There is an odd little moment in digital banking. You open an app to see how much money you have, and the app suggests somewhere to spend it. A restaurant. A shoe shop. A hotel. The inducement is cash back. The setting is your bank. Somewhere between those two facts sits Cardlytics.
- Brands pay to reach shoppers through bank and rewards channels.
- Purchase records help target offers and measure subsequent sales.
- Recent exits have narrowed the business around its core platform.
Most advertising begins with a guess about what you might want. Cardlytics begins with evidence of what you bought. That changes the conversation. A restaurant can look for people who eat out but have stopped visiting its locations. A retailer can pursue customers spending with competitors. The bank statement becomes a map of opportunities.
The coupon found a quieter address
Scott Grimes and Lynne Laube founded Cardlytics in 2008. Their banking background mattered: this idea required financial institutions to participate. The company needed an arrangement that made sense to the bank, the advertiser and the customer, while respecting the restrictions around transaction data. An appealing offer was only one piece of the machinery.
Consider the ordinary shopper’s version. You log into a participating banking app, activate an offer, and use the eligible card to buy something under its terms. The reward follows through the program. The merchant gets a purchase, the customer gets money back, and the bank gets another reason for you to keep using its card.
Cardlytics supplies the connection: audience selection, advertising delivery, campaign management and measurement. Its expertise lies in making transaction records useful to marketers inside bank-controlled environments. It says financial institution partners do not give it personally identifiable information. That is a specific description of data handling, rather than permission to stop thinking about privacy.
The sale after the sale
The distinction that matters is between counting a purchase and causing one. Someone who was already going to buy lunch can happily collect a reward. The advertiser has paid for generosity; whether it has bought additional business is another question.
Cardlytics describes a test-versus-control approach: compare customers exposed to advertising with a representative group held back. Its Purchase Intelligence dashboards also examine loyalty, geography and competitive spending. This places the company within commerce media, alongside networks that connect advertising to shopping. Its particular address is the financial account, with visibility across participating merchants rather than one retailer’s checkout.
A&W makes the proposition concrete. In a published case study, Cardlytics reports close to 250,000 purchases from new and lapsed guests and a $3 incremental return for every advertising dollar. Nielsen Sales Lift Measurement validated the results. Within 90 days, 45% of lapsed guests and 26% of new guests made another purchase.
“Seeing this level of detail”Liz Bazner, Director of Digital, A&W Restaurants
Those are campaign results, not a promise to every restaurant. But the repeat visits are worth watching. A discount can move today’s lunch. A second visit suggests the possibility of a habit. The useful question for a marketer is how much additional margin survives the reward and campaign fee.
A reward has to come from somewhere
Cardlytics sells performance-linked advertising. Its filings describe both served-sale and engagement-based pricing. Under served-sale pricing, a merchant can pay on purchases by customers shown an ad even when they do not activate the offer. A public rate card would make comparison easier; the disclosed mechanics instead point to negotiated campaign economics.
The money then splits. Consumer incentives reduce billings to revenue; partner payments take another bite. For 2025, the company reported $385 million in billings, $151.7 million in consumer incentives and $233.3 million in revenue. After partner share and other third-party costs, adjusted contribution was $130.3 million. That last figure still precedes substantial expenses.
This is why the business deserves a more careful description than software that serves coupons. Banks provide distribution and access to spending records, and they receive compensation. The same relationship that makes the network valuable puts an important part of its economics in somebody else’s hands.
The company’s published values include “Growth over Comfort” and integrity with partners and data. Its careers page describes analysts, creatives, developers and account managers working together. That mix fits a business whose promises have to satisfy both a marketing department and a bank.
The expensive search for a bigger picture
In 2021, Cardlytics bought Bridg for approximately $350 million in cash at closing, with potential additional earnouts. The attraction was product-level intelligence. Knowing somebody spent money at a grocer is different from knowing which products went into the basket. Bridg was meant to help close that gap.
The portfolio subsequently contracted. The Dosh consumer app was decommissioned in February 2025. Cardlytics’ Bank of America relationship ended in February 2026. In March, it sold Bridg’s assets to PAR Technology for 1,810,222 PAR shares. The original cash payment and later stock consideration are different measures; treating their difference as a simple realized loss would mislead.
Management explained the sale as a move to simplify operations and concentrate resources on the core Cardlytics platform. The sequence shows what changed: broader ambitions gave way to a narrower operating focus. Public disclosures establish the actions and management’s explanation, without requiring a fictional boardroom epiphany.
The reset remains unfinished. Second-quarter 2026 revenue was $36.9 million, down 36% year over year, with a $14.9 million net loss. Monthly qualified users were 185.4 million. That metric describes platform reach; it does not mean every person redeemed an offer. Positive adjusted EBITDA of $1.7 million did not erase the accounting loss.
Borrow the experiment. Check the landlord.
There is something practical to copy here: start with observed behavior, choose a specific audience, and reserve a comparison group. Ask whether customers return after the incentive disappears. Those principles travel beyond banking. The full model requires transaction access, participating publishers and enough merchant margin to fund rewards.
Competition complicates the arrangement. Chase is a Cardlytics partner and operates its own media business. Banks can be collaborators and alternatives. In October 2026, Cardlytics announced a Rove travel-rewards partnership, extending its offers to another distribution channel. The next question is whether that wider set of doors can support a healthier business behind each reward.