Method
How the calculator works
The starting point is an offer made to a 30-person operating company, about $120K [3]. That amount scales with company size, then moves for history, systems, documentation, field and status. Media you own is added at published per-unit prices [1].
- $Company size$120K at 30 staffScales with headcount
- ×Years of history×0.62 to ×1.61 to 30 years
- ×Connected systems×0.6 to ×1.31 to 8 systems
- ×Documentation×0.8 to ×1.2Little to extensive
- ×Field×1 to ×1.25General or specialist
- ×Status×0.5 to ×1Closed or operating
- =Most likely valueRange ×0.55 to ×1.8of the most likely figure
| Factor | Effect on value | Why |
|---|---|---|
| Company size | $120K at 30 employees, $260K at 100, $750K at 500, $3.3M at 5,000 | This is the heaviest factor. Buyers price whole companies rather than messages, and a bigger company spans more teams and workflows [2][3]. |
| Years of history | 1 year ×0.62, 5 years ×1, 10 years ×1.23, 20 years ×1.52 | Long-running records reveal how decisions played out over time; company age is itself a stated price factor [2]. |
| Connected systems | 1 system ×0.6, 4 systems ×1, all 8 ×1.3 | Work that can be traced from chat to ticket to code to customer commands more than one tool’s export [2]. |
| Written documentation | Extensive ×1.2, Some ×1, Little ×0.8 | SOPs, knowledge bases and decision records are the first things operational-data buyers ask for [3]. |
| Field | General business ×1, Specialist field ×1.25 | Ashgrove Data’s judgment: specialist work in finance, healthcare, legal and similar fields is scarcer in public data. |
| Company status | Operating ×1, Closed or winding down ×0.5 | A closed company sells exactly once and has nobody left to answer buyer questions; shut-down archives go for well below operating-company offers [2][3]. |
The range shown runs from ×0.55 to ×1.8 of the most likely figure. The size curve and every factor are Ashgrove Data’s fit to the public deals below, not a published formula.
Calibration
Checked against public deals
As a check, we run each public deal back through the calculator using the known facts about the seller. Every published figure lands inside the calculator’s range.
| Case | Inputs we used | Published | Calculator | Source |
|---|---|---|---|---|
| Shut-down startup archive | 20 employees, 4 years, Slack, email and Jira, closed | $10,000 to $100,000 typical | $17K to $56K inside | [2] |
| Qualified operating company | 30 employees (the minimum), 5 years, 4 systems | More than $100,000 | $66K to $220K inside | [3] |
| Large, multi-team operating company | 500 employees, 8 years, 5 systems, well documented | More than $500,000; more than $1M if highly unique | $620K to $2M inside | [3] |
| Spirit Airlines enterprise data | About 13,000 employees, 18 years, 6 systems, closed, airline operations | Bids $5M to $7.5M; Google agreed $10M | $4.4M to $15M inside | [5] |
Evidence
Published prices for company data
No public rate card exists, and most contracts stay confidential [6]. Only the figures already public appear here.
| Data | What counts | Published benchmark | Source |
|---|---|---|---|
| Operating company data | Chat, email, tickets, SOPs, CRM, project history; 30+ employees | More than $100,000 per qualified partnership; more than $500,000 for large datasets; more than $1M for highly unique data | [3] |
| Shut-down company archives | Slack, email and Jira from closed startups | $10,000 to $100,000 typical, across about 100 sales | [2] |
| Enterprise archive | Spirit Airlines: email, Teams, files, tickets, code since 2008 | $10M agreed by Google (bids from $5M); about 100M emails and 500M chat messages | [5] |
| Video footage | Recorded footage you own the rights to | $1.05 to $6.20 per minute (median $2.65), across 29 deals | [1] |
| Recorded speech | Calls, meetings, interviews, podcasts | $22 to $141 per hour of transcribed speech (median $58), across 17 deals | [1] |
| Images and photographs | Product, field or archive imagery | $0.41 to $3.48 per image (median $1.12), across 38 deals | [1] |
| Published articles | News, research notes, editorial archives | $0.18 to $1.30 per article (median $0.46), across 14 deals | [1] |
| Books and long-form titles | Titles you hold the rights to | $1,850 to $9,600 per title (median $4,300), across 21 deals | [1] |
Media ranges are the middle half of reported deals (interquartile range). Operational-data figures are published program tiers, typical payouts and a court-supervised sale.
