If your company holds data someone would pay for, the next question is how you want to be paid, and that's what a data monetization model decides. The same dataset can earn a single payment, a fee every year, or a share of every sale for years, depending on the model you choose. Read on to find out the six models companies use, how each one pays, which kinds of company fit each, and a five-step framework for picking the right one for your data.
What a data monetization model is
A data monetization model is the shape of the deal between you and whoever pays. It answers four questions. Who pays you, what they get, whether you keep ownership, and whether you're paid once or repeatedly. Strategy is the bigger decision about which of your data to put to work and for whom. The model is the mechanism underneath it, and picking the wrong one costs more than most companies realize, because a sale that pays once closes the door on every later deal.
The six data monetization models
| Model | How it pays | Who it fits | What it asks of you |
|---|---|---|---|
| One-time sale | A single payment, and the buyer owns the data | Companies closing down, or selling a finished archive they won't need | Nothing after the sale, and no later deals |
| One-time license | A single fee for a defined dataset and use, and you keep ownership | Owners of a finished archive who want to license it again later | A rights review and a clean package |
| Recurring license | An annual or multi-year fee for a living source, often with refreshes | Platforms and companies whose data keeps growing | Ongoing delivery and a relationship with the buyer |
| Marketplace revenue share | A share of every sale as the data licenses through a marketplace | Operating companies without a data team or a sales function | Scoping what's in, and a few hours on exports |
| Data product subscription | Customers subscribe to a feed, API or report you build | Companies whose data is already a product (weather, maps, financial feeds) | Building, packaging, selling and supporting the product |
| Indirect monetization | You earn by selling more or spending less, and nobody pays for the data | Any company with analytics capacity | A data team and a product to build into |
The first four are direct models, where someone pays for the data itself. The last two are what most data monetization advice covers, and they suit companies that already have a data team. Our guide to data monetization examples shows public deals under each of the direct models, with the prices.
How each direct model pays
The public AI licensing deals show the four direct models side by side.
- One-time sale. Spirit Airlines' operating records sold in bankruptcy court, where Google bid 10 million dollars for the whole archive in one purchase. Closing startups sell their code and workspace archives the same way, for roughly 10,000 to 100,000 dollars per deal at the small end of the market.
- One-time license. News Corp licensed its archive to OpenAI for a reported 250 million dollars over five years. The archive is still News Corp's, and it can license it again.
- Recurring license. Google pays Reddit about 60 million dollars a year for conversation data that Reddit's users refresh every day, and Reddit signed a separate deal with OpenAI on the same data.
- Marketplace revenue share. Thousands of creators and companies license through marketplaces like Troveo and are paid each time their data licenses. Typical full-company licensing deals for operating businesses start at six figures, and the first sale is the floor.
Our guide to data licensing as a business model goes deeper on the licensing structures, and our guide to exclusive vs non-exclusive data licenses covers the one term that changes the price more than any other.
Why the model matters more than the first number
Two companies can hold the same data and earn very different amounts, and the model is usually why.
- A sale pays once and ends. A license keeps ownership with you, and the same dataset can be licensed to more than one buyer.
- A recurring license and a revenue share both pay more than once, and they pay differently. Recurring fees need a living source and a relationship with each buyer. A revenue share needs a marketplace that finds the buyers, and it works for a finished archive too.
- The models stack. A publisher can run a recurring license with one lab and a one-time license with another, and a company can license through a marketplace while building an indirect model inside its own product.
The mistake companies make most often is taking a sale because it's simple, when a license would have paid several times over. Our guide to how much your company's data is worth to AI labs covers the six factors that set the number under any model.
A five-step framework for picking your model
Use this in order. Each step rules models out, and by the end one or two are left.
- Inventory what you hold. List the systems you run, how many years each one covers, and whether the data is finished (an archive) or still growing (a live source). A finished archive fits a sale or a one-time license. A live source can support a recurring license.
- Check what you own. Work out what your contracts let you license, what came from third parties, and what has to stay out. Anything you can't license cleanly is out of every model.
- Decide whether you're building or licensing. If you have a data team and a product to build into, a data product or indirect model is open to you. If you don't, license what already exists.
- Decide who does the selling. If you'll negotiate directly, you need a buyer relationship, a legal team and a sample the buyer can test. If you'd rather not, a marketplace handles the rights review, the packaging (with names and personal details replaced by pseudonyms), and every buyer conversation.
- Decide how you want to be paid. Once, every year, or on every sale. If you're not sure, non-exclusive licensing keeps every later deal open.
For most operating companies with 50 or more people on ordinary tools, the framework lands on a marketplace revenue share. There's nothing to build, you keep ownership, your company is never named publicly, and you're paid on every sale. Our owner's guide to licensing company data for AI covers what that process looks like step by step, and our guide to data monetization companies covers the marketplaces, exchanges and brokers you'd be choosing between.
Where to start
Step one of the framework is an inventory, and the quickest way to run it is our free data value assessment (ten questions about your systems, history and industry, about five minutes). You get an estimate of what your data could license for, which tells you whether the direct models are worth pursuing before you spend a day on any of them.
Where Troveo fits
Troveo is a data licensing marketplace, which is the revenue-share model above. We help companies understand what data they hold, protect what matters, and license what's valuable, selectively. We've paid rights holders more than 20 million dollars across video, audio, gaming, robotics and business data, we work with more than 40 active buyers, and we take no fees and no deductions from your payouts. Start with the data value assessment, read our business data page for how the process works, or talk to us about which model fits what your company holds.