Software Now Reports Its Own ROI, and the Vendor Keeps the Record

← Back · ACT II · Chapter 14: Who Controls Value Engagement?

PYMNTS ran a piece at the end of August on where B2B software pricing is heading, and one sentence in it names a problem I have watched from both sides of the table. “Companies can now deliver measurable value to customers without making that value particularly measurable by customers.”

The argument behind it is straightforward. Platforms that sit inside financial workflows, payments, accounts payable, receivables, procurement, treasury, already hold the transaction records that would show what the software did to costs, cash and working capital. So they can stop reporting how many invoices were processed and start reporting what the processing was worth. The article describes this as a move from modeled ROI to observed ROI, and says proof of value may become a core product feature. The three executives quoted in support of it work at payments companies, which is worth saying plainly: this is the supply side describing where it intends to go, and not evidence about what buyers are receiving today.

The direction is credible anyway, because it is a sensible thing for a vendor to build and the data is already sitting in the product. What deserves more attention is what it does to the buyer.

Where the Number Used to Come From

Until now the return on a software purchase was estimated in advance, by the seller, inside the business case that justified the deal. I have written about how that document gets built and who it actually reaches. The estimate had one useful property: everyone in the room knew it was an estimate. It was a projection, so it got argued with, discounted, and treated as an opening position rather than a finding.

An observed number behaves differently. It arrives with transaction records behind it, it updates continuously, and its arithmetic is not in question. It reads as a measurement. It remains the seller’s measurement, because the seller chose which transactions to count, what baseline to compare them against, and which of the buyer’s costs sit inside the calculation and which sit outside it. None of those choices is visible in the figure that comes out.

The Renewal Meeting Has One Instrument In It

Renewals are where this lands. A buyer deciding whether to keep paying needs an answer to a narrow question: did the result we bought this for arrive, and how much of it. If the only continuous record of that result is produced by the supplier being paid, the buyer is not evaluating the supplier. It is reviewing the supplier’s account of itself, with its own finance team checking someone else’s arithmetic instead of producing a number of its own.

That position gets weaker as the supplier’s reporting gets better. A crude vendor claim invites scrutiny. A precise, well sourced, continuously updated one discourages it, because assembling an independent version looks like duplicated work and the meeting has forty minutes.

The Same Asymmetry, One Step Later

This is an old problem extending its reach rather than a new one. The seller’s value model arriving machine readable is the pre-purchase half, where the supplier brings structure to the table and the buyer brings a reaction to it. A right to audit a supplier’s records is the contractual half, and it does not extend to establishing whether the promised result occurred. The record of what actually happened after the money moved is the third position, and until recently it was empty on both sides. What is changing is that the seller is filling it.

Once it is filled, the buyer’s options narrow. A record you did not define and cannot reproduce is not something you can negotiate against. You can accept it or dispute it, and disputing a number without an alternative number is how buyers concede the point while believing they pushed back.

What the Buyer Writes Down First

The answer is not to refuse the supplier’s reporting, which is genuinely useful and will often be more accurate than anything a buyer assembles by hand. The answer is that the buyer’s own record has to exist first, and it has to be defined before the purchase, while the buyer still has leverage and the definition is cheap. The Value Standard puts that declaration before the money moves: which outcome this purchase is being made to produce, what the baseline for it is on the day of signing, what evidence counts as proof that it moved, and who holds that evidence. A return requires a denominator, and the denominator is the part a supplier has the least reason to define strictly and the buyer has the most reason to define precisely.

With that written down, an observed supplier number is one input into a decision the buyer is equipped to make, and without it that number is the only account of the outcome anyone can cite, so the decision follows it by default.

There is a version of this that ends well for buyers. A supplier computing real outcomes from real transactions is an improvement on a supplier estimating them in a spreadsheet, and buyers should want the better instrument in the room. It only helps them if they arrive at the renewal with a record of their own to set beside it, written before the money moved and kept somewhere the supplier does not control.

The reason that dashboard has no rival inside the buyer is that the other half of the ledger is unowned.

The same asymmetry, applied to service contracts priced per resolved case, is the subject of Outcome-Based Pricing Works Only if the Buyer Does the Measuring.

Battery buyers have written a different rule into a published model contract. Its clause 6.6.6 says that performance, availability and degradation figures reported by the seller have no evidential priority over figures the buyer derives from the same data. The clause, and how it carries over to software and AI purchases, is the subject of Battery Buyers’ Contract Gives Seller Claims No Extra Weight.

Reference: PYMNTS, “B2B Pricing Power Changes When Software Can Prove Its Own ROI,” 27 August 2026, at pymnts.com/news/b2b-payments/2026/b2b-pricing-power-changes-when-software-can-prove-its-own-roi/. Quoted phrases are taken from the article’s highlights and body, read 6 September 2026. The article reports a direction in the B2B software market and quotes payments industry executives; it is not evidence about buyer practice, and nothing beyond what it states is credited to it.

Keep reading: ACT II · Chapter 14: Who Controls Value Engagement?