How to Build a Sales Forecast Your Leadership Can Trust

Sales forecasting is estimating how much revenue your team will close in a period, built deal by deal from the open pipeline. A forecast that leadership trusts combines two methods, usually the reps’ commit call and a weighted pipeline, and reconciles them weekly on clean close dates and amounts, in a spreadsheet or a CRM like Senitix CRM.

For sales managers, VPs of Sales and sales ops leads at B2B companies, the work comes down to six steps, each covered below:

1. Separate the forecast from quota, target and pipeline.
2. Choose two methods that check each other.
3. Weight open deals by your own stage win rates.
4. Put every deal in a forecast category by written rules.
5. Clean stale close dates, amounts and zombie deals before every call.
6. Review weekly, and measure accuracy against what you submitted.

What is sales forecasting, and what is it not?

A sales forecast predicts closed-won bookings for a defined period, such as a month or a quarter, with a named deal behind every dollar. It answers what the CEO and CFO ask: how much will we sign, and how sure are we? Hiring, spending and board commitments rest on that answer.

Three neighboring numbers get mixed up with it, and the mix-up costs credibility:

• Quota is what the company asks each rep and team to close: an assignment, not a prediction.
• Target is the plan the business is built around. A forecast that always lands exactly on target is a wish.
• Pipeline is the total value of open deals: inventory, most of which will not close this period.

Report the forecast as its own number and the gap to target as a separate line. Every method below reads from the deal record, so if your team is still deciding what that record should hold, start with what a CRM is and what it keeps.

Which sales forecasting methods should you use?

B2B teams usually choose among four methods. Each fails in a predictable way, so a trusted forecast runs two side by side.

Rep call

Each rep names the deals they expect to close and puts a total on them. It captures what no field does, such as a champion who went quiet, but it also carries each rep’s optimism or caution. Use it when deals are large and few, or the product is too new for win-rate history.

Weighted pipeline

Each open deal contributes its amount times a probability, and the forecast is the sum. It is mechanical, and two people running it get the same answer. It breaks when probabilities are guesses, amounts are stale, or one deal is a large share of the period.

Stage probability

A weighted pipeline in which the probability comes from the deal’s stage, not the rep’s estimate. Percentages typed in at setup are placeholders; replace them with your own stage win rates: of the deals that reached a stage and have since closed, the share that were won. Use your last four quarters, recalculate quarterly, and keep separate rates for new business and renewals.

Historical run-rate

Project recent bookings forward, adjusted for seasonality and known changes such as new hires. It needs no pipeline data and is hard to game, but it is blind to the present: a lost segment or a price change shows up only after the period ends. It leads for high-volume, short-cycle sales and checks the other methods everywhere else.

A decision rule: small, numerous deals that close within the period call for run-rate, checked by a weighted pipeline. Large, few deals call for the rep call, checked by stage probability. Teams in between run the rep call and stage probability together, and the gap between them sets the weekly agenda.

How do you calculate a weighted pipeline forecast?

The formula: weighted forecast = closed-won to date + the sum of (open deal amount × stage win rate), counting only deals whose close date falls inside the period.

Example: a 12-rep B2B SaaS team in Austin sells annual subscriptions. In week six of the fourth quarter it has booked $420,000 against a $1,500,000 target, and its stage win rates come from its own last four quarters of closed deals. All figures are illustrative.

Stage	Open deals	Pipeline amount	Stage win rate	Weighted amount
Discovery	34	$1,360,000	10%	$136,000
Evaluation	21	$1,050,000	25%	$262,500
Proposal	11	$660,000	45%	$297,000
Negotiation	6	$420,000	70%	$294,000
Total open	72	$3,490,000		$989,500

That is $420,000 closed plus $989,500 weighted: a forecast of $1,409,500, which is $90,500 short of target. Three checks turn the number into a call:

1. Coverage. The team needs $1,080,000 more and holds $3,490,000 of open pipeline closing this quarter, just over three times what it still needs. Coverage says there is enough to work with; win rates say how much converts.
2. Concentration. If one Negotiation deal is worth $180,000, weighting counts $126,000 for it, yet it will close for $180,000 or for nothing. Take any deal big enough to swing the quarter out of the weighting and call it in or out on its own.
3. Run-rate. If the last three quarters booked $1.25 million, $1.31 million and $1.38 million, the trend points to about $1.45 million, near the weighted $1.41 million, so the two methods agree. When they don’t, find the reason before anything goes to leadership.

Commit vs. best case: what is the difference?

Forecast categories put the rep’s judgment next to the math. Each open deal sits in one category, and the categories roll up into the numbers leadership sees:

• Closed: won and booked.
• Commit: the rep would stake the number on it. The buyer has confirmed the decision date and the steps to signature.
• Best case: could close this period if named things go right, such as a budget sign-off or a security review. The risk is written on the deal.
• Pipeline: a real deal that will not close this period unless something changes.
• Omitted: lost, or deliberately left out.

Leadership usually gets two numbers: the commit forecast (closed plus commit), a floor the team expects to beat, and the best-case forecast (closed, commit and best case), a realistic ceiling. In the Austin example, reps commit $620,000 of open deals and mark $380,000 as best case: a commit forecast of $1,040,000 and a best case of $1,420,000, against a weighted $1,409,500. The VP of Sales submits $1,200,000, the commit plus $160,000 of best-case deals that survived review.

