A recognisable customer will let you publish a case study with their name and a number in it. They want twenty percent off, permanently, in exchange.
Everyone in the room is valuing the logo. You are trading a price that resets at every renewal for a piece of proof that stops being interesting in eighteen months.
Why these three models
The decision is whether to discount in exchange for a public reference, and how to structure it if you do. The features that fire are a claim whose worth depends entirely on what it verifies, a piece of evidence that travels through a specific network rather than broadcasting, and two commitments with completely different lifespans.
Three lenses. Certification produces an equilibrium answer about what you are buying. Spread produces a cycle answer about where it goes and therefore whether you want it. Options produce a random answer about the asymmetry in duration, which is the part that makes an apparently fair trade expensive.
1. What does the case study actually certify
Start with what the artefact proves, because case studies vary enormously in that respect and are priced as though they do not.
A claim carries information in proportion to what it would cost to make falsely.1 A quotation saying the team has been great to work with costs the customer nothing and could be said about anyone, so it verifies almost nothing. A named person stating a specific number, which their own colleagues will read, is expensive to say untruthfully and therefore carries real information.
Which means the negotiation should be about the content rather than about the discount. The things worth paying for, in rough order: a number they will stand behind, their name and title attached, willingness to take a reference call from a prospect you choose, and appearing at something. The things not worth paying for: a logo on your website, and an approved quotation written by their marketing department.
A logo is a decoration. A number with a name on it is evidence, and they are not the same purchase.
The distinction has a practical test. Ask whether a competitor could obtain the identical artefact from the identical customer. If they could, and for most logo placements they could, you are buying something that does not separate you from anyone, and the discount is being spent on decoration.
The corollary founders miss: the customer usually knows which version costs them more, and asking for the expensive version first frequently produces a smaller discount request than asking for the cheap one, because the conversation has become about value rather than about price.
Asking for the expensive version has a failure mode worth seeing coming. A specific number attached to a named person is exactly the sentence a communications team strikes in review, because their job is to reduce what their employer can later be held to. You discover this six weeks in, after the discount has been agreed. The tell is available on the first call: ask who signs off on an external mention of a figure, and whether that person has ever released one. A company that has never published a number about itself is not going to start with yours, whatever your sponsor promises in the meeting.
Put the trade in currency to see its shape. A fleet tracking product sells at UGX 900,000 per vehicle per year. A distributor running 60 vehicles asks for twenty percent off in exchange for a case study, which is UGX 10.8m a year for as long as they stay. Against an artefact with an eighteen-month useful life, the first cycle alone costs about UGX 16m. Written down that way the founder stops negotiating the percentage and starts asking what the document has to contain to be worth that, which is the correct conversation and the one the discount frame prevents.
2. Who will actually see it, and does that matter
The second lens asks where the proof travels, because a reference does not broadcast, it moves through people who already talk to each other.
Evidence spreads along existing connections: a buyer asks someone in their own sector, a procurement lead checks with a peer, a prospect looks for someone like them.2 A case study is useful to the extent that it sits in the path of those conversations, which depends far more on who the customer is than on how well the document is written.
That reframes which customer to trade with. A large recognisable name in a sector you do not sell into is worth less than a mid-sized customer who is central in the exact segment you are targeting, and the second usually asks for a smaller discount because they are less aware of what their name is worth.
So before agreeing anything, write down the three prospects you expect this reference to influence. If you cannot name them, the case study is being bought for general credibility, which is a real thing and worth much less than a specific one, and should be priced accordingly.
What makes a reference work at all is worth naming, because it explains why the customer’s identity does more than the document’s quality. A prospect who cannot easily evaluate you substitutes someone else’s decision for their own assessment, which is why a name they recognise in a position like theirs settles the question and a well-written page from a stranger does not.4 You are not supplying evidence, you are supplying someone to copy.
The reference call is the part that actually travels, and it is the part usually left out of the agreement. A published case study is read by a few people who found your website. A customer who takes calls reaches prospects at the moment they are deciding, which is where the value is. Negotiate for the calls and treat the document as the by-product.
Then treat those calls as a consumable, because they are. A customer who enjoys the first two will decline the eighth, and the decline arrives in the quarter your pipeline is busiest, which is exactly when you were counting on it. Cap the number in the agreement at something you can say out loud, four calls a quarter is generous, and spend them on deals late in the cycle rather than on curious first meetings. Start earning the second reference before the first one tires. One reference customer is a single point of failure wearing the costume of an asset.
If they refuse publication outright, there is a cheaper version that keeps most of the value. Ask for permission to name them verbally on a call, and to show a one-page result to a named prospect in confidence. That route needs no communications review, no legal read and usually no discount, because nothing is published and nothing can be quoted back at them. It still reaches the three prospects you wrote down, and it removes the failure where you grant a rate concession for a document that never clears approval.
3. The two things you are trading have different lifespans
The third lens is the one that turns an even-looking trade into a bad one, and it is entirely about duration.
A permanent discount does not merely cost you twenty percent of that account. It sets a reference price for every renewal with that customer, it becomes the number their procurement team defends, and it establishes what your product costs for anyone who learns of it. It is close to irreversible: raising it later means an argument with a customer who is also your public reference, which is the worst possible negotiating position.
The case study decays. A result from two years ago is discounted by prospects, the named person changes jobs, the numbers stop matching your current product. Its useful life is measured in months and its cost is measured in years.
