it uses. The IAB/MRC retail media guidelines require attribution windows to be disclosed before a campaign runs. They also require the basis of any forecast that extends short results to 12 months to be disclosed (IAB/MRC, January 2024). How those guidelines apply to commerce media is covered in retail media measurement standards explained.

Banks cannot hand advertisers a customer file. Regulation P bars a financial institution from disclosing a card or account number to a nonaffiliated third party for marketing, with narrow exceptions (12 CFR 1016.12). Other customer data moves only under the notice, opt-out and service-provider rules covered in GLBA and advertising with bank data.

Networks design around that. Cardlytics says it analyses purchase data "without any personal data leaving" its bank partners and that the data it receives is anonymised (company statement) (10-K). The advertiser gets aggregate reports. It cannot join them to its own loyalty or e-commerce data at person level, so it cannot check the network's work directly.

Two responses are emerging. PayPal launched a measurement partner program with 20 certified third parties, including Circana, Experian and Kantar, so brands can validate results "with their own source of truth" (PayPal, 6 January 2026). Clean rooms let both sides match records under agreed thresholds, and the IAB/MRC guidelines require transparency in how such third-party matching is done (IAB/MRC, January 2024).

The table lists what four networks have published. Filings are audited or regulated disclosures; everything else is a company claim.

Two patterns stand out. Of the four, only Cardlytics publishes revenue for its network, in regulatory filings; the network profiles list the others' revenue as n/d. And only one result in the table is labelled incremental, with no published control design.

Chase Media Solutions revenue. We found no revenue figure for Chase Media Solutions on its own pages or in the other sources reviewed for this page on 9 November 2026. Its public numbers describe merchant spend driven by offers, which is the merchants' sales, not Chase's income. The Chase Media Solutions profile tracks any disclosure.

1. Ask for the counterfactual. Which customers were held out, how they were chosen, and how many. No holdout means the result is attribution.
2. Ask for coverage. What share of category spending the data sees, and whether the panel of banks or cards changed during the period.
3. Ask what was matched. Merchant locations, product data if any, and how missing locations are handled.
4. Ask when the numbers are final. Authorization or clearing, return windows and restatement rules.
5. Ask who else can check. Independent partners, clean room access and the thresholds that apply.

These five questions are part of the measurement group in our commerce media platform RFP, which scores vendor answers.

1. Board of Governors of the Federal Reserve System, "The 2025 Federal Reserve Payments Study: National payment volumes, top-line data (CY 2015-24)" (1 July 2026). https://www.federalreserve.gov/paymentsystems/frps_cy2015_24_topline.htm
2. Federal Reserve Financial Services, "2026 Diary of Consumer Payment Choice: consumer habits hold steady as payment options grow" (11 May 2026). https://www.frbservices.org/news/press-releases/051126-findings-2026-diary-consumer-payment-choice
3. JPMorgan Chase, "Chase Media Solutions" (accessed 9 November 2026). https://www.chase.com/mediasolutions/home
4. Cardlytics, Inc., Form 10-K for fiscal year 2025 (filed 4 March 2026). https://www.sec.gov/Archives/edgar/data/1666071/000166607126000010/cdlx-20251231.htm
5. Cardlytics, Inc., Form 10-Q for the quarter ended 30 September 2026. https://www.sec.gov/Archives/edgar/data/1666071/000162828026071002/cdlx-20260930.htm
6. Cardlytics, Inc., "Cardlytics Reports Third Quarter 2026 Financial Results" (Form 8-K exhibit 99.1, 5 November 2026). https://www.sec.gov/Archives/edgar/data/1666071/000162828026070992/cdlx_form8-kx2026q3xex991.htm
7. PayPal, "PayPal Ads Delivers Bottom-Line Growth Advertisers Can Measure" (22 June 2026). https://newsroom.paypal-corp.com/2026-06-22-PayPal-Ads-Delivers-Bottom-Line-Growth-Advertisers-Can-Measure
8. PayPal, "CES 2026: PayPal Ads Unveils Transaction Graph Insights and Measurement Partnership Program" (6 January 2026). https://newsroom.paypal-corp.com/2026-01-06-PayPal-Ads-Unveils-Transaction-Graph-Insights-and-Measurement-Partnership-Program
9. Marketing Dive, "Mastercard wants a bigger piece of a retail media pie" (1 October 2025). https://www.marketingdive.com/news/mastercard-bigger-grab-commerce-media-dollars-new-network/761568/
10. IAB and IAB Europe, "Guidelines for Incremental Measurement in Commerce Media" (3 November 2025). https://www.iab.com/wp-content/uploads/2026/01/IAB_and_IAB_Europe_Guidelines_Incremental_Measurement_Commerce_Media_November_2025.pdf