The spread between commit and weighted is the conversation. A weighted number near best case and well above commit means the reps are cautious or the win rates are generous, and the weekly review finds out which, deal by deal.

Categories work only if every rep applies them the same way, so write the commit test down as questions a manager can check on the record:

• Is the economic buyer engaged?
• Has the buyer agreed to a signature date?
• Are the legal and procurement steps scheduled?
• Has the price been accepted in writing?

A deal that fails one is best case at most.

What breaks a sales forecast?

Sales forecasting fails quietly when the inputs are wrong, whatever the method. Check for these five problems before every forecast call:

• Stale close dates. A date in the past, or one that rolls to every month-end, means nobody has asked the buyer. Count the pushes: a deal that has slipped twice leaves commit until the buyer confirms a new date.
• Stale amounts. List price after a discount, or last quarter’s seat count after the scope changed. The amount should follow the latest quote or the buyer’s latest number in writing.
• Sandbagging. Winnable deals kept out of commit so the quarter looks like an over-delivery. The symptom is a commit beaten by a wide margin every quarter, and deals that jump from best case to closed-won in the final week.
• Happy ears. Enthusiasm heard as a decision. The symptom is a commit deal with no next meeting booked, no contact above the champion and no reply in two weeks.
• Zombie deals. Months without activity, never marked lost. They inflate the weighted number, then vanish together at quarter end. Close them as lost, with a reason.

Sandbagging and happy ears are opposite biases, and the bias measure below exposes both. The pipeline hygiene routines in our guide to CRM best practices prevent most of the rest.

How do you run a weekly forecast cadence?

Sales forecasting earns trust when it is produced the same way every week, by the same people, in the same order. This cadence fits a 12-rep team with two frontline managers:

1. Monday by noon, each rep updates close date, amount, stage, next step and category on every open deal in the period. Output: current deal records.
2. Monday afternoon, sales ops runs the exception list: past close dates, deals with no next activity, amounts changed since last week, and deals that moved into or out of commit. Output: the deals to question.
3. Tuesday, each frontline manager meets each rep and reviews only commit deals, best-case deals and the exceptions, not the whole pipeline. Output: the manager’s call for the team, with every adjustment and its reason written down.
4. Wednesday, the VP of Sales rolls the teams up, compares commit, best case, weighted pipeline and run-rate, and sends the CEO and CFO one number with the five deals most likely to move it. Output: the submitted forecast.
5. Friday, sales ops saves a snapshot of the submitted number and the deals behind it. Output: the record accuracy is measured against.

Keep a manager’s adjustment separate from the rep’s call, so you can later see whose judgment was closer.

How do you measure forecast accuracy?

Measure the forecast you submitted at fixed points in the period, not the one revised on the last day. A common formula is forecast accuracy = 1 − (|actual − forecast| ÷ actual).

Example: the Austin team submitted $1,200,000 in week six and closed $1,290,000. The error is $90,000 ÷ $1,290,000, about 7%, so accuracy is about 93%. Take the reading at the same weeks every quarter, such as weeks two, six and ten, and judge it against your own last four quarters, not another company’s figure.

Accuracy hides direction, so track bias too: forecast minus actual, with the sign kept. Always under points to sandbagging; always over points to happy ears or generous win rates. Read it per rep and per manager, not only in total. The forecast-to-actual gap is also one of the measures in the benefits of a CRM and how to measure them.

How does Senitix CRM support sales forecasting?

A spreadsheet can do every calculation above. The hard part is keeping close dates, amounts and next steps current, and that is easier when they sit on the same deal record as the calls, emails and quotes that change them.

In Senitix CRM, deals move through stages your team defines, on a list or a Kanban board, each with an amount, a close date, a probability and an owner. Each stage can show guidance text and up to five key fields, so the commit questions your team agreed on appear on the deal at the stages where they apply; they are shown, not enforced. Reports start from templates and can be emailed on a schedule, so a report of open deals sorted by close date, oldest first, can reach each manager before the Monday review.

The forecast grid rolls open deals up by forecast category, such as Commit, Best Case and Pipeline, against each rep’s quota, records a manager’s adjustment and its reason alongside the calculated amount, and compares forecast with actual. Senitix AI can summarize a deal’s email thread before the review; the rep still makes the call.

The details are under reports, dashboards and forecasts. Compare plans on the pricing page, or talk to sales about the forecast process you run today.

Frequently asked questions

Should you forecast bookings, revenue or ARR?

Forecast the number leadership plans on, and define it once. In most subscription businesses, sales forecasts new bookings or new ARR, and finance turns bookings into recognized revenue on its own schedule. Whatever you pick, the amount field must mean the same thing on every deal, annual value or total contract value, or the weighted math quietly adds unlike numbers.

How far ahead should a sales forecast look?

Forecast the current quarter weekly, deal by deal. Next quarter can be forecast monthly, but most of its deals are still early, so lean on stage probability and run-rate and say so when you submit it. Beyond two quarters, deal data is too thin to forecast from; plan from capacity instead, meaning ramped reps, their quotas and the attainment they have reached historically.

Can AI forecast sales for you?

Partly, and only as well as the data underneath. AI sales forecasting models can help teams with a large volume of clean deal history, but they learn from the same close dates, amounts and stages, so stale data yields a confident wrong number. The forecast grid in Senitix CRM is not a predictive model: it rolls up the deals and categories your team maintains, so every calculated dollar traces to a deal.