How to Run a Deal Review That Actually Predicts Whether You'll Close

Most deal reviews report status. That's why they don't predict anything.

Ask a rep how a deal is going and you'll get a status update: two calls this week, deck sent, "feeling good about it." That's not a deal review. A deal review worth an hour of a manager's time answers one question only: does this deal survive contact with reality, or is it held together by a rep's optimism? But most deal reviews never ask that. They ask "where are we," get an answer built from memory and hope, nod, and move to the next deal on the list.

The fix isn't a better cadence or a longer question list. It's treating every claim in the review as something to verify, not something to accept. So a deal review process built that way predicts outcomes. One built to fill a slot on a manager's calendar doesn't.

Deal review vs. forecast call: two different jobs, constantly confused

Most sales orgs run one meeting and call it both things, and that's the root problem. A forecast call asks "will this close, and when." It's a roll-up exercise: leadership needs a number, reps supply confidence, everyone reconciles to a total. But a deal review asks something different: what do we actually know about this deal, and what are we assuming? One is about the number. The other is about the evidence underneath it.

Collapse them into one meeting and you get a predictable failure mode. Ask "will this close" in front of a VP and a rep will perform confidence instead of surfacing risk. Nobody wants to be the reason the forecast number drops in a room full of people staring at a dashboard. So the champion who's gone quiet becomes "just busy." The economic buyer nobody's met becomes "aligned, just needs to see the contract." The deal review vs forecast call distinction matters because the two meetings reward opposite behavior: one rewards a confident answer, the other rewards an honest gap.

Run them separately.

The forecast call happens with a number already built from verified deal reviews, not the other way around. And if your forecast call is where risk gets discovered, you're finding out too late to do anything about it.

The five things a deal review must actually verify

Every real deal review, whatever framework it borrows from, is trying to confirm five things. But skip one and the review becomes theater.

Business impact and the compelling event

"They have pain" isn't information. Pain without a number and a reason to act now is just a rep's read on a friendly conversation. What's the cost of doing nothing, in dollars or hours or whatever the buyer measures in? And why does that cost matter this quarter and not next year? A compelling event, a renewal date, a budget expiring, a competitor already live, is what turns pain into "why now." No compelling event, no urgency. No urgency, no reason the deal closes on your timeline instead of drifting.

The buying committee, verified not assumed

A rep saying "the champion is on board" and a champion who's actually put something on the line are not the same fact. Champion, coach, and economic buyer are three different roles doing three different jobs in the deal, and conflating them is its own failure mode worth reading up on separately. But what matters here is simpler: has anyone on your side actually spoken with the economic buyer, or is that a rep's assumption relayed secondhand? Has the champion done something that costs them something, forwarded an email internally, booked the next meeting, pushed back on their own team? If the only evidence of buying-committee engagement is a rep's paraphrase of a conversation nobody else heard, that's not verified. So multithreading isn't optional here. A deal with one confirmed contact is a deal with one point of failure, and scoring that risk properly deserves its own checklist.

Decision criteria vs. decision process

This is where most deal reviews stop short, and it's the one worth the most time. Decision criteria is what the buyer is grading vendors on: features, integrations, the security checklist, the reference calls. Decision process is how they'll actually get from "we like it" to a signature: who signs, what legal has to review, what procurement's cycle looks like, and how long each step has historically taken at that company. Every competitor in this field names the distinction. But almost none of them do anything with it.

Here's what doing something with it looks like. Decision process gets treated like a project plan, not a vague timeline. A named owner for each step. A dated milestone for each step. A verification method for each milestone that isn't "they said," but something you can point to: an email confirming legal received the redline, a calendar invite for the security review, a message from procurement about the budget cycle. So a deal with decision criteria met and no dated process behind it isn't close to closing. It's close to stalling.

Competitive reality and red flags

Knowing your competitors isn't the same as knowing whether you're losing to them. The specific signals that mean a deal is slipping: the champion stops initiating contact and only responds when you reach out, meetings get pushed without a reschedule attached, or the buyer starts asking questions that sound like they're building a case for someone else, or worse, for doing nothing at all. No decision, not a competitor, is usually what actually kills an enterprise deal, and a review needs to ask that directly: has the cost of doing nothing gone up?

The next step, owned by the customer

The mutual-action-plan test is blunt and it works: did the customer commit to the next step, or did the rep? "I'll follow up Thursday" is a rep commitment. "They're sending the security questionnaire back by Thursday" is a customer commitment. Only one of those tells you anything about momentum. If every next step in a deal's history was generated and owned by your side, the deal doesn't have buy-in.

