Salesforce sells the platform and the transformation. Nobody comes back afterwards and settles up on what either one earned. That gap is not a weakness in the platform — it exists around every enterprise platform ever sold. It is simply unoccupied.
ValueAnalytics sold a capability. Analytics across nine business areas — genuinely useful, genuinely hard to build, and impossible to attach to a moment. Nobody wakes up needing analytics. There is no date on the calendar when a capability becomes urgent, so the offering competed on interest rather than on necessity, and interest does not have a budget line.
ROI Replay sells to a moment. The renewal. The annual true-up. The AI business case coming up for its first review. The conversation that starts why are we still paying for this. Every one of those has a date, a person who owns it, and money already allocated to the decision.
A capability has to create its own urgency. A moment arrives on its own and brings a budget with it. The nine business areas were never the problem — they were the answer to a question nobody had scheduled.
Every Salesforce estate was implemented on a business case. Somebody built a model, put a number on it, and used that number to get the programme approved.
Then the deal was signed, and that document was never opened again. It is the only financial document in a large organisation produced to justify a decision and then permanently excused from ever being checked. Procurement knows this. It is exactly why the benefits half of any subsequent case gets discounted to zero before the conversation starts — not because the benefits were implausible, but because everybody in the room has learned that nobody ever comes back and settles up.
There is a second reason, and it is the one partners feel rather than name. Mosby and Weissman called it the Paradox of Excellence in 2005 — perform consistently well and customers stop noticing, until the relationship is repriced as a commodity. Their remedy, the Continuous Visibility Wheel, runs: discover the expectations, define your distinguishing value, select the metrics to render visible, uncover the best source of data to use, deliver the information. Twenty years on, implementations still stall at the fourth step — because consolidation work erases its own evidence. Retire a legacy CRM licence, collapse an integration, release the contractor who maintained it — and the cost stops appearing in the ledger. It is not recorded as a saving. It is recorded as an absence.
The same event, not two. You demonstrate worth by removing a cost, and a removed cost leaves no line to point at. Within a budget cycle the baseline has absorbed it, and the next funding round starts from the improved number as though it had always been true.
The wheel was never wrong. It was missing an instrument. Steps one to three are strategy and can be done in a room; step four is forensics and cannot.
ROI Replay re-runs the original business case against what actually happened — claimed versus delivered, line by line, against a baseline frozen from the customer's own pre-implementation records. Renew Report carries the earned figure forward into what the next term is worth.
The Paradox of Excellence is not ours. The second one is, and it is the reason the first is solvable at all.
A measurement the vendor controls is worth exactly as little as that control. A partner's own ROI figure is discounted on sight, and the discount is rational: everyone in the room knows the method was chosen after the answer was known. So the measurement runs regardless of sign. A Replay that comes back under the claim is published as one, in the same format, on the same four tests. That is not a disclaimer attached to the product. It is the product.
Which raises the commercial objection immediately: why would a partner commission an instrument capable of convicting them?
Because value does not fail at renewal — it fails quietly in month seven. ROI DriftWatch is the same measurement running continuously against the design intent the contract carries, so divergence surfaces while it is still divergence and not yet outcome. A gap found in month seven costs a conversation. The same gap found at renewal costs the renewal.
Reporting regardless of sign without watching continuously is reckless. Watching continuously without committing to report regardless of sign is theatre. Together they are the only arrangement in which a partner can afford to be audited — and the only one in which being audited is worth anything.
Gartner expects more than 40% of agentic AI projects to be cancelled by 2027. The interesting question is not whether that number holds. It is what a cancellation actually is.
A cancellation is a funding decision made in the absence of evidence. Nobody cancels a programme that can show what the last tranche earned. They cancel the one where the honest answer to did it pay is a shrug and a slide about adoption. The first agentic wave is now old enough to be asked that question, and almost none of it can answer.
A Replay on the prior wave is the precondition for funding the next one. That makes this offering countercyclical: it sells hardest exactly when AI budgets tighten, because tightening is what forces the question.
Four tests. Every line in a Replay is scored before it is allowed into the document, and the grade is printed on the line, so the customer's finance function can audit the case without asking anyone.
The first Replay on any account is reconstruction work — evidence has to be recovered from change logs, invoices, contracts and people's memories before anything can be scored. That is real, bounded, non-compressible effort, and it does not parallelise well.
So this books out. A limited number of Replays can be delivered before year end, and the honest way to launch it is to say so on the page rather than discover it at the third signature. Scarcity that is true is worth more than scarcity that is staged — and a waitlist with a date on it is a stronger commercial position than an open calendar.
The second Replay on the same account costs a fraction of the first, because the baseline is already frozen and the measures are already instrumented. That is the subscription, and it should be bought because the first one proved it worth buying.