For years, the customer data question was what a company could buy or track. The question that matters now is why a customer would choose to tell a company who she is.
A vendor contract answers the first question. Only the relationship answers the second.
A shopper browses a home goods retailer’s website on three devices, one of them her sister’s tablet. Then she pays cash for a lamp in the store. To the retailer she is four strangers, and marketing has no way to reach them or credit them with the sale. The next Saturday the cashier offers her free returns and points, and she gives her email address and picks one favorite style. By the time the receipt prints, the retailer has one member where it had four strangers. Every visit from here on lands on her record, and what she agreed to sits beside what she bought.
No tracking pixel made that happen. She did, because the program gave her a reason worth giving up her anonymity for.
The identity is the data, and the reason is the strategy.
The cookie was never the real reason
From 2020 to 2024, the end of the third-party cookie was the reason companies gave for caring about first-party data. A third-party cookie is a small file that a site other than the one she is visiting places in her browser. Then Google changed course and kept the cookie in Chrome, so the deadline turned into a setting. The Privacy Sandbox, the set of tools Google built to replace the cookie, is now being retired. Safari and Firefox had already blocked or fenced off third-party cookies by default.
If the deadline was the whole case for first-party data, that case is gone. It was never the right case anyway.
First-party data earns its budget on two jobs that both need a known customer. Relevance means knowing who is asking: her size, her last order, the reward she is saving for. Measurement means seeing the sale that follows an offer, in the store as well as online. A cookie sees one browser on one device, so it loses her the moment she picks up another device or walks into a store. A member record follows her from the store to the app to the service desk, because she carries the ID with her. Loyalty Juggernaut builds its patented first-party data layer on that premise, as a source of truth the whole business uses, not a loyalty silo.
AI raises the stakes on both jobs, because any competitor can license the same models while the records they work from belong to one company. The model is not the advantage. The member record is, so the first line of a first-party data budget belongs to the reason customers join.
Recognition now shows up on the income statement
Kroger’s annual report, filed in March 2026, carries a sentence worth reading twice. Because of its rewards program, over 95% of customer transactions are tied to a Kroger loyalty card.
Read that as a customer decision, because it is one. Visit after visit, shoppers choose to scan the card, and the filing draws the line from that choice to the business. Kroger Precision Marketing, the company’s media arm, uses the card data to link the ads a household sees to what it buys, in stores and online. That is retail media, a retailer selling advertising in its own channels to the brands on its shelves. It sits inside what Kroger calls its alternative profit streams, which earned $1.5 billion of operating profit in 2025.
A card scanned at nearly every visit makes the whole store measurable. It does more work than a discount card gets credit for, because it is the identity that both the media business and the offer run on.
Target and Walmart show how large retail media has grown. Advertising revenue at each rose by more than 40% in the fiscal year to January 2026, Walmart’s with VIZIO included.
That makes recognition a revenue number, and the share of known transactions belongs in the same report as sales. It is also the plainest sign that loyalty has become an enterprise growth engine.
Customers identify themselves when the exchange is worth it
Customers are not refusing to share. They are deciding who has earned it.
Deloitte asked more than 5,000 U.S. loyalty members whether they would share personal information for more tailored offers. A majority in every generation said yes. Sephora’s Beauty Insider program hit a record 45 million members in North America in 2025, each of whom chose to be known. For a loyalty team, that is permission to ask, on one condition: what she gets back has to be worth what she hands over.
Joining is only the first exchange, and the same survey found that the average member belongs to eight programs and actively uses five. A record from a program she has stopped using goes stale: her address changes, her tastes shift, and the style she saved for her old living room no longer fits the new one.
A data platform stores what she said once. A program she keeps using keeps it true.
That makes consent a decision she renews every time she comes back, and a box ticked at the register is only its first draft. The loyalty team controls the reasons she comes back, which means it owns that renewal as much as privacy counsel does.
The source of customer data decides whose terms apply
For a strategy, the test for any customer data is the relationship behind it: she dealt with the company directly, on its channel, under its terms. First-party data passes, because it comes from her purchases, accounts, and loyalty activity under the company’s privacy notice. So does zero-party data, what she tells the company on purpose, such as her favorite style. Second-party data is a partner’s first-party data, shared under contract or in a clean room. A clean room is a controlled space where two companies match customer lists without swapping raw records. Either way, the partner’s terms, never the company’s own, govern how it is used. Third-party data comes from brokers with no link to her, so it carries the weakest consent trail, the thinnest record of what she agreed to.
Bought data is not first-party data with a weaker label. It is someone else’s relationship, rented. The budget should follow that ranking: fund the reason customers give their own data first, and buy reach only where that data runs out.
Seven decisions build a first-party data strategy
Each decision needs an owner and a number to watch.
- Start with the decisions the data should change. Name the moments where knowing her changes the outcome: the offer, the service call, the media plan. The chief marketing officer should own this list with finance, because it sets what the business will do differently once it knows her. Data with no decision attached is storage.