It has a rep working hard.

The standardized question set: same 8 questions, every deal, every time

A deal review that asks different questions every time isn't a process. It's a conversation, and conversations drift toward whatever the rep wants to talk about. The same eight deal review questions to ask, asked the same way, on every deal, every review, is what turns a review into something that produces comparable answers across a pipeline instead of eight different vibes.

1. What's the quantified cost of not solving this, and who gave us that number?
2. What's the compelling event that makes this quarter different from next quarter?
3. Who is the economic buyer, and has anyone on our side spoken with them directly?
4. What has the champion done, beyond talking to us, that shows they're spending their own credibility?
5. What's the dated, named-owner process from here to signature, and what's the next verification point on it?
6. What does the buyer's decision criteria actually weigh, and where are we weak on it?
7. What's the specific evidence, not the feeling, that we're winning or losing against the alternative, including doing nothing?
8. What did the customer commit to as the next step, and is there a date attached to it?

Notice what's missing: "how do you feel about this deal." Feelings aren't evidence. So this set isn't invented from nothing, either. It's built by cross-checking Insight Partners' MEDDICC-based template, the Gap Selling-style diagnostic questions Keenan's firm publishes, and the seven practitioner categories SellingSherpa crowdsourced, then keeping only what survived all three and adding the verification language none of them had. Every one of the eight has a verifiable answer or it doesn't, and "I think so" isn't an answer. It's a placeholder for one.

Turning answers into a signal: a simple deal-health scoring rubric

Every question above produces an answer, but eight answers per deal don't tell a manager anything on their own. The mechanism that closes the gap: score each of five verification areas, Impact, People, Decision, Competition, Next Step, as Verified, Claimed, or Unknown, and be honest about what each status actually requires.

The rule that makes this more than a scoring exercise: any deal carrying two or more Unknowns comes out of the current-quarter forecast, full stop. Not deprioritized, not flagged for a follow-up call. Out. That's an uncomfortable rule the first time a manager applies it to a deal a rep swears is closing.

Whether you build this rubric into a shared doc or wire it into whatever CRM fields your team already touches is a build-vs-buy question worth its own answer. Either way, the rating has to survive someone disagreeing with it, or it isn't a system. It's a vibe with a table around it.

Cadence: who reviews what, and how often

Cadence should track deal segment, not habit. A flat "we review every deal weekly" rule wastes senior time on deals that don't need it and starves the ones that do.

Manager 1:1s cover every active deal in a rep's pipeline weekly. That's cheap, and it catches drift in the five verification areas before it compounds into a surprise. Team or executive-level reviews run biweekly to monthly, depending on deal count and average contract value, and they should spend most of their time on deals whose score changed since the last review, not just the ones due for a scheduled check-in.

A five-deal enterprise pipeline with a high ACV probably earns exec eyes every two weeks; there aren't many deals, and each one is worth the attention. But a fifty-deal SMB pipeline doesn't need executive review on each one. It needs the scoring rubric doing the sorting so a human only looks at what the numbers already flagged.

Why deal reviews rot between sessions, and what to do about it

Here's what actually happens between deal reviews. A rep updates "last contact with economic buyer" from memory, because that field lives in a CRM dropdown and updating it means stopping to think back through the week. By the next review, that memory has drifted: a meeting that was actually with the champion gets remembered as a meeting with the economic buyer. Not from dishonesty. Just from how memory works under a full pipeline. So the scoring rubric above is only as good as the inputs feeding it, and manually re-entered inputs decay a little more every cycle.

The fix isn't asking reps to be more careful. It's not making the data come from memory in the first place. Last contact with the economic buyer is a fact that already exists in an email thread and a calendar. Last decision-process update is a fact that exists in whatever thread confirmed it. Last competitive signal, if it's real, showed up on a call, and call activity increasingly lives as a transcript, not a memory, sitting in whatever tool recorded it, Fireflies or Fathom or the CRM's own call log. None of that needs a rep to reconstruct it from memory on review day. So it needs a system that pulls the fact from where it already lives.

That's the same argument that runs through most of GTM engineering: the manual re-entry step is where systems rot, and the fix is rarely "try harder." So it's removing the re-entry step entirely. It's also where the automation conversation tends to get overclaimed. An AI agent reading call transcripts isn't replacing the judgment call of whether a deal is Verified or Claimed. But it's making sure the person making that call is looking at what actually happened instead of what a rep remembers happening.