The forecast built on sand
The conversation repeats itself at the close of every quarter across revenue teams of all kinds. The forecast showed a 90% probability of hitting the target. The actual result was 65%. The gap can’t be explained by the market, the product, or the team.
It’s explained by what the forecast was made of: what each rep believed was going to happen, not what the buyer had actually signaled they were going to do.
This is not a methodology problem. It won’t be solved by changing your weighting model or layering AI on top of it. It’s a data origin problem: if what goes in is opinions, what comes out is a projection of opinions.
The problem isn’t the model — it’s the source
Forecast models have improved considerably. There’s automated scoring, stage-weighted pipeline, systems that cross-reference historical closed deals against the characteristics of the current pipeline. None of those advances solve the underlying problem if the data source stays the same.
Reps update the CRM with what they perceive. “I feel good about this account.” “They said it was a priority for this quarter.” “There’s a great relationship with the champion.” When the manager consolidates, they’re summing up those perceptions. When the VP reports to the board, that sum becomes a number that looks solid.
A sophisticated forecast model applied to opinion data is still an opinion forecast. The model’s precision doesn’t compensate for the weakness of its source.
Three signs your forecast is an act of faith
Before changing the process, it’s worth identifying whether the problem actually exists. These three signals appear consistently in forecasts that end up missing:
The percentages reflect how the seller feels about the deal, not observable buyer behavior. 75% might mean “the call went well” or “I’ve been working this account for three months.” Neither is a criterion.
There are no notes documenting what the buyer said about timing, urgency, decision process, or budget. Only the rep’s interpretation of that conversation exists. If the rep leaves, the information leaves with them.
The deals at the top of the forecast show no recent buyer activity. They’re still there because no one moved them down, not because the buyer is advancing. Age in the pipeline gets mistaken for momentum.
None of these three signals are invisible. All of them are detectable in a pipeline review if you know the right criteria to look for.
Buyer evidence vs. rep belief
The central distinction in this framework isn’t complex, but it does require discipline to apply consistently: every input to the forecast is classified according to its origin.
- —“They told me it was a priority for this quarter”
- —“I feel like they’re ready to decide”
- —“There’s great chemistry with the champion”
- —“The company is a perfect fit for our ICP”
- —They said: “we need this live before year-end, legal requires it”
- —The CFO joined the second demo without being invited
- —They asked about the onboarding process and implementation timelines
- —They sent their IT team to validate the integration with their current stack
The difference between the two columns isn’t about size: it’s about traceability. Evidence can be traced back to something the buyer said or did. Belief cannot. When you ask for the source of a belief-based data point, the answer always starts the same way: “the rep said that…”
What changes when you anchor your forecast to what buyers say
When the forecast is anchored to the actual language and behavior of the buyer, three things change in a notable way.
Reps start having different conversations. If advancing a stage requires buyer evidence, sellers start actively looking for that evidence during calls — instead of moving a deal forward because “the conversation went well.” The questions change: they stop being general and become specific about timing, decision process, and the consequences of not solving the problem.
Managers can distinguish between deals with real momentum and deals that simply have seniority in the pipeline. It’s not the same to have a deal in the “Proposal” stage because the buyer asked when implementation could start, versus one at the same stage that has had no buyer activity for six weeks.
The forecast stops being a weekend ritual and starts being an actual management tool. You don’t need more information — you need the right information: what the buyer said, when they said it, and what behavior backs it up.
A recalibration for this week
There’s a practical exercise that works well as a starting point. Take the ten highest-value deals in your current forecast. For each one, answer these three questions without reading the rep’s notes — only with what you can verify:
- What did the buyer say, specifically, about when they want to solve this problem?
- What action did the buyer take in the last two weeks that demonstrates real progress?
- Do you know, by name, who is going to sign this deal?
If the answers for most of them depend on what the rep interpreted rather than something documented that the buyer said or did, that deal isn’t qualified as “forecast”: it’s qualified as hope. The number you’re assigning a high probability to isn’t a data point — it’s a wish.
The worksheet that accompanies this article turns those three questions into a table to run with your team this week, deal by deal, marking in real time where the evidence is and where the assumption is.
Why this shouldn’t be done manually
Everything above works. The problem is that it depends on someone remembering to run the audit, deal by deal, week after week. In teams of ten reps or more, that process doesn’t scale — and when it doesn’t scale, only a fraction of the pipeline ever gets audited: the deals someone decided to review that week. The rest go through unchallenged.
That’s exactly what Performy solves automatically. Every call is analyzed for real buyer signals: mentions of timeline, urgency indicators, references to the decision process, questions about implementation. When the buyer says “we need this before year-end” or “we have to go through legal first,” that gets captured as evidence — not as a note the rep decided to write, or not write, after hanging up.
The result is a forecast that reflects what the buyer is communicating in every conversation, across 100% of your team’s calls — not just the ones someone decided to look at this week.