Sofía Herrera used to open her Friday forecast calls the same way every week: "Okay, be honest with me." She meant it as a joke, mostly, the way a sales leader says it to break the tension before eight reps take turns explaining a number that everyone in the room already suspects is fiction. But she wasn't really joking, and everyone on the call knew it. The forecast her team produced every Friday was less a prediction than a negotiation — a document assembled from eight people quietly deciding, independently, how much truth was safe to put in front of their manager that week.
Two kinds of fiction lived in that number, and Sofía had learned to spot both. The first was sandbagging: a rep with a deal at 90% closed probability calling it 60%, because coming in over quota looked better than coming in exactly on it, and because a lowballed forecast this month bought slack for a rockier month next quarter. The second was the opposite problem and, in its way, more dangerous — happy ears. A rep excited about a deal, genuinely believing in it, reporting a client's polite interest as a near-certain close because the story felt good to tell and nobody wants to be the one dragging the mood of a Friday call down with realism.
"Every forecast I ever inherited was a negotiation dressed up as a spreadsheet. Reps weren't lying to me. They were protecting themselves from a number that had real consequences and no actual data behind it."— Sofía Herrera, Sales Leader
The Friday Call Everyone Dreaded
Sofía's team sold enterprise software subscriptions to mid-market manufacturers across central Mexico, a sales cycle long enough that a single quarter's forecast rested on maybe fifteen live deals per rep, each one a judgment call about a human relationship that was, by definition, impossible to reduce cleanly to a percentage. The Friday call existed to true that number up before it went to her VP, and it usually took ninety minutes of gentle interrogation — "walk me through why this is 70% and not 50%" — to get anywhere close to a defensible figure. Reps hated the call. Sofía didn't love it either. Everyone left knowing the number was still, at best, an educated compromise between eight different incentives to shade it one way or another.
What made it worse was that the compromise wasn't even stable. A sandbagger having a good quarter might suddenly stop sandbagging to build a cushion of goodwill before a slow one. A happy-ears rep who'd been burned by an over-promise the previous quarter might swing hard the other way and start under-calling everything. The forecast wasn't just inaccurate — it was inaccurate in a different direction every month, which made it nearly useless for the one thing a forecast is actually supposed to do: tell leadership, with some confidence, what was about to happen.
Two Kinds of Fiction: Sandbagging and Happy Ears
The deeper problem, Sofía came to believe, was that both fictions were rational responses to the same flawed setup: asking a human being to self-report a probability about their own performance, with their own incentives sitting directly on top of the answer. No amount of coaching fixes that math. You can tell a rep to "just be honest" every week for a year and you will still get sandbagging from the cautious ones and happy ears from the optimistic ones, because the number was never actually about the deal. It was about how the rep felt about reporting the deal.
The fix she eventually landed on wasn't a better spreadsheet or a stricter policy about rounding probabilities to the nearest ten. It was removing the self-report entirely — building the probability from what a deal was actually doing, not from what a rep was willing to say about it out loud on a call with their boss listening.
What an Activity-Based Forecast Actually Looks At
An activity-based forecast doesn't ask a rep how they feel about a deal. It looks at what's actually happening inside it — how recently the buying committee engaged, whether the conversation is still moving forward or has quietly stalled, how long the deal has actually sat in its current stage compared to deals that historically closed from that same position, whether the economic buyer has ever actually been in a room versus only a champion two levels down. None of that requires a rep's optimism or pessimism as an input. It requires the deal's own behavior, captured automatically from the calls, threads and notes already being logged, and rolled up into a number that updates itself the moment something in the deal actually changes — not once a week, under gentle interrogation, on a call everybody dreads.
To be clear about where the human still sits in this: the system doesn't decide the forecast is final and walk away. It surfaces the signal and the reasoning behind it — this deal looks softer than the rep's stage suggests, because engagement dropped off nine days ago — and a manager or rep can flag context the system doesn't have, like a client's internal reorg that explains the quiet spell. The number is a starting point built from evidence, not a self-report built from incentive. That distinction is the entire fix.
See a forecast that updates itself
The Number Got Smaller. Then It Got Trusted.
The first month Sofía's team ran the activity-based forecast alongside the old self-reported one, the two numbers didn't match — and the new one was lower. Several deals her happy-ears reps had been calling 70% showed up closer to 35% once actual engagement was the input instead of enthusiasm. That was an uncomfortable Friday. But the deals that closed that quarter tracked the new number far more closely than the old one, and by the second quarter, nobody on the team was negotiating with the forecast anymore, because there was nothing left to negotiate. The number wasn't a story anyone was telling. It was just what the deals were doing.
Sofía's Friday calls are twenty minutes now, not ninety, and they're about coaching the deals that need it, not interrogating the number. "I didn't get into sales leadership to be a lie detector," she says. "Give me a forecast that's already honest, and I'll spend my time actually helping my team close, not cross-examining them about a percentage."
Stop feeding the CRM. Watch it feed you.