- Give her a reason to be known at every touchpoint. Member prices, points, free returns, and saved preferences all give her a reason to sign in at the register as well as online. A reason that does not work at the register leaves it anonymous. The loyalty lead should track the join rate, the share of shoppers who sign up, at every touchpoint, the store included.
- Ask for consent where the value is. Ask for her email address when she wants a digital receipt, because that is when she can see what she gets for it. Then honor what she agreed to in every system, which is where first-party data stops being only a marketing question. In June 2026, the International Association of Privacy Professionals counted 23 states with comprehensive privacy laws, four of them enacted that year. Alabama’s privacy bill, passed in April 2026, expressly allows different prices for members of a loyalty program they chose to join. Privacy counsel belongs beside the loyalty lead whenever the program sets member prices.
- Resolve every touchpoint to one member record. Identity resolution is the work of matching each sale and service contact to the right person. Point of sale, app, website, contact center, and partner transactions should all land on the same member ID. A customer data platform, the software that stores and joins customer records, holds the record. The loyalty ID makes the joins reliable, because she presents it herself. Identity resolution starts with the reason she gave to be known, and the technology carries that reason into every system.
- Keep the record current. Track how many members used the program in the last 90 days and how old each saved preference is. Both numbers slip before anyone notices the record is wrong. Give members small reasons to confirm what the program knows, such as a size check before a sale. A size she confirmed at her last visit beats one a model guessed from old orders.
- Put the record to work where she and the business can see it. Build segments, groups of members who behave alike, and let them decide the offer and the message each one gets. With her consent, the record also reaches paid media. Google’s Customer Match targets ads to an advertiser’s own customers and to other customers like them. A clean room lets two companies measure a shared campaign without trading their lists.
- Measure recognition, then measure lift. Track the share of transactions tied to a known customer; Kroger’s is over 95%. Then measure lift, the extra sales an offer caused, with a holdout: a random group of similar members who get no offer. A sale from a member is not proof the program caused it. Analytics and finance should report incremental margin: the profit the offer added beyond what members would have spent anyway. Recognition shows the sale, and the holdout shows the cause.
GRAVTY® generates first-party data from what members do
Back to the lamp. She joined at the register with contact details and one saved preference. In GRAVTY®, Multi-Dimensional Behavior Tracking, a patented approach to generating first-party data, builds her profile from what she does next. She signs in to the app and starts saving toward a reward on a side table.
Some weeks she still pays cash at a smaller store in the chain. She photographs the receipt, and GRAVTY® AI-Scan, AI-powered receipt scanning for offline retail, credits her points on the spot and adds the purchase to her record. A cash sale that used to belong to a stranger now belongs to the relationship. The same record works at national scale: GRAVTY® runs Spin Premia, FEMSA’s loyalty program, in real time across 33,000 stores.
Say the team wants to know whether members who join at the register behave differently from members who join online. Agentic AI Compass, the multi-agent intelligence suite launched in October 2025, works the question like a team of analysts. It finds the pattern and tests a first-month offer in simulation before it goes out.
GRAVTY’s patented AI-Driven Mass Individualization then shapes a 1:1 version of that offer for her: bonus points toward the side table she is saving for. The team holds out a random group of similar members and counts what the offer added.
The strangers became one member, and one member became a relationship the business can measure.
She should feel known, never watched
A customer never experiences a first-party data strategy. She experiences what the company does with the things she shared.
The team should put one plain test to every use of her record before it ships. Does it make her next visit easier or her offer more useful? If all it does is make her easier to target, the company has spent her trust and given her nothing she can see.
What she notices is small. The receipt knows her name. Her favorite style is already at the top of the page when she shops, and the offer brings the side table closer. Each detail tells her that the information she gave is working for her.
The old strategy asked what a company could collect. The next one asks what a customer would choose to share, and whether the company keeps deserving it.
Recognition is hers to give. The program’s job is to keep earning it.
The numbers behind first-party data
Each figure below links to its primary source.
Retail media runs on the loyalty card
Kroger’s own filing ties its media business to the loyalty card. Target and Walmart show how large retail media has grown.
