Deal economics
Dataset exclusivity opportunity cost: a scenario worksheet for owners
By SourceX Editorial · Reviewed by Noah Loul ·
10 sections
Short answer
Dataset exclusivity opportunity cost is the value of the licenses, internal uses and exit options you give up while one buyer holds exclusive rights. Estimate it as years locked, times the qualified licensees you would expect per year in the locked scope, times each later license's value relative to the first. Then test the assumptions that move it most and compare the result with the premium on offer.
License terms
Dataset exclusivity
Key takeaways
- Exclusivity costs the later licenses, own uses and exit options it rules out, not the records themselves.
- Estimate it as years locked x licensees per year x share of demand in the field x relative value of a later license.
- Express the result in first-license equivalents and compare it with the exclusive-to-non-exclusive offer ratio.
- Expected licensees per year is usually the assumption that moves the decision most.
- Own-use restrictions and successor-binding grants need separate review, not arithmetic.
- Narrowing term, field and covered records usually beats negotiating a larger premium.
How do you estimate the opportunity cost of granting data exclusivity?
Count what the exclusive grant stops you from doing, for how long, and how likely each of those things was to happen. In practice that means three numbers for each scope you would lock: how many years, how many other qualified licensees you would realistically expect per year, and what each of those later licenses would be worth compared with the first one.
Multiply them and you get the opportunity cost in first-license equivalents: a unit where 1.0 means the value of the license in front of you. Working in that unit keeps the discussion about structure rather than money, and it lets you compare an exclusive offer directly with a non-exclusive one from the same buyer.
Two costs sit outside that multiplication and have to be judged separately: restrictions on your own use of the records, and the effect of the exclusive grant on a future sale of the company. Both can outweigh the arithmetic when they apply.
Why exclusivity has a cost at all when data can be copied
Data is nonrival: one company's use of a dataset does not use it up for anyone else. Charles Jones and Christopher Tonetti's paper on the economics of data, published in the American Economic Review, builds on that property and argues that firms worried about rivals may hoard data that could be used by many firms at once. A non-exclusive license keeps that multi-use value available to you; an exclusive one sells it to a single buyer for a period.
That is the whole source of the opportunity cost. The records are not consumed by the first buyer, so the only thing an exclusive grant removes is your right to license the same records again, and sometimes to use them yourself. If no second buyer would ever appear, exclusivity costs little; if several would, it can cost more than the first license is worth.
The DOJ and FTC Antitrust Guidelines for the Licensing of Intellectual Property describe the same mechanics in contract terms: an exclusive license, or a partially exclusive one limited by territory or field of use, limits the licensor's ability to license others and possibly to use the property itself.
The exclusivity scenario worksheet with editable assumptions
Fill one copy of this worksheet for each exclusivity structure under discussion. The illustrative column shows the values used in the worked scenario further down; replace every one with your own estimate and a note on where it came from.
| Input | What it means | How to estimate it | Illustrative value |
|---|---|---|---|
| Locked scope | The records, fields and time periods the exclusive grant covers | Read the draft grant clause; list included systems, date ranges and any carve-outs | All work-order history, 2019 to 2025 |
| Field of use | The purposes for which only the buyer may use the records | Compare the field with the range of AI uses other buyers ask for | Any AI model training |
| Years locked | How long the exclusive period lasts, or the useful life of the data if perpetual | Use the term; for perpetual grants, use the years before the records stop being current | 2 years, or 5 years useful life if perpetual |
| Qualified licensees per year | Expected number of other buyers per year who would license this scope on acceptable terms | Count real inquiries, comparable datasets licensed in your sector and buyer requests that match your records | 0.6 per year for the full scope |
| Share of demand in the field | Fraction of those licensees whose use falls inside the locked field | Map past inquiries to the field definition | 100% for all AI training; one third for a single use case |
| Relative value of a later license | A later non-exclusive license as a fraction of the first license | Ask what a non-exclusive version of the first offer would be; adjust for data aging | 0.5 |
| Internal use reserved | Whether you may still use the records for your own AI and analytics | Read the restrictions on licensor use in the grant clause | Yes |
| Exit encumbrance | Whether the grant binds affiliates or a future acquirer | Check assignment, change of control and affiliate definitions | Binds successors; flagged for counsel |
The formula, and what each term leaves out
Opportunity cost in first-license equivalents equals years locked, times qualified licensees per year, times the share of their demand inside the field of use, times the relative value of a later license. Then compare: the exclusive offer, expressed as a multiple of the non-exclusive offer, should exceed 1.0 plus that opportunity cost before the arithmetic favors exclusivity.