What the calculator leaves out
It comes down to who owns the rights, the permissions in your privacy notices, record cleanliness, and exclusivity of the license. Any one of them can swing an actual offer well beyond the range, up or down. The calculator prices records. It says nothing about your right to sell them.
From estimate to offer
A firm valuation follows a set sequence: inventory and a rights check, then de-identification, then a buyer process. Send over what you entered and we come back with its likely worth and the buyers who would take it.
Request a valuationQuestions
About the calculator
How accurate is the company data value calculator?
You get a range, not a price. The model is tested against four public data points, from shut-down startup archives to the Spirit Airlines sale, and every one falls inside its range [2][3][5]. That range is wide on purpose: rights, data quality, exclusivity and buyer demand move real offers.
Why does company size matter so much?
Buyers price a company’s records as a whole, not message by message. Larger firms span more teams, workflows and edge cases, so size lifts the price. The climb trails headcount, though: 10 times the staff works out to roughly 4.5 times the value.
Do I need 30 employees to sell company data?
Not necessarily. The biggest operational-data program requires 30 or more employees and documented workflows [3], yet smaller and closed companies still change hands, typically via brokers and wind-down buyers at $10,000 to $100,000 [2].
Is a closed company’s data worth less?
Usually. The sale happens once, and no one stays behind to field buyers’ questions. Even so, size drives the price: the bankrupt Spirit Airlines attracted bids of $5 million to $10 million for its records [5].
Are video, audio and text priced the same way?
No. Media sells by the unit at rates from disclosed licensing deals: a median $2.65 per minute of video, $58 an hour for transcribed speech, and $1.12 per image [1]. Factor these in at step 4 of the calculator.
Sources
- [1] AI Industry Reviews, “AI Training Data Licensing Deal Statistics: 2026 Report”, 6 October 2026. 312 licensing agreements signed January 2023 to September 2026; per-unit prices are medians and interquartile ranges from deals that disclosed unit pricing.
- [2] Forbes (Anna Tong), “AI's New Training Data: Your Old Work Slacks And Emails”, 16 April 2026. SimpleClosure reported about 100 sales of shut-down companies’ Slack, email and Jira archives, typically paying $10,000 to $100,000 each; a rival wind-down firm said price depends on company size, age and how rich the data is.
- [3] The Next Web, “micro1 commits $1bn to company data for AI agents, financed by Citi”, 9 October 2026. micro1 plans to spend $1 billion in 12 months licensing operational data from companies with 30 or more employees: more than $100,000 per qualified partnership, more than $500,000 for large or multi-team datasets, more than $1 million for highly unique data. Companies keep ownership.
- [4] Reuters, “Google to buy Spirit Airlines business data for $10 million”, 17 August 2026. Google agreed to pay $10 million for internal business data from the bankrupt airline, de-identified before transfer, to improve its products and AI models.
- [5] The Next Web, “Google picked and paid for the firm anonymising the Spirit Airlines data”, 18 August 2026. The package: about 100 million emails, 500 million Teams messages, SharePoint and OneDrive files, IT tickets and 516 code repositories, with records back to 2008. Bids: Google opened at $5 million, Mercor reached $7.5 million, Google won at $10 million.
- [6] Troveo, “AI Training Data Cost: What Buyers Actually Pay”, 27 August 2026. There is no public price list for AI training data and most contracts are confidential; Reddit’s licences reportedly bring in about $60 million a year from Google and $70 million a year from OpenAI.