How a decision model can support CRM forecastingEvery point of probability has a reason.

Most sales forecasts are a sum of hopes. Each rep puts a probability on each deal, the CRM adds them up, and the quarter ends somewhere else. A decision model does not replace those estimates. It sits next to them, and shows where they are too hopeful, and why.

The problem with the rep's number

A rep's estimate mixes two things: what happened in the deal, and how the rep feels about it. The feeling is not useless. But it moves with mood, with targets and with the last call that went well. Nobody can check it, because it comes with no reasons.

You see it at the end of the quarter. Deals marked 80% close far less often than 8 times in 10. The forecast was not wrong because people were careless. It was wrong because nothing ever compared the numbers with what really closed.

What a decision model does

A decision model turns evidence into a probability, one step at a time, and keeps every step visible. In Parley Notes it works in three steps.

1. It reads every conversation. Buying signals, objections, who was in the room, how firm each promise was, and how long the deal has been quiet. Parley Notes already writes up your calls, so the evidence is there without anyone typing it in.
2. It scores the deal. It starts from how often your deals close at that stage. Then it moves the probability a little with each signal, up or down.
3. It checks itself. Every won or lost deal, and every Confirm or Fix you tap, makes the next score more accurate.

Enterprise plan: how its probability is builtIllustrative example

Base rateat Proposal, from your deals

31%

Budget confirmedKick-off

+7

Decision maker in the roomQ3 review

+5

Price objectionQ3 review

−8

Quiet for 21 daysNo reply since

−13

Parley Notes probability

22%

Rep estimateWhat the rep entered

80%

0%25%50%75%100%

Read the example from top to bottom. At the Proposal stage, 31% of your deals usually close. The budget was confirmed at the kick-off, and the decision maker joined the pricing review. Both push the number up. Then came a price objection, and three weeks of silence. The deal ends at 22%. The rep still says 80%.

Neither number is the truth. But only one of them tells you what to do next: get a reply, and answer the price objection.

Calibrated instead of hopeful

A probability is only useful if it means what it says. When the model says 40%, about 4 in 10 of those deals should close. This is called calibration, and you can test it: group past deals by the probability they were given, then count how many closed.

Parley Notes runs that check on your own history. Each closed deal adjusts the base rates and the weight of each signal. While your history is short, the model leans on what it learned from companies like yours, and the forecast range stays wide. As your own wins and losses add up, your history takes over and the range narrows.

Beside the rep and never in place of the rep

Rep estimate and Parley Notes, deal by dealIllustrative example

Deal	Stage	Value	Rep	Parley Notes	Gap	Biggest reason
Enterprise plan	Proposal	EUR 24k	80%	22%	−58	Quiet 21 days
Clinic rollout	Negotiation	EUR 16k	90%	41%	−49	No next step
Brand refresh	Qualified	EUR 14k	50%	58%	+8	Budget confirmed
Second site	Negotiation	EUR 10k	85%	79%	−6	Next step agreed
Fleet tracking	Qualified	EUR 21k	40%	33%	−7	Competitor mentioned

The gaps are where the forecast meeting should go. Enterprise plan: 80% against 22%, because nobody has replied for 21 days. Clinic rollout: 90% against 41%, because the deal has no next step. And sometimes the model is the optimist. Brand refresh scores 58%, above the rep's 50%, because the budget is confirmed.

A rep who disagrees can say why, and fix the record. That fix is evidence too, and the model learns from it.

A forecast with a range

The quarter, two waysCalibrated on your last 48 deals

Closed so far

EUR 118k

Calibrated forecastParley Notes

EUR 265k

Rep estimateSum of the reps' numbers

EUR 410k

0150k300k450k

Adding up the calibrated probabilities gives a forecast of EUR 265k, somewhere between 238k and 295k. The reps' estimates add up to EUR 410k. The target is EUR 300k. One view says you will beat the target easily. The other says you are about 35k short, and shows which deals could close the gap.

That is the point of a forecast you can trust. It does not just give a better number. It tells you which conversations to have this week.

What it cannot do

• It cannot see what was never said. A deal discussed only over lunch has thin evidence, and its score comes with a wider range.
• It does not know your market better than you do. It knows your history, and it asks you to confirm or fix.
• It does not decide for you. It shows the reasons. You choose what to do with them.

How to start

Record your sales calls and let Parley Notes add the deals from your notes. Confirm or fix what it suggests. You get a first forecast once a few deals are in, and it gets sharper with every deal you win or lose.