The formula is deliberately simple so that every term can be argued about separately. It ignores the timing of later licenses, which matters when the exclusive period is long; a later license received in year four is worth less to you today than one received in year one, so discount it if your finance team uses a discount rate. It also ignores the cost of selling each later license, which is real staff time.
GAO's Cost Estimating and Assessment Guide recommends varying one assumption at a time to see which inputs move an estimate most, and presenting a range with a confidence level rather than a point. Apply that here before anyone treats the result as a number to negotiate on.
Which assumption has the largest effect on the decision?
Usually the expected number of qualified licensees per year, because it is the least observed input and it multiplies everything else. A company that has never been approached by a second buyer is guessing; a company that has turned down inquiries has evidence.
Years locked is next, and it behaves differently for fixed and perpetual grants. A two-year grant has a known ceiling. A perpetual grant has a ceiling set by how long the records stay useful, which for operational data depends on how fast workflows, products and regulations change in your sector.
Evidence for the arrival rate is closer to hand than most owners think: inbound inquiries in the last two years, requests from software vendors you already work with, and whether peers in your sector have announced data licenses. Write down where each number came from so the estimate can be revisited when a new inquiry arrives.
Show uncertainty by running a low, likely and high case for licensees per year and for useful life, and by reporting the break-even multiple each one implies. If the decision flips between the low and high cases, the honest conclusion is that you need more evidence about demand, not a sharper estimate.
Worked scenario: three exclusivity structures for the same records
Illustrative example only: the company, request and numbers are invented to show how the worksheet behaves. A field service company with seven years of work orders receives an offer for a license and is asked to consider exclusivity. Using the inputs in the worksheet above, it compares three structures, each expressed in first-license equivalents.
The broad perpetual grant needs the exclusive offer to be about 2.5 times the non-exclusive one before it breaks even, and that requirement swings from about 1.75 to 4 times as the arrival rate moves from 0.3 to 1.2 licensees per year. The narrow, one-year grant barely moves the result in any case. The structure, not the premium, is where the negotiation has most leverage.
| Structure (illustrative) | Calculation | Opportunity cost | Break-even multiple | Main uncertainty |
|---|---|---|---|---|
| Perpetual, all AI training, all records | 5 useful years x 0.6 x 100% x 0.5 | 1.5 first-license equivalents | 2.5x the non-exclusive offer | Useful life and arrival rate |
| 24 months, all AI training, all records | 2 years x 0.6 x 100% x 0.5 | 0.6 first-license equivalents | 1.6x | Arrival rate |
| 12 months, one use case, refresh data excluded | 1 year x 0.6 x one third x 0.5 | 0.1 first-license equivalents | 1.1x | Field definition |
| Perpetual grant, low arrival case | 5 x 0.3 x 100% x 0.5 | 0.75 first-license equivalents | 1.75x | Whether 0.3 is too pessimistic |
| Perpetual grant, high arrival case | 5 x 1.2 x 100% x 0.5 | 3.0 first-license equivalents | 4.0x | Whether demand evidence supports 1.2 |
Costs the worksheet cannot multiply: your own use and a future sale
Check whether the grant limits your own use of the records before you compute anything. If an exclusive field of use reads as covering your own AI features, the cost is no longer a set of foregone licenses but a constraint on your product roadmap, and that belongs in a separate conversation with the people who own it.
Then check what the grant does at exit. A grant that binds affiliates and successors travels with the company, and a buyer of the business will treat it as an encumbrance on the data. Time-limited, field-limited grants are easier to explain in diligence than open-ended ones.
In U.S. copyright law, an exclusive license counts as a transfer of copyright ownership, unlike a non-exclusive license, and a transfer generally must be in a writing signed by the owner. Business records are often only partly copyrightable, so whether this matters depends on the content; counsel should review the grant language in either case.
