When a lab asks for your data, you are the only one in the room without a price in mind.
Dividata represents the seller. We assess what your operational data is worth, define what leaves and what never does, and put it in front of buyers who are used to negotiating against nobody.
No retainer. The assessment is free and there is no obligation to proceed.
The buy side is smaller than you think.
The set of organisations buying operational data at scale is small, they speak to each other, and most of them have bought from an unrepresented company before.
Dividata knows who is buying, what they have been paying attention to recently, and which of them will bid against each other on a given package.
A single interested buyer is not a market. Our job is to make sure there is never only one.
Counts indicate active coverage. No organisation is identified.
How a deal runs.
Five defined stages, each with written outputs and one decision that is yours before the next begins.
- 01
Approved sample assessment
You share a small, pre-cleared slice.
We assess workflow depth, domain rarity, volume and how messy and real it is, because clean synthetic-looking data is worth less than the genuine article.
Outputs
- A written assessment
- An indicative structure
- A view on whether there is a deal at all
Your decision
You decide whether to run a process.
- 02
Package definition
We work out what is in, what is out, and what never leaves the building.
Customer data, HR files and anything under client NDA are ring-fenced at this stage rather than negotiated later.
Outputs
- A defined package
- A written exclusions list
- A rights position on each category
Your decision
You sign off the scope before anyone sees it.
- 03
Competitive process
The centre of the deal, and the reason to have a broker at all.
We take the package to labs, data vendors and RL gym builders, and solicit bids in parallel rather than one after another, because sequential conversations let each buyer set the price alone.
Outputs
- A bidder list for your approval
- Indications of interest
- A comparison of terms
- A written comparison of every bid received, including the ones you rejected
Your decision
You choose who stays in the process.
Bid comparison, illustrative
Bidder Afrontier lab3-yr licenceshortlistedBidder Bdata vendorone-off saledeclinedBidder CRL gym builder2-yr licenceshortlistedBidder Dfrontier lab3-yr exclusivedeclinedBidders are referenced, never named, until you approve the list.
- 04
Specialist review
Privacy counsel and a redaction partner work on the package before anything changes hands.
This happens before transfer, not after a problem.
Outputs
- A redaction pass
- A privacy review
- A clean rights position
Your decision
You approve the reviewed package.
- 05
Buyer vetting and close
We establish who is receiving the data, what they will do with it, and negotiate scope, permitted uses, exclusivity, retention, deletion and payment terms.
Outputs
- A signed licence
- A deletion schedule
- An audit position
Your decision
You sign or you walk.
Five decisions are yours. Everything between them is ours, and nothing moves without your written sign-off.
The same data, at four different prices.
A one-off sale and a recurring licence are the same data at very different prices, and which one you agree is the single biggest lever in the deal. Move the terms and watch the relative position change.
Structure
A licence with annual refresh makes your ongoing operations part of the asset, and gives you a reason to speak again next year.
Exclusivity
Exclusivity is worth paying for, and it removes every other bidder from the table for the term.
Permitted use
Training only confines the data to model development inside the buyer.
Retention
A fixed term with a deletion schedule keeps the grant reversible and dated.
Relative value indicator
40 / 100
Relative illustration only. Not a valuation.
These are the arguments the buyer's counsel is prepared for. Our job is to be on your side of that conversation.
What labs actually pay for.
Workflow depth
The sequence of steps, including the corrections. A record of how a decision was reached is worth more than the decision.
Domain rarity
Data nobody else holds, from a field with few operators.
Volume with continuity
Eleven years from one operation beats eleven months from twenty.
Mess
Real annotations, abandoned attempts and inconsistent formats. Tidied data looks synthetic and prices accordingly.
Most owners assume their data is worth less than it is, and tidy it up before showing anyone, which makes it worth less than it was.
What a closed deal looks like.
A composite illustration showing the shape of a closed deal. It is not a client and no figures are actual.
- Seller
- Mid-sized industrial services business, 22 years trading
- Package
- 340,000 inspection records, 11 years, one territory
- Composition
- 212,400 reports / 94,800 notes / 28,300 photographs / 4,500 schematics
- Ring-fenced
- Customer identities, pricing, anything under client NDA
- Structure
- 3-year non-exclusive licence, annual refresh
- Permitted uses
- Model training, internal evaluation. No derivative distribution
- Retention
- Term plus 90 days, then deletion
- Deletion
- Written certification, per-refresh schedule, audit right retained
- Buyer
- Frontier lab, described by category only