Skip to content
Outrich Data

What buyers say moves the price of company data

Published 2026-11-17 · Updated 2026-11-17 · 7 min read · By the Outrich Data team

Three buyers of company data publish what they look at when they price it: DataPaid, Polyshares and micro1. Their lists overlap on years of history, clean rights and a specialized domain. They also agree on what lowers value: records held for clients, heavy personal data and content that is already public.

Why this is worth reading closely

Most pages about the price of company data start with a range. A range is the least useful part. It tells you where deals have landed. It does not tell you why one company sits at the bottom and another at the top.

The better question is what a buyer looks at when it prices a dataset. Three buyers answer it on their own pages. We read each page on November 17, 2026, and we report only what the page says, with its date. We do not repeat a claim we could not read ourselves. The comparison of ranges, minimum company size and payment timing is on our page about who buys company data for AI. This post is about the factors.

What DataPaid lists

DataPaid's page on what company data is worth is dated October 13, 2026. It names five factors: size and volume, workflow complexity, data quality, domain expertise and uniqueness. It then expands on several of them.

  • Depth over volume. Connected workflows are worth more than disconnected fragments.
  • Years of history. Longer records show how decisions change over time.
  • Specialized expertise. Work that takes trained judgment, such as underwriting or claims review, is valued higher.
  • Multiple teams. Data that spans support, sales, finance and operations is worth more.
  • Clean rights. Clear ownership that allows licensing without client consent problems makes a deal easier and raises the price.

What DataPaid says lowers a price

The same page lists what pulls value down. These are knowledge held mostly in people's heads and not written down, non-English documentation, heavy personal or regulated data that takes extensive removal, records held for clients and not owned, and generic content that resembles what is already public.

The page states ranges as well. It says published payouts for established companies start at $100,000 and can reach $2 million for the largest and most unique datasets. It cites Fast Company, dated April 17, 2026, for smaller archive sales of $10,000 to $100,000. We have not opened those secondary sources and give them only as DataPaid reports them.

What Polyshares lists

Polyshares' answer page on what company data is worth is marked updated September 2026. It says its deals typically land between $100K and $2M paid to the company, and that the figure quoted is the figure the company receives, with nothing taken out of it.

It frames value around the record more than the company. In its words, the length and texture of the record matter more than headcount or revenue. The factors it names that raise value are the depth of operating history, visible reasoning, a specialized domain with few written records, and clean rights, meaning provenance that does not have to be litigated. One of them is unusual: visible disagreement and correction, a record of people being wrong and then getting it right.

What Polyshares says caps or removes value

The page lists four cases. A short operating history makes the record too thin. Records the company holds but does not own belong to the client under a master agreement. Personal information woven through the record that cannot be cleanly stripped lowers value. Material that is already public, published, scraped or widely syndicated adds little.

What micro1 asks for

micro1 does not list factors in prose. Its data estimate page shows what it asks, which is a different kind of evidence. We read the page on November 17, 2026. The estimate uses six answers: number of employees, years of data, number of business entities, share of data in English, headquarters location, and whether the data includes regulated or sensitive information.

The page labels its output as an indicative estimate only, not a quote or binding offer. It showed a figure range when we opened it, but we do not repeat the numbers, because they came from inputs we chose and mean nothing for any other company.

Where the three agree

Set side by side, the lists share a core. Each treats these as price drivers.

DriverDataPaidPolysharesmicro1 form
Years of historyListedListedAsked
Clean rights and ownershipListedListedNot asked
Specialized domainListedListedNot asked
Personal or regulated dataLowers valueLowers valueAsked
Records held for clientsLowers valueLowers valueNot asked
Language of the dataNon-English lowers valueNot statedAsked
Company sizeNot statedMatters less than the recordAsked

What this means for a company preparing data

Most of what moves the price is not a technical property of the files. It is a property of the company's rights and records, and a company can settle those questions before it talks to anyone. Our guide to AI training data pricing and data valuation methods go through the mechanics.

Three steps cost little and affect the offer. The first is a rights review. Find out which records you own and which you only hold for clients, using the client data vs company data guide and the free data rights pre-check. The second is an inventory of what you have, by system and by year, using the data inventory worksheet. The third is a de-identification plan. See how to anonymize company data.

The data valuation calculator on this site shows an illustrative range and the method behind it. A written offer from Outrich Data follows a rights review, never before it.

How to use the lists before you ask for an offer

A list of factors is most useful as a self-check. Go down the shared drivers one at a time and write an honest answer for your own company. How many continuous years of records do you hold in each system. Which of those records belong to you, and which belong to a client under a master agreement. Which work in the company takes trained judgment, and is any of it written down in tickets, reviews or documents.

Then do the same for the factors that lower value. Estimate how much of the text is not in English. Note which systems hold regulated or sensitive data and how much removal each would need. Mark anything that is copied from public sources, because buyers say it adds little. The answers will not give you a price. They will show you which questions a buyer is about to ask, and which of them you can settle first.

Keep the answers in one place. A short written summary, with a date, makes a rights review faster and gives your counsel something concrete to check. It also protects you from describing your data more favorably in a call than your records support.

What we cannot tell you

These pages are marketing pages. They describe what each buyer says it values, and they do not publish its pricing model or any deal. A buyer's list of factors tells you what is in the conversation, not how much each factor is worth. Dates matter too: DataPaid's page is from October, Polyshares' from September, and each can change.

Our own page is no different. Outrich Data does not publish a price list, because an honest offer depends on the title search. What we publish is the process and the commitments. See how it works and sell your company's data.

Common questions

What raises the price of company data most?
Buyers name several factors without ranking them. DataPaid and Polyshares both list years of history, a specialized domain and clean rights. Neither page says how much each factor adds, so treat the list as a guide to what a buyer will ask about.
Do records held for clients count?
Both DataPaid and Polyshares say records a company holds for clients, and does not own, lower value or remove it. They generally belong to the client under the master agreement. This is general information, not legal advice.
Does company size set the price?
Polyshares says the length and texture of the record matter more than headcount or revenue. micro1's form asks for employee count, so buyers do not agree. Eligibility minimums differ by buyer and are listed on our comparison page.
Are the ranges on these pages reliable?
They are what each buyer says about itself on a dated page. They are not audited, and they describe the buyer's own deals. Use them to see where offers tend to start, then ask for a written offer.

Find out if your data qualifies.

Eleven questions, about five minutes, no call required. You get an answer on screen and a written follow-up.

esc