A forecast is not a pipeline report. It is a committed number, submitted on a cadence, rolled up through the org, and tracked against actual bookings every quarter.
Sales forecasting is the process of predicting future revenue a company will book over a specific period, typically monthly or quarterly. It combines pipeline data (deals, stages, close dates) with historical close rates and rep judgment. Modern forecasting uses weighted pipeline math, forecast categories (commit, best case, pipeline), a weekly submit-and-lock cadence, and AI-generated risk signals. The output drives board commits, hiring plans, and cash flow projections.
The six things every operator should know before building, running, or buying a sales forecasting process.
Sales forecasting predicts bookings over a defined period, usually a month or quarter. The number is a commitment reps, managers, and leaders are held to, not a wish list. It feeds board updates, hiring plans, cash flow, and quota planning. A forecast carries accountability a pipeline report does not.
Modern forecasts split expected revenue into three bands. Commit is the number a rep bets their quarter on. Best case is the stretch if everything breaks right. Pipeline is the universe of deals that could still land. Reporting all three shows the confidence band, not just a single point estimate that hides risk.
Bottoms-up asks reps to call their deals. Top-down starts from a target and divides it. Pipeline-weighted multiplies open deal amounts by stage probability. AI-assisted overlays historical patterns onto rep calls. Most mature teams combine methods, bottoms-up as the primary input, pipeline-weighted as a sanity check, AI as a tiebreaker.
Reps submit a number by Friday evening. The number locks for the week. Managers roll up by Monday morning. The forecast meeting runs against the locked number, not whatever got edited at 10:52 that morning. At quarter end, the forecast is reconciled against actual bookings and rep accuracy scores update.
A manager sees every direct report's submitted number stacked as rows, with drill-down to the deals behind each number. The VP sees every manager's rollup stacked above that. The entire org pipeline rolls up on one page, with the quarter target line drawn across the top. No pivot tables, no exported CSVs.
A forecast is only useful if it is accurate. Mature revenue teams track rep accuracy as a scored metric over the quarter, forecast variance week-over-week, and category movement (deals sliding from commit down to pipe or omit). These signals tell leadership which reps to trust at face value and which calls to pressure test.
Every sales forecasting process is some mix of the four methods below. Picking the right blend depends on data maturity, deal size, sales cycle length, and how much the business can afford to miss. None of the methods is wrong on its own. The wrong move is picking one and ignoring the others.
The most common method. Each rep categorizes their open deals into commit, best case, pipe, or omit, submits a weekly number, and the manager sums the rollup. Strongest where deals are large, cycles are long, and reps know their accounts well. Weakest where reps sandbag, inflate, or guess. Requires a submit-and-lock cadence to work.
The quarter target comes from the board or finance, then gets split across regions, teams, and reps based on quota and territory. Fast, easy, requires almost no CRM data. Weakest for predicting the actual quarter, strongest for setting the quota and modeling the shape of the business. Most teams use it to set the target, not to call the quarter.
The classic formula. A $100k deal at 60% probability contributes $60k. Sum every weighted deal in the quarter and the result is the pipeline-weighted forecast. Works best when stage probabilities reflect reality, which almost nowhere they do out of the box. Teams that run this method well recalibrate stage probabilities every quarter against actual conversion.
Look at the same point in prior quarters and extrapolate. If week 4 of Q3 last year carried $2.1M of commit and the quarter closed at $5.4M, the ratio is roughly 2.6x. Apply that to this quarter's week 4 commit to get a projection. Lightweight sanity check, not a primary method. Breaks when seasonality or sales motion changes year to year.
Modern CRMs run machine learning across deal history, engagement signals, email cadence, and stage velocity to flag which commit-category deals are quietly slipping. The AI does not replace the rep call, it pressure tests it. The best use is a side-by-side view: rep's commit number next to the model's number next to the delta, debated in the forecast meeting.
Top-down sets the target. Bottoms-up produces the primary forecast. Pipeline-weighted and AI-assisted run as sanity checks against the rep call. When the three agree, confidence is high. When they diverge, the forecast meeting has a specific list of deals to pressure test. The blend is the strongest signal any one method is wrong.
Forecasting is a workflow, not a report. The teams that run it well have a weekly rhythm every rep, manager, and leader can describe without looking at a doc. Here is the shape of that cadence from Friday afternoon to the following Monday review.
