A predicted close date is when a model expects a deal to resolve, based on the behavior of deals like it. The rep-entered date says when a seller intends to close. The predicted date says when comparable deals have historically closed. Both belong on the record, and the distance between them is the useful part.
Where the Prediction Comes From
Timing is predicted per group rather than per deal. A single opportunity has no history to fit against, so ORM groups each opportunity with a machine learning model and predicts a close-timing curve for the group. Those curves run from 1 to 80 weeks. Most of the expectation lands before week 12, and very few groups carry meaningful expectation past 52 weeks.
That structure makes deal age interpretable. A deal is not old because the calendar says so. It is old relative to the curve of the group it belongs to, so ninety days in a fast group means the window has already passed while ninety days in a slow group is unremarkable.
Why Rep-Entered Dates Drift
Forecast processes ask sellers for a quarter, so dates cluster at quarter boundaries regardless of what the deal is doing. The result is a pipeline that looks loaded for the current period and resolves differently. ORM's read is that roughly 20 percent of the pipeline dated inside a quarter on day one actually closes inside it, which leaves most of the visible in-quarter value unrealized.
Movement in that date is the strongest warning you get. ORM treats a rep changing the close date as the best slippage signal available, and a deal that slips from one quarter into the next becomes less likely to close even when it stays in commit.
Reading the Gap
Sort open deals by the distance between predicted and entered dates and the inspection list writes itself. A commit deal predicted to land next quarter is exposure the roll-up hides. A deal the model expects sooner than the rep does often means the seller is sandbagging or has not updated the record.
Watch the silent ones too. ORM counts a deal as touched only when its stage, close date, or amount changes, and applies a twelve-month rule for most customers. The earliest warning on any deal is the absence of a signal, not a bad one.
Using It in the Forecast Call
Bring the gap list, not the whole pipeline. Ask for evidence on each deal where the model and the seller disagree, and record the reason so the pattern can be graded later.
Over a few quarters the gaps show which segments and which reps run optimistic, which is a better input to coaching than a total. Pair this with deal slippage tracking and check it against forecast accuracy rather than against pipeline coverage, since coverage tells you nothing about when anything lands.
Frequently Asked Questions
What is a predicted close date?
It is a model's estimate of when a deal will close, derived from the resolution timing of comparable deals. It exists alongside the rep-entered date rather than replacing it in the CRM.
How is a predicted close date calculated?
Deals are grouped with similar records and each group gets a timing curve. ORM's curves run from 1 to 80 weeks, with most of the expectation falling before week 12 and very few groups carrying meaningful expectation past 52 weeks.
Why are rep-entered close dates unreliable?
They cluster at period boundaries because forecast processes ask for a quarter, not a probability. ORM finds that about 20 percent of the pipeline dated inside a quarter on day one actually closes in that quarter.
What does the gap between predicted and entered dates tell you?
It quantifies exposure. A commit deal whose predicted timing falls outside the quarter is carrying risk the roll-up does not show, and that deal should be inspected before the number is submitted.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like predicted close date into prescriptive action for your team.
Schedule a Demo