The company that monetised purchase data without ever selling it

A company you have never heard of sees one in every two card swipes in America

Cardlytics reaches nearly 215 million shoppers and observes roughly one in two card swipes in the United States — $5.8 trillion in annual global spend as of the end of 2024. It has no app on anyone's phone, no loyalty card in anyone's wallet, and no relationship with a single one of those shoppers.

It sits between two parties who could never have worked together directly: the banks that hold purchase data, and the brands that would pay a great deal to understand it. That position is the entire company. It is also the clearest illustration we know of a principle that matters enormously in Saudi Arabia right now.

The problem Cardlytics solved

A consumer-goods company faces an old and expensive blindness. It sells through retailers, so it never meets the person who buys its product. It knows how many cases left the warehouse. It does not know who bought them, whether the same person came back, or whether the campaign it ran last month changed anything at all.

The bank, meanwhile, sees the purchase — but cannot act on it. Handing customer transactions to an advertiser would be a betrayal of the relationship the bank exists on, and in most jurisdictions unlawful besides.

So both sides sat on halves of a picture neither could assemble. The obvious move — the bank sells the data — was the one move nobody could make.

What they built instead

Cardlytics never takes the data out of the bank. Its software runs behind the bank's own firewall. The bank strips identifying details before anything is processed, so no name and no account number leaves the building, and Cardlytics itself cannot tell one customer from another; transactions are attributed to opaque identifiers, not people.

The advertiser gets less than you would assume, and it turns out to be enough. It never receives identity. It receives the ability to describe an audience by behaviour — people who buy this category, people who buy a competitor, people who used to buy from us and stopped — and to measure, afterwards, whether the campaign produced real purchases rather than clicks.

Everyone in the arrangement gets the part they actually wanted:

• The brand gets to reach real buyers and see whether spending worked, without ever holding a stranger's personal data
• The bank gets rewards that make its card the one customers reach for first, and a revenue line that costs it nothing to produce
• The customer gets an offer on something they already buy, inside an app they already trust, and their data never leaves it

The lesson, stated plainly

Purchase data is enormously valuable and cannot be sold. Both halves of that sentence are true, and the second is not a limitation to be engineered around — it is the reason the first one holds. The moment the data leaks, the trust that produced it evaporates, and with it the asset.

So the value has to be delivered without the data moving. That is a harder thing to build than a data sale, which is precisely why the company that builds it ends up occupying the middle.

Where the receipt goes further than the card

Here is the limit of the card. A card transaction knows the merchant, the amount, and the time. It does not know what was in the basket.

The bank can tell that a customer spent 214 riyals at a supermarket on Thursday. It cannot tell whether that basket held your brand of coffee or your competitor's, whether the shopper switched sizes, whether the promotion moved units or just moved margin. For a consumer-goods company, that missing layer is the whole question.

The receipt is the layer. It carries the line items — the product, the quantity, the price paid, the branch, the time — which is the difference between knowing that someone shopped and knowing what they chose.

Why Saudi Arabia is a different starting position

Cardlytics was built in a market where the ground rules were written afterwards. Saudi Arabia has the opposite sequence: the Personal Data Protection Law is already in force, and documented consent is the entry ticket rather than a compliance chore bolted on later.

That sounds like friction. It is closer to an advantage. A consent captured properly — with the exact wording the customer saw, the moment they agreed, and a record that survives an audit — is an asset that a scraped dataset can never become. It is defensible. It can be built on.

And there is a second difference. A card number is issued by a bank; a receipt is issued by the merchant, to a customer standing in front of them. The relationship is already there. What was missing was a lawful, practical way to record it — because a phone number recited to a cashier carries no consent with it: no wording shown, no timestamp, nothing to produce when someone asks.

What this means for a merchant, a brand, and a bank

Wateer does not issue your invoices; your point of sale does that, exactly as it does today. Wateer is the layer above the transaction — unifying receipts from every system a merchant runs, tying each payment to the receipt it belongs to, reconciling that against what actually landed in the bank, and delivering the receipt to the customer with their consent recorded alongside it.

Which puts the same three-sided position within reach here, on richer ground:

• The merchant finally has a lawful way to build a customer relationship instead of a spreadsheet of numbers collected at the till
• The brand can learn what actually moved off the shelf, at the line-item level a card can never reach
• The customer keeps every receipt in one place, with a consent they can withdraw as easily as they gave it

Cardlytics proved the shape works at the scale of half a country's card swipes. The receipt is the deeper layer, and Saudi Arabia is the market where it can be built on consent from the first day rather than retrofitted onto it.