Buyers Judge Vendor Claims by Size and Ignore the Track Record

← Back · ACT II · Chapter 11: Outcome Credibility Markets

A corporation that promises to repay one billion dollars in ten years has made a claim. Financial markets do not accept that claim at face value. They price it. The same promise from two different issuers trades at different yields, because the market has built an entire apparatus for answering a separate question from how large the promise is: how likely is it to be kept.

That apparatus took decades to build and it changed the behavior of everyone inside it. Issuers with strong records borrow cheaply. Issuers with weak records pay for their history. The record follows the promise.

Enterprise buying has no equivalent. A vendor claiming a 40 percent improvement and a vendor claiming a 25 percent improvement are compared on the size of the claim. Nothing in the standard comparison asks how often each one has been right before.

Magnitude Is Visible and Credibility Is Not

Buyer intelligence is not the constraint here. Available information is, and the shortage produces a predictable distortion.

Expected value from any claim is the magnitude of the outcome multiplied by the credibility of the claim. Buyers can observe the first term precisely, because it is printed on the slide. The second term is not observable at all in the ordinary course, so in practice it gets treated as though it were the same across vendors. When credibility is assumed constant, comparison reduces to magnitude, and the largest claim wins.

Work the arithmetic against realistic numbers. A vendor claiming 40 percent with a historical realization rate of 35 percent carries an expected value of 14. A vendor claiming 30 percent with a realization rate of 80 percent carries an expected value of 24. The smaller claim is worth substantially more. Under a comparison that cannot see the second column, it loses.

That is the shape of the problem. Buyers are not choosing badly. They are choosing correctly against an incomplete price.

The Cost Shows Up After Signature

The evidence of the gap is in the realization data rather than the selection data, because the distortion does not become visible until the promise comes due.

World Commerce and Contracting finds average contract value erosion of 8.6 percent, and more recent work from the same body with the contract platform provider Ironclad puts the loss at 11 percent of contract value once agreements move into delivery and ongoing management, with complex industries frequently higher. Gartner’s 2023 survey of technology buyers involved in renewal and expansion decisions found that 60 percent regret nearly every purchase they make, which Gartner reports as a 6 percent increase on 2020.

Those are the numbers of a market that prices magnitude and settles for whatever credibility turns out to have been. A market that priced credibility in advance would not produce a sixty percent regret rate among the buyers who renew and expand these agreements.

The Party That Controls Measurement Controls Value

There is a reason the missing rating has never appeared on its own, and it is worth stating directly, because it explains why waiting for one is not a strategy.

Credit ratings exist because a third party with no stake in the issuance produces them and the market compels their publication. In enterprise buying, the party best positioned to measure realized outcomes is the party whose claim is being measured. Vendors hold the case studies, the benchmarks, the reference customers, and the definition of what counted as success. A market cannot price credibility using evidence supplied by the promiser.

The buyer, however, holds something the vendor does not: the actual outcome. The buyer knows what the deployment delivered against what it projected. Every organization is already sitting on the raw material for the rating that does not exist, unrecorded, distributed across sponsors who moved on and business cases nobody reopened.

Constructing the Missing Column

A buyer does not need the market to build this. A single organization can build its own, and the version that works is smaller than it sounds.

It requires that every claim entering evaluation be recorded in a form specific enough to be checked later, which mostly means a number, a date, and a definition. It requires that the check actually happen on the date, with a named owner. And it requires that the result attach to the vendor and stay attached, so that the next evaluation opens with a realization record rather than a blank second column.

Do that for three years and the comparison changes character. Claims stop being judged on ambition and start being weighted by history. A vendor with a strong record can promise less and win, which is the incentive change that matters, because it makes accuracy commercially rational for the first time.

Enterprise buying will keep rewarding the largest claim until someone starts keeping count of which claims came true. The buyer is the only party with the evidence to do it, and the only party with a reason to.

This argument runs alongside AI’s Second Wave Is Being Funded by Savings Nobody Checked, which follows the same unverified claim into the next funding round.

References: World Commerce and Contracting, contract value erosion research, 8.6 percent average erosion, worldcc.com; World Commerce and Contracting with Ironclad, 11 percent of contract value lost in delivery and ongoing management; Gartner, “Gartner Survey Reveals 60% of Technology Buyers Involved in Renewal Decisions Regret Nearly Every Purchase They Make,” June 14, 2023, gartner.com. Illustrative expected-value figures are worked examples, not measured data.

Keep reading: ACT II · Chapter 11: Outcome Credibility Markets