Where one side of a trade is permanent and the other expires, the structure should be fixed rather than the price argued.3 Pay with something that has the same lifespan as what you are getting. A discount for twelve months, renewing only if the reference is renewed. A one-time credit rather than a rate change. Additional service, extra seats, early access, or a feature they want, all of which cost you once. Any of those is better than a permanent rate for a temporary asset.
The term-limited discount has a failure of its own, and it arrives at month twelve. The customer has budgeted a year at the lower rate and will experience the return to list as a price rise, which is the argument you were trying to avoid, now happening later and with a public reference sitting on the other side of the table. Write the step back into the original contract as a schedule: twenty percent in year one, ten in year two, list from year three. A scheduled increase is an administrative event. An unscheduled one is a negotiation, and you will lose it.
When they will not accept any expiry at all, price the permanence rather than refusing it. A permanent rate is defensible when it buys something equally permanent: a three-year term, a committed annual minimum, or the year paid upfront in one transfer instead of monthly. Each of those converts a concession into a purchase you can account for. What you must not do is hold a permanent price against an obligation the customer can walk away from in a quarter, which is precisely what an open-ended case study agreement is.
And put an expiry on their side too. The case study is current for twelve months, after which you both agree to refresh the numbers or retire it. That is normal, it protects them from a stale claim in their name, and it is much easier to agree at the start than to raise later.
What the three say together
- Negotiate the content before the price. A number with a name and a willingness to take calls, not a logo.
- Ask whether a competitor could get the same artefact from the same customer. If yes, it separates you from nobody.
- Name the three prospects this reference should influence. If you cannot, price it as general credibility.
- Pay with something that expires: a term-limited discount, a one-time credit, or service. Never a permanent rate.
Where they disagree
The certification argument and the option argument recommend different things when the customer refuses the structure.
Certification says the expensive version of the reference is worth a great deal, particularly for an early company being priced off its category, and that a genuinely strong proof point can be the difference in every subsequent sales conversation. If a permanent discount is what it takes, the arithmetic may still work.
The option argument says a permanent rate concession compounds in ways nobody models, spreads to other customers when it becomes known, and is defended by the customer at every renewal, and that no eighteen-month asset justifies it.
The reconciliation is that the answer depends on how early you are. A company with no references at all is buying its first proof, and the first one is worth disproportionately more than the fourth, so paying badly for it can be correct. A company with three case studies is buying an increment and should not pay permanently for it. Founders usually apply the early-stage logic well past the point where it stopped being true, because the first negotiation set the pattern.
What none of them contain
None of the three prices the customer’s own risk. Putting their name on your product is a small bet by them on you, and if you fail publicly they carry some of it. That is why the good ones negotiate carefully, and it is a reason to treat the request as a partnership rather than as a purchase.
None of them handles the approval chain. The person agreeing this is rarely the person who can authorise it, and a case study that dies in their legal or communications team six weeks later has cost you a discount you already granted. Make the discount conditional on publication, not on agreement.
None of them prices the work. A case study is not a document you write and send for approval; it is two calls, a draft, a review cycle through their communications team and usually a legal read. That is real time from you and from them, and a discount granted for something that never gets published is the most common failure in this whole area.
And one property the ensemble will not produce: your other customers will see the case study and some will conclude that a public reference is worth a discount. That is a rule you have just created, and the second customer to ask will be less useful and equally expensive, which is an argument for having decided the standard trade before the first one.
The one action that survives the ignorance: before agreeing anything, make the discount conditional on publication and time-limited to twelve months. Both are normal, both are usually accepted, and together they convert a permanent cost for a decaying asset into a matched trade.
Who has to move
The founder, at the point the customer first raises it, because the structure is easy to set before a number has been discussed and hard afterwards. The instinct is to negotiate the percentage, which accepts the shape of the trade and argues about its size. The cheapest first test is to ask what they would want in exchange for a reference call rather than a case study. The answer is usually much smaller, and the call is the part that actually reaches the people you are trying to convince.
Sources and notes
- Jeffrey Carpenter and Andrea Robbett, Game Theory and Behavior, MIT Press. Costly signalling and certification, and the result that a claim separates types only in proportion to what it would cost to make falsely, are developed in the chapters on signalling and asymmetric information. Used in section 1, including the practical test that an artefact a competitor could obtain identically from the same source carries no separating information.
- Albert-László Barabási, Network Science, Cambridge University Press. Spreading processes and the dependence of propagation on existing contact structure rather than on broadcast are developed in the chapters on spreading phenomena. Used in section 2 for the claim that a reference is worth what its position in the buyer network is worth, which is a property of the customer rather than of the document. No claim is made about any specific market’s topology.
- Sanjit S. Dhami, The Foundations of Behavioral Economic Analysis, Oxford University Press, and the real-options literature it references. The value of preserving flexibility rises with the irreversibility of a commitment, which is why a permanent rate concession and a decaying asset are badly matched. Used in section 3. No option value is computed; the argument requires only the asymmetry in duration.
- Jeffrey Carpenter and Andrea Robbett, Game Theory and Behavior, MIT Press. Information cascades, in which buyers with weak private signals follow the visible choices of others rather than running their own assessment, are treated in the chapters on social learning. Cited for the mechanism underlying section 2: a named reference works because it lets a prospect substitute someone else’s decision for their own evaluation, which is also why the identity of the customer does more work than the content of the document.
A note on a number this article does not give. There is no standard discount for a case study. The figures quoted in sales training come from markets with different margins and different reference economies. What transfers is the structure: match the duration of what you pay to the duration of what you get.
Joshua Agonya Pi’Rwot, Founder.