What is data monetisation?

Data monetisation is turning the data a business holds into money: by using it to run the business better, by building it into what the business already sells, or by selling it, or something made from it, to someone else.

In short. Almost every business now holds data worth money to someone. Very few capture that money. The data on its own is the cheapest part of the business; the value goes to whoever does the work of turning it into an answer a buyer will pay for. Data monetisation is the discipline of being that party.

What counts as data monetisation?

Three routes cover the field. The first is to improve: use the data inside the business to lower cost or raise performance, and take the money as margin. The second is to wrap: build the data into the product the business already sells, so the customer pays more, buys more or stays longer. The third is to sell: put a price on the data, or on a product made from it, and sell it to someone outside. Barbara Wixom, Cynthia Beath and Leslie Owens set out these three routes in Data Is Everybody's Business, and they are the right starting map.

The register on this site holds examples of all three from the past year. Customers Bank, a $25.9 billion regional lender, is using its own operating data with an AI partner to bring commercial loan cycles from 30 to 45 days down to about seven, with a stated target of moving its efficiency ratio from about 49% to the low 40s. That is improve: the money never appears as data revenue, it appears as cost that is no longer there. Bank of America added peer benchmarks on payments and working capital to its corporate treasury portal in September 2026, drawing on the 213 million payments the platform handled in the first half of the year, and named no separate price for them. That is wrap: the data makes the existing product harder to leave. PhonePe launched PulsePro in July 2026, a paid intelligence product built on aggregated payment flows from more than 700 million registered users, sitting above a free portal it has run since 2021. That is sell.

Most people mean the third route when they say data monetisation, and most of the money is in the first two. Selling is the visible route because it produces a revenue line with the word data in it. Improving and wrapping produce money that shows up as margin, retention and price, and nobody writes a press release about a churn rate.

Is data monetisation the same as selling data?

No, and the confusion costs companies money. Selling raw data is the narrowest and usually the least valuable way to monetise it. A raw data file is easy to copy, hard to price and, once delivered, out of your hands; the buyer learns what it is worth, and you never do.

What companies actually sell, when they sell well, is something made from the data: a benchmark, a score, a forecast, a placement, an answer. Citi's commerce media business, launched in September 2026, is built on 6.5 billion card transactions a year across more than 700 spending categories, and brands buy advertising placements from it, never the transaction file. McDonald's media network, piloted in 450 restaurants from August 2026 with a stated ambition of a $1 billion business, sells reach against roughly 70 million daily customers; the visitor and transaction data stays with McDonald's. In both cases the data is the reason the product is worth buying, and in neither case does the data leave.

Most data monetisation does not involve personal data changing hands at all. It runs on data used inside the business, on aggregated and anonymised data turned into benchmarks, or on data that never leaves the building because the buyer is paying for an outcome rather than for records. Where personal data is involved, the law, the contracts and the buyer's own scrutiny set hard limits, and a well-built data business treats those limits as part of the design rather than as an obstacle to it.

Who monetises data?

Any business that records what its customers, suppliers or operations do. The register shows the range in a single year. Banks turn card and payment flows into advertising and treasury benchmarks. Retailers, restaurants, fuel stations and shopping centres turn their customer traffic into media networks. Accounting platforms turn their customers' books into peer comparisons: Xero launched Industry Benchmarks in June 2026, scoring each small business against its industry and country cohort on nine financial metrics, with the accountant as the person who acts on it. Media companies license their archives to AI developers. Health networks license de-identified records to pharmaceutical research. Information companies, whose whole business is data, are bought and sold as assets.

The common thread is a by-product. Very few of these businesses set out to produce data; they recorded it while doing something else, and later found that someone else would pay for what it shows. The by-product is the point and the trap. Because it was never designed as a product, it arrives without the work that would make it one.

Why doesn't holding data make you money?

Because data on its own is the cheapest part of a data business. In August 2026 Google won a bankruptcy auction for the corporate data of Spirit Airlines, a defunct US carrier: about 100 million emails, 500 million internal chat messages, 30 million lines of code and 3.4 million payroll records across some 80,000 accounts, de-identified before transfer, with the passenger and loyalty records carved out. The winning bid was $10 million, against a single backup bid of $7.5 million. An entire airline's internal record, gathered over years, fetched about $125 per account.

The contrast sits in the same register two months later. In October 2026 LSEG agreed to buy Octus, a credit-intelligence business, for $3.535 billion in cash. Octus collects hard-to-find credit documents, covenants and deal terms, and turns them into structured, searchable data used by more than 40,000 people at buy-side firms, banks and law firms; it is expected to generate over $300 million of revenue in 2026 at an adjusted EBITDA margin of about 40%. Octus does not write the documents its product is built on; they are other people's paperwork. It owns the work that turns them into something a credit analyst will pay for every year.

