Sell or license your company data? How each deal works and what it pays

Most companies that "sell" their data don't sell it at all. They license it.

In a license you keep ownership. The buyer gets the right to use the data, under rules you set. How you shape that deal decides how much you earn, for how long, and how much control you keep.

Here are the five common deals, how each is priced, and the contract terms that matter most.

1. Outright sale

You hand over the data and ownership of it. You may have to stop using it yourself.

When it fits: you're closing a product line, or the data is an old archive you no longer need.

The catch: you get paid once. You give up all future income from the data. Few companies sell data they still run on.

2. Non-exclusive license

You keep ownership. You license the data to one buyer and stay free to license it to others.

When it fits: most of the time. If several kinds of buyers want your data, you can license to each one. See who buys company data.

The catch: each buyer pays less. They know rivals may hold the same data.

3. Exclusive license

Only one buyer may use the data. The deal is often limited to one field, such as "exclusive for AI training" or "exclusive in banking."

When it fits: one buyer values the data far more than anyone else, and will pay well to keep it from rivals.

The catch: you lock out other buyers for the length of the deal. Keep the exclusive narrow and short. Make sure the extra money covers what you give up.

4. Data feed or subscription

The buyer gets ongoing access instead of a single file. Delivery is usually a regular file drop or an API. The buyer pays monthly or yearly.

When it fits: your data keeps updating and freshness matters. Investors and research firms often want this.

The catch: steady income means steady work. You must deliver on time, in the same format, for the life of the contract.

5. Revenue share or partnership

The buyer builds a product with your data and pays you a share of what it earns. Sometimes there is a smaller fee up front.

When it fits: no one knows yet what the data will earn, or the buyer can't pay much up front.

The catch: you may earn more over time, but your income depends on the buyer's success. You need the right to audit what you're owed.

How data licensing deals are priced

The price model follows the deal. Four are common:

• Flat fee for a set of data or a year of access.
• Per record or per thousand records. Brokers often pay this way.
• By usage, such as number of users, API calls or products built.
• Revenue share, a percentage of what the buyer earns from the data.

Whatever the model, price on what the data is worth to the buyer. Not on what it cost you to collect.

Data licensing terms that decide what you earn

Two deals with the same price are not worth the same. These terms make the difference:

• Permitted use: what the buyer may do with the data, in writing. Say whether AI training is allowed.
• Resale: can the buyer pass your data on? The answer should usually be no.
• Length and renewal: how long the license lasts, and how the price changes when it renews.
• Deletion: what happens to the data when the deal ends. Data inside a trained AI model may not be deletable.
• No re-identification: for data with personal details removed, the buyer must agree not to identify anyone.
• Liability: who pays if something goes wrong, such as a privacy complaint or a breach at the buyer.
• Audit rights: a must for revenue share and usage pricing.

Frequently asked questions

Is it better to sell or license company data?

For most working companies, license it. You keep ownership, can earn from several buyers, and can renegotiate as the market changes.

How long do data licenses last?

Often one to three years, with an option to renew. Reddit's 2024 filing to go public, for example, described data licensing deals of two to three years.

Should I give a buyer an exclusive license?

Only for a clear premium. Keep it as narrow as you can: one field of use and a fixed term.

How do I price my company data?

Start with an estimate based on similar data and the buyers who fit yours. Get a free estimate of your company data's value.