No Signatures, No POV

Here is a deal autopsy I have performed more times than I want to count. Six weeks of proof-of-value work. Two solutions engineers, mostly full-time. The customer's data, extracted, cleaned, and modeled by your team because the customer's engineers were busy. A final readout deck with genuinely strong results. And then the sentence that kills the quarter: "This is impressive. We'd like to extend the evaluation to a second use case before we take it upstairs."

The seller calls it momentum. It is not momentum. It is the invoice arriving for a decision the team never made: they started proving before anyone agreed what proof would mean.

A proof of value without a signed success-criteria agreement is free consulting. You are donating your scarcest resource — SE hours — to educate a buyer who has committed to nothing, against criteria that don't exist, judged by people you haven't named, with a finish line that moves every time you approach it.

The fix is one page and one rule: no signatures, no POV.

Why POVs drift

An unscoped POV fails in a predictable sequence, and every stage of the failure feels like progress while it's happening.

First, the scope creeps. Without written criteria, "success" is whatever the last stakeholder in the room wants to see next. Week two adds a data source. Week four adds an integration test nobody mentioned at kickoff. Each addition is small, reasonable, and fatal in aggregate.

Second, the judges multiply. When no one has agreed who evaluates the results, everyone does. The skeptic who never attended a session shows up at the readout with a new objection, and because nothing was signed, the objection is admissible.

Third — and this is the one that costs you the quarter — even a flawless technical result produces no decision. The team proves the product works, and the customer agrees the product works, and nothing happens, because "it works" was never connected to "and therefore we will buy." The POV answered a question nobody with budget authority had asked.

Every one of those failure modes is a criteria-design failure. None of them is an execution failure. Your SEs did not lose that deal in week six. The AE lost it at kickoff, by starting without a contract for what the evaluation would prove and what would happen when it did.

The agreement: what goes on the page

The instrument is the POV Success Criteria Agreement — Template 5 in the Field Toolkit — and it fits on one page. That is deliberate. A twenty-page evaluation plan is a document nobody reads; a one-page agreement is a document everybody signs.

Three to five criteria. No more. Every criterion is a promise of SE hours and customer attention. A list of twelve criteria is not rigor — it is an unscoped POV wearing a spreadsheet. If a stakeholder insists on criterion number nine, something else comes off the list, and the negotiation about what stays is itself discovery: it tells you what this buying committee actually cares about.

Each criterion carries four fields, and an empty field is a finding, not a formality:

The criterion, specific and measurable. Not "the model performs well." Something like: "Alert triage precision of at least 85% at a fixed 40% workload-reduction threshold." If two reasonable people could disagree about whether it was met, it is not a criterion yet — it is a hope.
The target. The number, with units, agreed in advance. Targets negotiated after results are in are not targets; they are arguments.
The data set and method. Which data, from which period, evaluated how. This field is where readout disputes go to die. "Your results were on a sample" is unanswerable in week six and trivially preventable in week zero.
A customer owner. A named person on their side who owns this criterion — supplies the data, attends the sessions, attests to the result. A criterion without a customer owner is a criterion the customer doesn't believe in, and it will be the one they dispute later.

At least one business criterion. This is the field most technical teams skip and the one that decides whether the POV moves money. Technical wins alone don't move money — they move technical opinions. If every criterion is precision, latency, and uptime, then the best possible outcome of your POV is a technical recommendation, which then goes into a business-case process you have no hand in. Put the business on the page: "Evaluation results, extrapolated across the full alert volume, demonstrate at least $2M in annual analyst-capacity recovery, using the customer's own workforce numbers." Now the readout is a budget conversation, because you designed it to be one.

Four signatures, before kickoff

The agreement is signed by four people, and each signature buys you something specific:

The champion — the signature that says "I will spend internal capital defending these criteria when someone tries to move them." This is also a champion test, and a good one. A champion who won't sign a document they helped write is telling you something you need to know now, not in week six.
The EB sponsor — the signature that connects proof to money. It sits directly above the conditional commitment (more on that in a moment). Without it, you are proving things to people who cannot buy things.
The technical gatekeeper — the signature that disarms the readout ambush. The skeptic who signed the criteria in week zero cannot introduce new ones in week six without visibly moving the goalposts in front of their own leadership. You are not silencing the skeptic; you are converting them from a judge with unlimited jurisdiction into a judge bound by the law they wrote.
Your SE lead — the signature your own organization owes itself. It says the criteria are achievable, the timeline is real, and engineering is committing hours to a scoped effort, not an open-ended one. Any SE leader who has burned a quarter on a drifting POV will sign this one with feeling.

And the sequence is not negotiable: signatures before kickoff. A criteria agreement signed in week three governs nothing — the scope has already drifted, and you are asking people to sign a description of the drift.

The sentence that turns proof into pipeline

The most important lines on the page are not the criteria. They are the conditional commitment beneath them, in the Economic Buyer's words, in writing:

"If the evaluation meets the criteria above, we will issue the purchase order by the end of Q3."

Getting that sentence is the real test, because asking for it forces the conversation demo-first sellers spend entire quarters avoiding. If the EB won't complete it, the deal is not ready for a POV — and that is the single most valuable thing you can learn before spending six weeks of SE time. The business case isn't built yet. Go build it. The POV was about to fail slowly and expensively; the unsigned sentence lets it fail in one meeting, for free.

Sellers resist asking because it feels like pressure. It is the opposite. You are offering the buyer the most honest deal in enterprise software: we will prove exactly this, measured exactly this way, judged by your own people — and if we do, you act. Serious buyers respect it. Unserious buyers reveal themselves. Both outcomes are wins.

The objection you're already forming

"Our buyers won't sign anything before an evaluation." Some won't — and notice what that refusal is. A customer unwilling to define success is telling you either that they don't know what success looks like (discovery isn't done; go back) or that this evaluation exists to satisfy a process, check a box, or pressure an incumbent's renewal price (you are the free consultant; leave). The signature request isn't bureaucracy. It is the cheapest disqualification instrument you will ever deploy — one page, one meeting, and it finds the deals that were never deals before they consume your best engineers.

"This will slow the deal down." It slows it down by roughly one week at the start and speeds it up by a quarter at the end, because the readout stops being the opening argument of a negotiation and becomes the closing evidence of one already agreed.

Where to start Monday

Audit every active POV for a signed criteria agreement. Not "we discussed criteria" — a document with four signatures and a conditional commitment. Any POV running without one gets paused until the page exists. Yes, paused. The discomfort of that conversation is smaller than the write-off you are currently accruing.
Check every criteria list for a business criterion and a customer owner per row. All-technical criteria means the readout will produce a technical opinion, not a purchase order. Fix the list now, while it is still cheap to fix.
Make the conditional commitment a gate, not a preference. No EB sentence — "if the criteria are met, we will ____ by ____" — no SE hours allocated. Put it in the stage exit criteria and let the gate do the disqualifying, so your sellers don't have to summon the courage deal by deal.

A proof of value is the most expensive artifact your sales organization produces. Spend it like it costs what it costs: on buyers who have signed their name to what proof means and what happens when they see it.

No signatures, no POV. Everything else is consulting — and you forgot to send the invoice.

Go deeper. The POV Success Criteria Agreement is Template 5 in the free Field Toolkit, companion to The Value Engine: How Elite Enterprise Sales Teams Turn Buyer Pain into Forecastable Revenue by Rudy M. Celekli — where the full evaluation stage is demonstrated end-to-end on one $8.9M enterprise deal. Get the book, download the template, and put it in front of your next POV before kickoff.