Guide
How to sell data to companies
How to sell data to companies that buy it: who the business buyers are, what they test, and how to pitch and pass diligence.
Updated
The short answer
To sell data to companies, start with the buyer type: investment firms, AI labs, marketing and sales teams, and product or analytics teams each want different things. Pitch the decision your data helps them make, send a documented sample under NDA, answer their diligence on rights, privacy and quality, then agree a licence that fits their use.
Know which company you are selling to
"Companies" covers buyers with almost nothing in common. A quant fund wants a long, consistent history it can test against prices. A machine learning team wants volume, variety and clean rights. A sales team wants accurate contacts for a defined market. Pitching the same dataset the same way to all three wastes every meeting.
| Buyer type | What they use data for | What they test first | Common deal-breakers |
|---|---|---|---|
| Investment firms (alternative data) | Signals about companies, sectors or the economy | History length, consistency, point-in-time accuracy | Possible inside information, gaps, changing methodology |
| AI labs and model builders | Training and evaluating models | Rights, uniqueness versus public data, format | Unclear licence, scraped content, personal data |
| Marketing and sales teams | Finding and reaching customers | Match rates, freshness, coverage of their market | Consent problems, stale contacts |
| Product and analytics teams | Filling gaps in their own product or research | Fit with their schema, update schedule, support | No ongoing feed, no documentation |
How to pitch and close
- 1
Write a one-page data sheet
Cover what the data describes, coverage (time, geography, entities), size, format, update schedule, how it was collected, and what has been removed. Leave out adjectives; buyers skim for facts.
- 2
Lead with the decision it improves
Instead of "10 years of shipment records", say what the buyer can do with them: "see shifts in import volumes for 40 retailers weeks before they report". Then prove it with the sample.
- 3
Find the person who owns the budget
At an investment firm that is often a data sourcing or data strategy team; at an AI lab, a data partnerships or acquisition lead; at a corporate, the head of the team that will use it. A general sales inbox rarely leads anywhere.
- 4
Sign an NDA, then share a real sample
A sample should be representative, not your best rows. Buyers who find the full delivery worse than the sample remember it.
- 5
Prepare for diligence before it starts
Have the rights documents, collection description, privacy review and data dictionary ready. Answer in writing, and say "we don't know yet" rather than guess.
- 6
Agree the main terms in a letter of intent
Scope, fields, history, price and payment schedule, exclusivity, trial period and delivery. Getting this on one page before lawyers draft the licence saves weeks.
- 7
Sign the licence and deliver
The licence sets permitted use, whether the buyer may share derived results, how long they may keep the data and what happens at the end. Deliver through a logged channel and confirm acceptance.
Diligence questions to expect
Where did each field come from?
Have a short collection history for every source. Buyers want to know whether data was generated by you, supplied by customers, bought from a vendor or gathered from the web.
Do you have the right to license it to us for this use?
Point to the clause or policy that permits it. If rights differ by source, say which records are covered.
Does it contain personal information or confidential third-party information?
Explain what was removed and how. Investment buyers also ask whether any data could be material non-public information about listed companies.
Has the method changed over time?
A change in how data was collected can look like a real-world trend. Document each change and when it happened.
What happens if you stop supplying it?
Buyers building products on a feed want notice periods and sometimes the right to keep historical data.
What this guide can't tell you
Sources
| Source | What it supports | Link | Checked |
|---|---|---|---|
| SEC press release 2021-176 | App Annie settlement, first alternative data fraud case, more than $10 million | https://www.sec.gov/newsroom/press-releases/2021-176 | 2026-11-07 |
Questions people ask
What kinds of companies buy data?
Mainly investment firms, AI labs and model builders, marketing and sales teams, and product or analytics teams filling gaps. Governments and researchers buy too, under their own procurement rules.
How do I find a buyer's data team?
Look for job titles such as data sourcing, data partnerships, data acquisition or alternative data on professional networks, and for people who speak at data industry events. A warm introduction from a broker or a shared contact helps most.
Should I offer a free trial of my data?
A limited trial or sample is normal, especially for investment buyers who need to test a signal. Put it under an NDA and a short trial agreement that limits use and requires deletion if they don't buy.
Will a company pay upfront for data?
Some pay on signing or on delivery; recurring feeds are often billed periodically. Payment timing is negotiable and belongs in the letter of intent and the licence.
What makes a company walk away from a data deal?
Unclear rights, personal data that can't be explained, a sample that doesn't match the full data, or a seller who can't answer diligence questions. Most of these can be fixed before you pitch.
Related
- How to sell data your company holdsHow to sell data your business already holds, from listing what you have to signing the licence and handing it over safely.
- How to sell data to AI companiesSell data to AI companies only after you can prove where it came from, that you may license it, and why it isn't online already.
- Where to sell dataWhere to sell data depends on the dataset: compare cloud marketplaces, listing sites, brokers, direct deals and asset sales on fees and fit.
Become a data broker
The DataBroker course and tools: find companies with data to sell, price it, license it and close the deal.