Forecast categories are the four buckets every deal lives in during the sales cycle: Commit, Best Case, Pipeline, and Omitted. They answer one question — how confident is this rep that this deal closes in the current period? — and they aggregate up from rep to manager to CRO to board. A Commit category that closes at 65% is a miscalibrated system; the whole forecast model breaks. A Commit that closes at 90–95% is a machine the CFO can plan around.
The close-rate bands cited throughout this page (Commit 90–95%, Best Case 50–70%, Pipeline 10–30%, and so on) are typical practitioner-calibration ranges, not published benchmarks. Treat them as starting points and calibrate on your own historical close rates over 6–8 quarters.
When to use this framework
Use forecast categories as the primary mechanism for period-level forecast construction. Every deal in your CRM has a close date and a stage; the forecast category is the human confidence layer on top of that structural data. Categories are what reps update weekly, not stage (stage changes when the deal moves forward; category changes when confidence shifts). Without explicit category discipline, your “forecast” is just a sum of close dates, which is always optimistic.
The four categories: explicit definitions
Commit
Definition. A deal the rep is willing to bet on closing in the current period, barring a rare and unexpected event. The rep should not Commit a deal if a single predictable obstacle — procurement delay, legal review, budget approval — could slip it.
What it requires. Verbal or written agreement from economic buyer on price and timeline. A mutual action plan (MAP) with a specific signed-paper date. Internal champion confirmed. No unresolved blockers that could move the close date.
Typical close rate benchmark. 90–95%. If your team’s Commit category closes below 85%, either reps are sandbag-committing (under-promising to over-deliver) or the category definition is too loose. Investigate which.
Common mistake. Committing deals where the legal or procurement cycle is not confirmed. A deal where paper still needs to go through legal and procurement has a 30–45 day slip risk even after a verbal yes. That is a Best Case, not a Commit, until the MAP confirms the close date is locked.
Best Case
Definition. A deal that could close in the current period if conditions go well, but still carries meaningful risk. The rep is not willing to bet their number on it, but it is reasonable to expect a portion of Best Case deals to close.
What it requires. Qualified opportunity: clear pain, engaged buying group, budget confirmed or directionally confirmed. Active engagement — the deal is moving, not stalled. A likely close timeline within the period, but without a confirmed MAP or signed-paper date.
Typical close rate benchmark. 50–70%. Some teams run lower (30–50%) if Best Case is used liberally for mid-funnel deals. The useful metric is not the absolute rate but the stability — Best Case conversion should be predictable within ±15 percentage points quarter-over-quarter. Wild swings indicate the category is being used inconsistently.
Common mistake. Treating Best Case as a catch-all for “deals I’m kind of working.” A deal should be in Best Case because the rep has a plan to close it this period, not because it’s a nice-looking logo.
Pipeline
Definition. Deals that are actively being worked but are not expected to close in the current period. They are real, qualified, progressing — just not this quarter.
What it requires. A qualified opportunity (pain, champion, budget directionally confirmed) that is early enough in the cycle that a current-period close would be unusually fast.
Typical close rate benchmark. 10–30% within the current period. The more useful number for Pipeline is the conversion rate to next-period Commit — a healthy pipeline converts 40–60% of its current-quarter value into next-quarter Commit.
Common mistake. Allowing unqualified deals to accumulate in Pipeline. If a deal has been in Pipeline for two or more quarters with no movement, it is not Pipeline — it is dead or dormant. Pipeline inflation is the primary driver of inaccurate forecast models.
Omitted
Definition. Deals explicitly excluded from the current forecast. This includes closed-lost, churned, opportunities on hold, and deals where the rep has decided the close date is too speculative to include.
What it requires. An active decision to exclude. Omitted is not a default state; it is a deliberate editorial judgment. A deal should be Omitted when including it would add noise, not signal, to the forecast.
Common mistake. Using Omitted as a limbo for deals the rep doesn’t want to close-lose in the CRM. Omitted deals should still have active close-date tracking and appear in pipeline reviews. The only deals that disappear entirely are closed-lost.
How the categories roll up
Category roll-ups at the manager level follow a simple rule in most teams: manager forecast = (Commit × 95%) + (Best Case × 50%) + (Pipeline × 15%). These coefficients should be calibrated against your own historical close rates, not borrowed from benchmarks. Run the regression on 6–8 quarters of data; your Commit coefficient might be 88% and your Best Case coefficient might be 40%. Use the real numbers.