| Company | What it reported | Source |
|---|---|---|
| Kroger | Over 95% of customer transactions tied to a loyalty card, credited to its rewards program. About 63 million households a year. Kroger Precision Marketing links ad impressions to household transactions. Alternative profit streams, retail media among them, earned $1.5 billion of operating profit in 2025 | Annual report, March 2026 |
| Target | Advertising revenue of $915 million in fiscal 2025, up 41% from $649 million, while merchandise sales fell from $104.8 billion to $102.7 billion | Annual report, March 2026 |
| Target | Roundel, its media business, drove nearly $2 billion in value in 2024, when more than 13 million members joined Target Circle | Press release, March 2025 |
| Target | Members spend three times more than non-members, a gap that also reflects which shoppers choose to join, so it is no proof of what the program caused | Fourth-quarter 2024 summary |
| Walmart | Global ad business grew 46% in the fiscal year, to nearly $6.4 billion with VIZIO | Earnings release, February 2026 |
Members will share when the exchange is worth it
| Source | Finding | What it means for a program |
|---|---|---|
| Deloitte 2025 Consumer Loyalty Program Survey, 5,564 U.S. members, January 2026 | 89% of Gen Z, 87% of millennials, 78% of Gen X, and 64% of baby boomers are willing to share personal information for more tailored offers or experiences | Members will answer when the offer that follows is worth it |
| The same Deloitte survey | The average consumer belongs to eight loyalty programs and actively uses five | A program has to earn a place among the five she uses, or her record goes stale |
| Sephora, January 2026 | Beauty Insider hit a record 45 million members in North America in 2025 | A retail program can hold tens of millions of known customers |
Four kinds of customer data and whose terms cover each
| Type | Whose terms cover it | What it does for the strategy |
|---|---|---|
| First-party data (purchases, accounts, app and web visits, service contacts, loyalty activity) | Your company, under your privacy notice and her settings | Relevance and measurement tied to real sales |
| Zero-party data (what she tells you on purpose: sizes, preferences, goals) | Your company, for the use she agreed to | Offers and recommendations she asked for |
| Second-party data (a partner’s first-party data, shared under contract or in a clean room) | The partner that collected it; your use stays inside its terms | Reach and measurement beyond your own customers, as in a loyalty ecosystem |
| Third-party data (assembled by brokers with no link to you) | The aggregator, under permissions gathered elsewhere | Broad prospecting and modeling, with the weakest consent trail |
Each decision has an owner and a measure
| Decision | Owner | Measure |
|---|---|---|
| 1. Name the decisions the data should change, starting with personalized offers | Chief marketing officer, with finance | Decisions with a known-customer version, and the revenue each one touches |
| 2. Give her a reason to be known | Loyalty lead | Join rate at each touchpoint, the store included |
| 3. Ask for consent where the value is | Privacy counsel and the loyalty lead | Members reachable by email or text, with a recorded consent |
| 4. Resolve every touchpoint to one member record | Data and IT teams | Share of activity in every channel that lands on a member ID |
| 5. Keep the record current | CRM team | Members active in the last 90 days; age of saved preferences |
| 6. Put the record to work in customer segments and paid media | Marketing and media teams | Offers and ad audiences built from member data |
| 7. Measure recognition, then lift | Analytics team, with finance | Share of transactions tied to a known customer; incremental margin from holdouts |
Frequently asked questions
What is a first-party data strategy?
A first-party data strategy is a company’s plan to earn data from its own customers with their consent and put that data to use. It sets the decisions the data should change and the reason customers have to identify themselves. It also sets how consent is kept and how lift is measured. A loyalty program can supply both the reason and the ID, which is why the loyalty lead belongs in the room when the strategy is set.
What is 1st party, 2nd party, and 3rd party data?
First-party data comes from your own customers, and second-party data is a partner’s first-party data shared under contract. Third-party data is bought from firms with no link to the customer, and the same data is sold to many buyers at once. First-party data includes her purchases and the points she earns and spends in your loyalty program. Zero-party data is what she tells you on purpose, such as her size. For a brand, the practical difference is whose terms govern each use, and the first-party data guide defines each kind.
How do you collect first-party data?
Collect first-party data by giving customers a reason to identify themselves, then asking for more when the value is clear. Loyalty memberships, accounts, member prices, digital receipts, and service requests all turn an unknown visit into a known customer. Google’s guidance to advertisers lists loyalty programs with rewards among the ways to generate first-party data. On a website, progressive profiling asks one question per visit. Keep a record of what she agreed to next to the data.
Is Google still phasing out third-party cookies in Chrome?
No, Google is keeping third-party cookies in Chrome. In July 2024 it dropped its plan to end them. In April 2025 it kept the existing user choice in Chrome’s settings and dropped a planned standalone prompt. In October 2025 it said it would retire most Privacy Sandbox tools because adoption was low. Safari has blocked third-party cookies by default since March 2020. Relevance and measurement still need a known customer, whatever browsers do. A brand should build its plan on the member record and leave browser policy out of the business case.
Is loyalty program data first-party or zero-party data?
Loyalty program data is both first-party and zero-party data. What members do in the program is first-party data, from purchases to redemptions. What they tell it on purpose is zero-party data, such as sizes and birthdays. A program gathers both under one consent and one member ID, and it refreshes them each time she comes back. The same record can then shape the offer and measure what it did.
Do you need a customer data platform for a first-party data strategy?
No, a first-party data strategy can start without a customer data platform. A customer data platform or a data warehouse stores and joins the data. It works only as well as the ID and the permission behind it, and the program supplies both. Start with the reason customers identify themselves. Then pick the storage that fits the decisions the data has to support.
How do you measure a first-party data strategy?
Measure a first-party data strategy by recognition first, then lift. Recognition is the share of transactions tied to a known customer and the share of members active in the last 90 days. Lift comes from a holdout test: a random group of similar members gets no offer, and the gap in what they buy shows what the offer added. The loyalty program ROI guide sets out the method.