When exclusivity terms need competition or legal review
Most exclusive arrangements are ordinary commercial terms. The FTC's guidance on exclusive dealing says such contracts are generally lawful and that the potential for harm grows with the length of the term, the share of sources covered and the scarcity of alternatives. The DOJ and FTC licensing guidelines add that exclusive licenses raise concerns mainly where the parties are in a horizontal relationship, such as competitors.
Ask for review when the buyer competes with you, when your records are among very few sources of their kind, or when the grant would also bar you from licensing similar data you have not yet created. These are the cases where scope wording matters more than the multiple.
- The buyer operates in your market or plans to.
- Your records are one of a small number of comparable sources.
- The grant reaches future records, not only the delivered history.
- The grant binds affiliates, portfolio companies or an acquirer.
- The field of use is undefined or reads as all uses.
Narrowing the grant before negotiating the premium
Each worksheet input is also a negotiating lever, and narrowing the grant usually beats arguing about the multiple. Shorter terms cut years locked, a defined field of use cuts the share of demand, and excluding refresh data or future records protects the value of what you have not yet delivered.
The OECD's work on measuring the economic value of data distinguishes market-based, cost-based and income-based approaches. This worksheet is a simple income-based view, so pair it with a market check: ask the buyer what the non-exclusive version of the same offer would be, because that ratio is the cleanest evidence of how they value the lock.
- Replace perpetual with a fixed term, or a term that ends when a refresh is delivered.
- Define the field of use by model type or task rather than all AI uses.
- Limit the grant to the delivered records and exclude future periods.
- Reserve your own internal use in writing.
- Exclude affiliates and make the grant end or narrow on change of control.
How exclusivity is handled when licensing through SourceX
For owners who license through SourceX, the worksheet feeds a process in which SourceX runs rights review, contracting, preparation and delivery. Exclusivity and its term are set out in each deal's signed agreement, and the company approves the scope and terms and signs before any data is delivered, so the worksheet's inputs can be settled on paper first.
The first step is a short company assessment that needs no data uploads, which is a sensible point to raise any exclusivity question you already expect.
Frequently asked questions
What is a first-license equivalent?
It is a unit equal to the value of the license currently on the table. Expressing foregone licenses in that unit lets you compare structures without putting a money figure on them, and makes the break-even multiple easy to read.
Is perpetual exclusivity ever reasonable for business records?
It can be when the records age quickly, demand from other buyers is unlikely and your own use is reserved, because the useful life caps the cost. It is hardest to justify when the records stay current for years or several buyers have already asked about them.
Does an exclusive license stop me using my own data?
It depends on the grant clause. Some exclusive grants restrict only licensing to third parties, while others also restrict the licensor's own use within the field. Read the clause for licensor restrictions and ask for internal use to be reserved explicitly.
How should uncertainty be shown to a board or co-owner?
Show the low, likely and high cases for the inputs you are least sure about, usually licensees per year and useful life, and the break-even multiple each implies. If the decision changes across that range, say so and describe what evidence would narrow it.
Can exclusivity be limited to part of a dataset?
Yes. Grants can be limited by field of use, territory, record type, date range or term, and refresh deliveries can be excluded. Each limit lowers the opportunity cost in the worksheet and is often easier to agree than a larger premium.
Sources
- Jones and Tonetti's paper on the economics of data argues data is nonrival and that firms may inefficiently hoard data that many firms could use at once. Source
- The DOJ/FTC IP licensing guidelines say an exclusive or partially exclusive license limits the licensor's ability to license others and possibly to use the property itself, and raises concerns mainly in horizontal relationships. Source
- FTC guidance says exclusive dealing contracts are generally lawful and potential harm grows with contract length, sources covered and fewer alternatives. Source
- 17 U.S.C. 101 defines a transfer of copyright ownership to include an exclusive license but not a nonexclusive license. Source
- 17 U.S.C. 204(a) requires a transfer of copyright ownership, other than by operation of law, to be in a writing signed by the owner or an authorized agent. Source
- The OECD note on measuring the economic value of data describes market-based, cost-based and income-based valuation approaches. Source
- GAO's Cost Estimating and Assessment Guide recommends varying one assumption at a time and presenting a range with a confidence level. Source
See if your company qualifies
A short company assessment. No data uploads are needed.