Recurring data revenue is the monthly income from licensing each new month of your company's records to buyers who already bought your history. The first license pays for the past, and a refresh term in that license pays again every time a new batch of records is delivered and accepted.
Most data sales are one-off: a company exports an archive, a buyer pays once and the relationship ends. A business that keeps operating keeps creating records, though, and labs want current data as well as history. A refresh term turns that steady activity into a payment schedule.
Two prices in one license
Each of our licenses has two prices:
- A price for the history, paid once for all past records up to the first delivery
- A refresh price, paid for each monthly batch of new records while the refresh term is active
The history price is usually the larger single payment, because it covers years of records. The refresh price is smaller per month but repeats for the length of the term. Over a multi-year license, refresh payments can add up to more than the history did. Our data licensing terms guide covers the other clauses that sit alongside these prices.
The large public deals follow a similar shape. Reddit disclosed $203 million in data licensing contracts with two to three year terms, and News Corp's OpenAI deal was reported at more than $250 million over five years. In both cases the seller keeps producing new content, so the buyer pays over time for a flow of data.
How the monthly batches work
After the first license, our SDK does the monthly work on its own:
- It runs on your machine on a schedule, monthly by default.
- It pulls only the records created or changed since the last successful run.
- It scrubs them locally, removing names, contact details, account numbers and secrets, and rounding rare values.
- It uploads the scrubbed output and a scrub report as a new batch for that month.
Raw data never leaves your machine through the SDK. Only the scrubbed output and the report are uploaded. Each batch covers one dataset and one month, and moves through review until it is accepted or rejected.
Every accepted batch pays the refresh price once for each license with an active refresh term. Two licenses with refresh terms on the same dataset mean two refresh payments for each accepted batch. The recurring revenue docs list the exact statuses and payout rules.
No work after setup
Once the SDK is scheduled, there is nothing to do each month. You don't export files, chase buyers or send invoices. The portal shows each batch as it moves from received to accepted, and you come back when we ask you to approve a new buyer or when something needs fixing, such as an expired access token.
Setup itself is a short job for someone technical: install the SDK, connect the source systems you chose and run the schedule command. The quickstart walks through it, and our SDK page covers what it connects to.
Why refresh batches grow
Refresh payments track how much your team does. A firm that sends more invoices, closes more deals or handles more tickets creates more records each month, so each batch carries more data. As your team grows, so does each refresh.
Fresh records also matter to buyers. Practices change, software changes and new kinds of cases appear. A model trained on how work was done five years ago misses what changed since, so a dataset that stays current is worth more to a lab than one that stops at the day of the first sale.
Who keeps what
We take a commission on each license, including each refresh payment. Payouts record your share in basis points. Your seller agreement sets the share. If nothing sells, you pay nothing.
When recurring revenue doesn't apply
A refresh term only makes sense when new records keep arriving. It doesn't fit:
- A company that has closed or stopped using the system
- A one-time archive, such as records from a discontinued product line
- Data where every new batch would need fresh customer consent you can't get automatically
In those cases the history license is the whole deal. The Forbes report on shut-down startups selling archives for $10,000 to $100,000 each describes that kind of one-time sale.
Rights apply to every batch
Monthly batches go through the same rules as the first delivery. If your contracts allowed the history to be licensed, they usually allow new records of the same kind, but check whether any new customer contracts have different terms. You approve every buyer, and you can end a refresh term under the terms of your seller agreement. Data already delivered and used for training can't be pulled back out of a trained model, so decide carefully before each new buyer.
Model your recurring revenue
Our recurring data revenue page shows how history and refresh payments add up over 24 months, and our calculator estimates what your history is worth today. If both numbers make sense, book a call and we'll look at your systems and contracts. Our guide on how much your data is worth covers the factors behind the estimate.
Frequently asked questions
Is recurring data revenue the same as a subscription?
It works like one from your side. The buyer pays the refresh price for each accepted monthly batch while the refresh term is active, and you receive your share of each payment.
What work does my team do each month?
None, once the SDK is scheduled. It pulls new records, scrubs them on your machine and uploads the batch on its own. You only step in to approve new buyers or fix an expired credential.
What if a monthly batch is rejected?
A rejected batch doesn't pay. Batches are reviewed after upload and either accepted or rejected, and the portal shows each batch's status so you can see when one needs attention.
Can I license the same data to more than one buyer?
Yes. Our licenses are non-exclusive by default, so several buyers can license the same dataset, and each license with an active refresh term pays for every accepted batch.