# Forecasting bottom-up: key accounts, renewals, new business > Build a sales forecast in three parts: key accounts judged one by one, the base of renewals, and new deals weighted by pooled stage conversion; then commit, best case and expected. - Course: [B2B Marketing and Sales: Win Your First Company Customers](https://productovi.com/courses/b2b-marketing-and-sales.md) - Module: The pipeline - Lesson 11/32, ~21 min - Source: https://productovi.com/courses/b2b-marketing-and-sales/forecasting-bottom-up ## Leo's number **Alex Tutor:** You're still running Klasna for Teams' pipeline, now with stages that mean something. This week Marta has to tell Leo what Teams will sign this quarter. **Marta:** Leo's method: each owner's probability times the deal's value, add it all up. His sheet says over ten million for the quarter. **Sofia:** Isn't that a normal weighted pipeline? **Alex Tutor:** The arithmetic is fine; the inputs aren't. Those probabilities are feelings. We build the **sales forecast** bottom-up instead: key accounts one by one, the base of renewals, and new deals weighted by what history says. > **Exercise:** Why is «each owner's probability × deal value, summed» a weak forecast for Teams? ## Key accounts, one by one **Sofia:** Key accounts first: Veloria, Cobalt, Orchid? **Alex Tutor:** Each gets a judgement with evidence, not a percentage: can this deal sign this quarter, given the buyer's own calendar? **Marta:** Veloria wants a paid pilot first and a start early next year. Cobalt is under contract with its current school until March. Orchid is still in discovery. **Sofia:** So none of them signs before January, whatever the sheet says. **Alex Tutor:** Right. Big deals distort a forecast the most, so they get names and reasons. One key account in the wrong quarter can be most of the error. > **Exercise:** Cobalt is at the proposal stage. Its contract with its current school runs until March, and the owner has put it in November at 60%. How should the forecast treat it? ## New deals: weight by history **Alex Tutor:** Third, the many small new deals. Here a percentage helps, if it comes from history: of the deals that reached each stage, how many were won? **Marta:** Last four quarters at the proposal stage: 2 won of 3, then 1 of 5, 1 of 1, 1 of 4. **Sofia:** The average of those rates: 66.67, 20, 100 and 25. That's 52.92%. **Alex Tutor:** That average lets a quarter with one deal count as much as a quarter with five. Pool them instead: 5 won out of 13 reached, **38.46%**. **Marta:** Fourteen points lower. On a few million of proposals, that's real money. **Alex Tutor:** Use pooled rates, and only on live deals. A stuck deal at full stage odds is the oldest lie in a pipeline: it goes to the honest question first. > **Exercise:** At the contract stage, Teams won 2 of 2, 1 of 2, 1 of 1 and 1 of 2 deals over four quarters. What pooled conversion should the forecast use? **Alex Tutor:** Now put a pooled rate to work on one deal. > **Exercise:** A live new deal worth 960,000 UAH sits at the proposal stage, where Teams' pooled conversion is 38.46 %. How much does it add to the expected forecast? In UAH. **Alex Tutor:** And a deal that has gone quiet? > **Exercise:** A new proposal-stage deal worth 720,000 UAH has had no buyer activity for 40 days. What should it add to the expected forecast this week? ## The base and slippage **Alex Tutor:** Back to the second part, the base: renewals and expansions of current clients. Like key accounts, judge each one. **Marta:** The Lviv studio renews in December, and the contract is already with their accountant. They also want to add their support team. **Sofia:** Then the renewal is base, and the expansion is still a proposal. And a renewal whose client hasn't replied in five weeks? **Alex Tutor:** That isn't base, it's a risk: call them this week. And watch **slippage**: a close date that keeps moving is a deal telling you it isn't ready. Our rule flags two moves. > **Exercise:** **True or false:** a renewal is always safe to count in full, because the client already knows the product. ## Commit, best case, expected **Alex Tutor:** Then three numbers. **Commit**: what you'd bet on. Live renewals ready to sign plus new deals at the contract stage with recent buyer action, at full value. **Alex Tutor:** **Best case**: commit plus live deals at the proposal stage that could close if things go well. **Expected**: the base plus every other live deal times its pooled stage rate. **Sofia:** Why three? One number is simpler. **Marta:** One number hides the range. I want to tell Leo what's nearly certain, what's possible and what's likely. > **Exercise:** Match each forecast number with how it is built. **Alex Tutor:** One last check before the numbers go to Leo. > **Exercise:** **True or false:** if the expected number comes out above best case, the forecast needs checking before it goes to Leo. **Sofia:** And the key accounts go nowhere? **Alex Tutor:** They go in the quarter they can actually sign, with a line each in the note. Moving them isn't pessimism; it's the buyer's calendar. **Marta:** Leo will hate seeing his ten million shrink. And he'll ask why I didn't just average the quarters. **Alex Tutor:** Show him why, deal by deal. A forecast that's honest in October saves an awkward conversation in December. > **Exercise:** Write Leo two or three sentences explaining why the forecast uses the pooled proposal-stage conversion (5 of 13, 38.46%) rather than the average of the four quarters' rates (52.92%). Over the last four quarters, proposal-stage deals were won 2 of 3, 1 of 5, 1 of 1 and 1 of 4. > **Key ideas** > > - Build the sales forecast bottom-up in three parts: key accounts one by one, the base of renewals and expansions, and new deals weighted by history. > - Judge each key account against the buyer's calendar: a switching buyer under contract, a pilot first or an early stage moves it to a later quarter. > - Weight new deals by pooled stage conversion (total won ÷ total reached), not by an average of quarterly rates, and only live deals. > - A renewal with weeks of silence is a risk to act on, not base. A close date moved twice is slippage. > - Report three numbers: commit (nearly certain), best case (possible) and expected (likely, weighted). > - Explain every move deal by deal; an honest forecast early beats an excuse later. **Alex Tutor:** In the case, you'll do it for real: thirty open deals, the stage history, and Leo's ten million. - Previous lesson: [Qualifying a deal: four honest questions](https://productovi.com/courses/b2b-marketing-and-sales/qualifying-four-questions.md) - Next lesson: [Case: the quarter's forecast review](https://productovi.com/courses/b2b-marketing-and-sales/quarter-forecast-review-case.md) --- Productovi Learning. Free, no sign-up needed to read.