Govern before you sell: everyone wants your archive
$100,000. $500,000. $1,000,000+. Those are published entry tiers on data-purchase programmes, aimed at ordinary companies with ordinary operations.
Someone is buying what your company knows.
TL;DR
A market now pays companies cash for their operational data: published tiers run from $100K to $4M+, across dozens of buying programmes.
Almost none of those programmes' own pages address the hard questions: what may leave, under what consent, with what provenance.
The consent gaps are visible in the buyers' own words.
Govern before you automate, and before anyone sells.
The wave
The figures here come from a public register of data buyers and sellers, which tracks 159 vendors and 38 buying programmes across 29 companies. All amounts are the programmes' own published terms, as of early October 2026, and they move, so treat them as a snapshot rather than a price list.
The tiers, in the programmes' own words:
Programme | Published tiers |
|---|---|
micro1, Enterprise Data Partnerships | $100K+ / $500K+ / $1M+ |
Handshake AI | $100K to $4M+ |
Turing, "Project Lazarus" | up to $1M |
Labelbox, "Prism" | ~$50K to $1M+ per licence |
Mercor, Mode, Scale, Polyshares | similar ranges ($10K to $5M across programmes) |
There's also a bounty layer: $10K to $50K+ paid per company that signs. micro1's own bounty moved from $25K to $50K between June and October. These are vendors competing for your door.
What they buy, and what nobody is buying
Read the programmes' own descriptions, because they're specific about what they want:
micro1 buys "SOPs, knowledge bases, CRM data, project histories and QA processes", along with patterns of "decision-making".
Turing wants "the company's operating record: communications, tickets, wikis, knowledge bases, databases and source code."
Handshake asks for "business process flows, hiring logic and document-level communications."
Notice what that is, collectively: the archive. What happened to be recorded: messages, docs, tickets, footage of how things went.
And notice what's missing. Across all 38 programmes we reviewed (as of this writing), none buys verified execution as a category. That would mean a trusted record of how the work actually runs, and who approved it. They buy what you happened to have lying around, and nobody is paying a premium for knowledge that holds up.
That gap is going to close. When it does, whatever you've governed, and whatever you've sold, is the position you'll be defending.
Calibration, before anything else
Most coverage skips this part. Doing it early.
"A listed amount is not an offer." That's the register's own disclaimer, and it's the right lens. Published tiers are marketing until a signed contract says otherwise.
Some hype markers to keep visible: one broker claimed "$4m MRR" and, when pushed, corrected it to "$2.8m unrealised". An "ARR" claim of $48M surfaced two weeks after a company launched. Check incentives everywhere: the list that surfaced much of this wave carries referral codes, and the register itself is maintained by a data-masking vendor (disclosed on the site) that may sell services to the companies it lists. None of that makes the market fake. It does mean the comfortable version of the story, where your operational data is a new oilfield, is being sold to you by people holding shovels.
The consent gap
This is where the programmes' own pages get quiet.
Read the listings closely and you find the gaps in the sellers' own words: one doesn't mention consent from employees whose messages sit inside workspace data. Another anonymises automatically, with no manual review. A third simply "removes identifying details on its side."
The questions almost nobody's FAQ answers:
Whose knowledge is it: the company's, or the people who made it?
What exactly leaves, and in what form: the export, or a reconstruction of the workflow?
What happens to your edge after the deal? What, precisely, is the moat you didn't just sell?
And what may never leave, at any price?
What to do (govern first)
Six questions to answer in writing, with names against them, before anyone entertains an offer:
- Inventory. What operational records exist, where, and in what form?
- Red lines. What never leaves, under any price? (Write this down before there's money on the table. It's much harder after.)
- Consent. What have employees, customers and contracts actually agreed to?
- Provenance. Who reviewed the record, which version is current, and what may it be used for?
- Quality. Is the archive stale, or does it match how the work runs today? (Buyers are already asking this. You should ask it first.)
- Escalation. Who approves: one name, not a committee.
Govern before you automate, and before anyone sells.
Where capture fits
One footnote from where I sit. I build operational memory capture tooling, and MimeOps is my service; this wave lands on the same desk as capture. Masking at capture and session boundaries exist in MimeOps today; the broader evidence-and-governance layer is our direction of travel, not a shipped claim. This part isn't about product: if you start capturing operational knowledge, with any tool, govern it from the first session, because the market will eventually find you whether you went looking or not.
Your operating record is suddenly an asset. Make sure it's an asset you govern.
The Recovery Advantage series. Previous: "The dirt path: why nobody documents anything, and what to build instead". Next: "From SOPs to recovery infrastructure" (publish URLs pending).