Reps who have not submitted their weekly number get an in-app and email nudge at 3pm local time. The nudge includes a one-click link to the submit page. The point is not to punish, it is to make submit the path of least resistance so the manager has every number by close of business.
By 5pm every rep has submitted. The number locks. Any resubmit after lock is audit-logged with timestamp, reason, and the delta from the prior submit. The lock turns the weekly cadence from a cultural habit into a product feature. Managers run their rollup against a frozen snapshot, not whatever the rep edited over the weekend.
An overnight job scores every commit-category deal against deal velocity, stage age, engagement signals, and historical outcomes. Deals that look like historical slip patterns get a risk flag. Monday morning, the manager opens the forecast page and sees the AI flags lined up next to the rep calls. Nothing is auto-removed. The flags prompt a conversation.
The manager has already walked the rollup table, challenged the rep calls on at-risk deals, and submitted the team number. The VP opens the leader view and sees every manager stacked, every rep inside each manager, every deal inside each rep, with the quarter target line across the top. No pivot table, no exported CSV, no "give me a few hours."
The forecast meeting is thirty minutes. Everyone is looking at the same page. The agenda is: at-risk commit deals, deals pulled forward, deals pushed, and variance from last week. The meeting is not a status update. It is a working session on the small number of deals that will swing the quarter.
When the quarter closes, every weekly submit gets reconciled against actual bookings. Rep accuracy scores update. Managers see which reps called the quarter inside the band, which sandbagged, and which inflated. The scores feed the following quarter's weight on each rep's call. Over time, trusted reps earn more weight on their commit number.
Strkr's sales forecast module ships the submit lock, the three-tier categories, the manager rollup, variance tracking, rep accuracy scores, and Strkr AI risk flags out of the box. Fourteen-day trial, no credit card, every forecast feature on every plan.
A pipeline is a list of open deals in various stages. A forecast is a committed revenue number for a defined period. The pipeline is the input, the forecast is the output. A pipeline report shows "here are the deals we are working." A forecast says "here is what we will book this quarter." Pipeline is a universe, forecast is a commitment.
Most B2B revenue teams forecast weekly during the quarter, with a tight submit-review-rollup cadence every Friday-Monday. Transactional or high-velocity teams may forecast daily. The deeper the sales cycle, the longer the cadence can stretch, but weekly is the baseline. Monthly-only forecasting leaves too much room for deals to slip between checkpoints.
Pipeline-weighted forecasting multiplies each open deal's amount by its stage probability, then sums the results. A $100k deal at 60% probability contributes $60k to the weighted forecast. The method works only when stage probabilities are recalibrated against actual conversion data every quarter. Out-of-the-box CRM probabilities almost never reflect reality.
Forecast categories (typically commit, best case, pipeline, omit) separate deals by confidence. Commit is highly likely, best case is possible if conditions break right, pipeline is still being worked, omit is out. Reporting all three columns separately gives leadership a confidence band instead of a single number that hides underlying risk. The band is the signal.
Mature revenue teams target rep-level forecast accuracy above 85% and quarterly variance under 10%. Below 85%, the forecast is noise and leadership stops trusting it. Variance over 10% week-over-week means the underlying pipeline is unstable, not just the call. Tracking accuracy as a per-rep metric over time creates a feedback loop that improves calls each quarter.
No. AI pressure tests the rep call, it does not replace it. The best setup runs the model and the rep call side by side with the delta flagged. When the model and the rep disagree, that specific deal gets debated in the forecast meeting. Reps know account context AI cannot see. AI catches patterns reps cannot. The combination outperforms either alone.
Smaller teams run forecasting out of spreadsheets. Mid-market and enterprise teams use the CRM's forecast module or a dedicated forecasting tool. The must-have features are: weekly submit lock, forecast categories, manager rollup through the org tree, variance tracking, rep accuracy scoring, and ideally AI risk flags on at-risk deals. Spreadsheets break the moment the org has more than a few reps.
Reps own their individual number. Managers own the team rollup. The VP of Sales or CRO owns the full org forecast and the commit back to the board. Finance and RevOps run the reconciliation and scorekeeping. Everyone is accountable to the number they submit, which is why the submit lock and audit log matter so much.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
