This series is brought to you in partnership with Airspeed. They build AI that reads what customers actually say and writes it back into your CRM. That's the subject of the next article. This one is about everything you can do before you spend a dollar on it, using the data you already have.
The Forecasting Series:
There was a deal that sat in commit for three consecutive quarters.
Every forecast call, there it was. Committed.
Nobody questioned it, because on paper it looked perfect: the requirements call had lined up exactly with what our product did, feature for feature. It was the deal that was obviously going to close, right up until someone finally did a deep dive and discovered we hadn't heard from the buyer in six months.
We weren't lying to ourselves on purpose. We were holding onto hope, because the fit was so clean it felt inevitable. So when they went dark, instead of closing the opportunity out and starting fresh if they ever came back, we just kept it open. And kept forecasting it.
Here's the part that actually hurt. When that buyer did re-engage, we weren't looking at a warm nine-month-old opportunity. We were looking at a deal that needed to start over at day one, because the rep had to requalify everything from scratch. The buying committee had almost completely turned over. The people who loved us were gone. That opportunity should have been closed and reopened as net new, and instead it spent three quarters propping up a forecast it had no business being in.
That deal did not lie to us. But the board of investors was not impressed when they found out we had been relying on a long shot for several quarters. It looked like we did not know how to manage a forecast.
This article is about building the friction that catches a pipeline problem before it costs somebody their quarter.
Let's be fair to the reps for a second, because our instinct is to blame them and that instinct is wrong.
Reps are optimists who get paid to be optimists. They have to be able to withstand hearing “NO” all day long and still have the resilience to keep trying.
They are also managed by people whose own careers were built on that same optimism, and everyone reports up to an exec who has to say a number out loud to a board. Optimism compounds on the way up, because at every rung, nobody wants to be the one who admits the goal cannot be hit with the pipeline currently in the system.
You've seen what happens when a forecasted number drops.
The disappointment is palpable, and it rolls through the entire organization, from the board and CEO all the way down. Nobody wants to deliver that news. And here's the kicker: your salespeople have pulled a Hail Mary out of the hat before. They closed the uncloseable deal in the last week of Q2 that one time. So maybe they'll do it again?
The whole system was designed to keep salespeople in a pressure cooker, and that's on purpose. Goals have to be enough of a stretch that a seller feels genuinely motivated to do the work. Set the goal too low and you get a rep running a side hustle, taking their weekly pipeline review beachside while peeling off a wetsuit. (Yes, that happened. No, I won't say more.)
Seller optimism is a feature, not a bug. You are not going to train it out of people, and you shouldn't want to. What you can do is put two things in place to catch it early: a manager savvy enough to run a genuinely thorough pipeline review, and system signals that automate the work of combing through notes and next steps to find the deals that don't add up.
The rest of this article is how to identify and trigger based on those signals.
None of these require a budget, a new vendor, or a procurement cycle. Every one of them is buildable in your CRM right now, out of data you're already collecting. Here are the ones worth the effort:
Any one of these in isolation can be innocent. The power is in combining them, which is exactly what the slip report does.
The individual indicators are the ingredients. The slip report is the dish.
Start with what changed in the quarter. Did the deal move into this quarter from a prior one? Was it created in-quarter? Did it push out to a new quarter entirely? Then layer in stage movement: did the stage go up, down, or stall while the close date moved?
Now combine the signals, because that's where the truth lives. A close date pulled in and a stage bumped up, with no activity logged and no clear next step? That's not progress. That's wishful thinking. A deal sitting in commit whose meetings never had more than one person from the buyer's side on the invite? Also wishful thinking.
The exercise is to define, up front, which signals all have to be present for a deal to count as healthy. Activity within X days, more than one contact engaged, a scheduled next step, a stage that matches the age of the deal. When a deal fails the combination, it surfaces on the report, and someone has to explain it on the Monday pipeline review call. If you missed the piece on building the meeting cadence a trustworthy forecast needs, it's here.
This combined, quantitative view will almost always be more useful than a score a rep or manager typed in by hand. That said, some managers swear by their qualitative gut-score, and I'm not going to pretend that judgment is worthless. My recommendation: if you're going to roll out a qualitative score, pair it with an objective quantitative one that looks only at the signals. When the two disagree, that gap is the most interesting conversation you'll have all week.
Everything above catches problems in the current quarter. This catches the people who cause them, quarter after quarter.
Track forecast accuracy by rep, by manager, and by SVP, across trailing quarters. Not the deal detail. The accuracy of the call itself: what did they commit, what actually closed, and how wide was the gap.
The way to build this is with regular snapshots. Take a picture of the forecast at week one, week two, week five, and week ten, and compare each against where the quarter actually landed. Do that consistently and a pattern emerges: you learn how each rep's, each manager's, and each SVP's number drifts across the quarter. Some are dead-on by week two. Some are fantasists until week nine and then panic. The snapshot history tells you which is which.
This is the artifact that ends arguments. When someone wants to know whether to trust a director's commit, you don't offer a philosophy. You show that this director has landed within five percent for six straight quarters, or that this one runs twenty percent hot every single time, without fail. History turns "I have a feeling about this number" into "the record says discount it by fifteen percent."
Publishing this without creating conflict or enabling finger pointing across managers is its own skill. Don't drop a rep-shaming leaderboard into a Slack channel. Roll it up to managers first, frame it as calibration rather than judgment, and let the people with hot forecasts see their own numbers privately before anyone else does. The goal is a more accurate forecast, not a public shaming.
What you're really giving leadership is the chance to hear the bad news quietly from RevOps in week three, instead of discovering it on the earnings call in week thirteen. That's a huge win for your executive team.
I'd rather you hear this from me than discover it the hard way. (Been there. Done that.)
Every single indicator in this article is a proxy. Close date pushes, activity gaps, single-threaded deals, stalled stages: none of them are the actual problem. They are the shadow the problem casts on your CRM.
The real information lives in what the customer said on the call, wrote in the email, and pointedly did not say when your rep asked about timeline.
The indicators only catch the deals where the customer's disengagement finally showed up as a missing next step or a stale activity date. By then you're often reacting to a fire that started weeks ago.
The three-quarter commit deal from the top of this article is the perfect example. Every quantitative signal eventually flagged it: no activity, single-threaded, close date pushed. But those signals lit up months after the actual moment of death.
Without conversational AI analyzing your call data, and some organizational effort to train an LLM on what your deals actually sound like, these quantitative scores miss serious context.
Here's the good news: once you layer that conversational context in, you can retire the old qualitative gut-score entirely and lean on something far richer. Even better, most of the raw material is already sitting in tools you've integrated with your CRM. Call and meeting recordings, synced email, attached presentations. That is an enormous amount of context, and until recently there was no practical way to read it at scale. Our new AI friends, like Airspeed, changed that.
Every indicator in this article is a shadow of what the customer actually told your rep. Article four goes to the source, and shows how to get more out of the tools you already own.
Your turn. What's the deal you held in commit way too long, and what finally made you close it out? Post the horror story in the community. We've all got at least one.