Spirit's archive was a large pile of recorded data that nobody had turned into an answer. Octus is other people's paperwork with years of work on it. The market priced the first at $10 million and the second at three and a half billion.

The book behind this site puts the rule in one line: value goes where the work is, not where the data is. The data a business holds before any work is done on it is the substrate. It becomes worth something only when someone builds the capability to turn it into an answer a buyer pays for: cleaning it, joining it, modelling it, packaging it into a decision. Whoever does that work captures the value, whatever the contract says about who owns the data. Most of the strategic choices in data monetisation are choices about who that party will be.

How big is data monetisation?

Market-size estimates for "data monetisation" vary by an order of magnitude, because nobody agrees on what to count: the improve and wrap routes never appear as data revenue at all. The more useful evidence is what companies disclose and what buyers pay.

What they disclose is starting to appear in results. CuriosityStream, a factual-media streaming company, reported $14.1 million of licensing revenue in the second quarter of 2026, up 48% on the year, driven by licensing its video archive and a private code corpus to AI developers, with gross margin rising to 72.8% from 53.4% a year earlier. The same library that earned inside a subscription is now priced directly against a model developer.

What buyers pay shows up in acquisitions. S&P Global completed the purchase of With Intelligence, a private-markets data business with expected 2025 revenue of about $130 million, for $1.8 billion in November 2025. LSEG's $3.535 billion for Octus followed in October 2026. The two prices come to about 14 times With Intelligence's expected revenue and about 12 times Octus's. Buyers pay that multiple for a business that turns data into an answer clients renew every year, not for the data on its own.

What decides who keeps the value?

A business can have valuable data, real buyers and a capable team and still keep a small share of what its data creates. The share is rarely decided in the negotiation. It is decided earlier, by a handful of design choices that usually nobody recognises as choices at the time, and by the time a contract is on the table they have already set the ceiling on what the business can keep.

When those choices are inherited rather than designed, the result is under-capture: the business leaves value on the table in deal after deal, in a place the commercial team cannot reach, because the fix sits upstream of the negotiation.

What those choices are, how to recognise them in a business and how to make them deliberately is the subject of my book, Value Capture: The Hidden Economics of Data Businesses, out in 2027. More about the book.

Why does data monetisation matter more now?

Two things changed, and they changed at the same time.

The first is on the cost side. Turning raw data into an answer used to take a team of specialists, years of engineering and a budget few companies would approve for a side business. The data had the value in it. The capability to get it out was expensive. Now it is not. Work that once needed a dedicated data business can be built by a small team in months, which means many more companies can do the work on their own data instead of handing it to a partner who does.

The second is on the buyer side. AI developers have become buyers of a kind the market never had: they want large, clean, rights-cleared bodies of text, code, images, audio and expert judgement, and they pay for them in cash. A media archive, a code base or a library of solved problems now has a price that did not exist a few years ago. The same developers are also the party most able to do the enrichment work themselves, so the business that sells them data has to decide carefully what it is selling and what it is giving away.

Cheaper capability and a new class of buyer pull in opposite directions. One lets data holders do the work and keep the value. The other tempts them to sell the raw material to the party that will.

Questions people ask

What is an example of data monetisation? A bank turning its card transactions into an advertising business, sold to brands as placements rather than as data (Citi, 2026); an accounting platform turning its customers' books into industry benchmarks inside its product (Xero, 2026); a media company licensing its archive to AI developers (CuriosityStream, 2026). The first two keep the data inside the business; the third sells access to it.

What is the difference between direct and indirect data monetisation? Direct monetisation puts a price on the data or on a product made from it and sells it to an outside buyer. Indirect monetisation uses the data to improve operations or to make the existing product more valuable, and takes the money as margin, price or retention. Indirect monetisation is usually larger and almost never reported as data revenue.

Is data monetisation legal? Yes, within the law and the contracts that govern the data. Most of it uses data internally or sells aggregated, anonymised outputs rather than records about people. Where personal data is involved, privacy law, consent and contracts limit what can be done, and the limits belong in the design of the product from the start.

Who should own data monetisation in a company? The executive who owns revenue and margin for the business line the data comes from. Data monetisation is a commercial decision about who captures value; the data and technology teams build what it needs, but they cannot decide the price, the buyer or what leaves the building.

Where should a company start? With the buyer's decision, before the data: whose decision the data could improve, what that decision is worth to them, and who would have to do the work to answer it. The Data Monetisation Readiness Assessment on Monetaise is a structured way to run that first pass.

---

Next in the guide: The ways data makes money, the full map of models, from licences and data products to media networks and partnerships. All the deals cited here are in the register, each with its reading.