At the CRO level, forecast typically incorporates three inputs and ships the number when all three agree within 5–8%: rep-up category roll-up, AI-driven deal score (Clari, Aviso, or BoostUp), and historical conversion model against this point in the quarter. When all three diverge, the CRO investigates before submitting to the board.
How Clari, Aviso, and BoostUp handle categories
The category names are the same across all three platforms. The differences are in how each platform augments category data and where it disagrees with rep-submitted categories.
Clari. Clari ingests rep-submitted categories from Salesforce and overlays an AI deal score from conversation, activity, and CRM signal. Its primary value is showing where the AI score diverges from the rep’s submitted category — a deal the rep has in Best Case that the AI scores at 20% win probability is flagged for manager review. Clari’s forecast roll-up is the dominant view at the CRO level for most enterprise RevOps teams using it. Category definitions in Clari mirror Salesforce defaults; the platform does not rename or restructure categories.
Aviso. Aviso takes a different approach: it runs its own time-series model independently from rep categories and produces a “most likely” number alongside best-case and worst-case scenarios. The key distinction is that Aviso’s AI forecast does not start from rep-submitted categories — it starts from historical pipeline data and activity signals, then validates or contradicts rep categories. This means Aviso’s output is useful specifically when you want an AI estimate untainted by rep optimism bias. The tradeoff is that Aviso’s model is harder to explain to a rep in a 1:1 (“the AI says 60%” is less actionable than “your deal is in Best Case but hasn’t had buyer activity in 14 days”). Aviso also supports consumption-based and usage-based forecast models that Clari and BoostUp do not, which matters for SaaS with variable ARR components.
BoostUp (rebranded as Terret on September 9, 2025). BoostUp focuses on deal-level inspection depth and audit trails rather than top-of-funnel category management. Its differentiator is tracking every forecast category change with a timestamp and reason — so when a rep moves a deal from Commit to Best Case, the system records when, who made the change, and what the associated activity signal was. This creates an accountability layer that prevents last-minute category-changes from distorting the forecast without a visible record. BoostUp also supports custom hierarchical roll-ups — by product line, territory, and customer segment — that are more flexible than Clari’s enterprise-oriented structure. For teams with complex overlay structures (e.g., product specialists, overlay AEs, regional splits), BoostUp’s roll-up model is typically easier to configure.
Category discipline: the rules that make it work
Categories fail when the definition is ambiguous, not when the tool is wrong. Three rules that eliminate most category drift:
Rule 1: Define entry criteria, not exit criteria. Tell reps what a deal needs to have to move into Commit, not what makes them move it out. “A deal is Commit when we have a signed MAP with a specific paper date and the economic buyer has confirmed the budget” is an entry criterion. Reps can check it. “A deal exits Commit when something bad happens” is unenforceable.
Rule 2: Weekly category review in the 1:1. The deal stage updates automatically as deals progress; the category update is a judgment call that requires a weekly human conversation. Manager-rep 1:1s should walk every deal at or near Commit category and confirm the entry criteria are still met.
Rule 3: Measure category accuracy by category. Track Commit-close-rate, Best-Case-close-rate, and Pipeline-close-rate as lagging indicators each quarter. If the rates move more than 10 points quarter-over-quarter without a known pipeline quality reason, the category discipline is breaking down — audit a sample of deals for each category to find out where the definition is being stretched.
Common pitfalls with paired guards
Sandbag committing. Reps who consistently commit low and over-deliver look accurate, but destroy capacity planning. Guard: track commit-versus-actual by rep. A rep who beats their Commit by more than 20% every quarter is sandbagging, not forecasting.
Pipeline inflation from zombie deals. Deals in Pipeline for 2+ quarters that never move inflate the model. Guard: auto-flag any Pipeline opportunity without a CRM activity update in 30 days. Require reps to either move it or close-lose it.
Recency bias in Best Case. Reps move deals to Best Case after a good call, then forget to move them back when the deal stalls. Guard: require Best Case to have a closing activity (proposal sent, next call scheduled) within the last 14 days to remain in category.
AI score as the only input. Using AI deal scores alone without rep category input loses ground truth on negotiation context. Guard: flag deals where AI score and rep category diverge by more than two tiers, and require a manager note explaining which to trust.
Related
- Forecast accuracy — the metric forecast categories are in service of
- Deal stage definitions — the structural layer below categories
- Pipeline velocity — the upstream metric that determines whether Commit is achievable
- Clari — enterprise forecast layer, category roll-up
- Aviso — independent AI forecast, consumption model support
- BoostUp — deal-level audit trails, flexible roll-up hierarchies