> ## Documentation Index > Fetch the complete documentation index at: https://b2-b-playbook.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Forecasting **Last reviewed:** 2026-08-29 · **Reading edit:** 2026-09-12 A forecast combines what is in the pipeline with evidence about whether and when deals will close. Use a method that fits your sales process, keep the underlying records current, and make changes in confidence explainable. The review should reveal risks while there is time to act. Check the pipeline; Test timing and confidence; Update the forecast *Reading guide: check the pipeline → test timing and confidence → update the forecast.* ## Use this when * The board number and the CRM number cannot be reconciled in one sitting. * Managers “feel” the quarter and cannot say why a deal is in or out. * Pipeline coverage is a screenshot, not a gap with a strategy. * You are about to hire a second seller and the only forecast is a spreadsheet tab named FinalFinal. ## Do not use this when * There is no [ICP](../01-strategy-and-buyers/icp.md) and no stages with entry/exit criteria. You will forecast noise. * The founder still sells three logos. A CRO forecast stack is costume. * You need to *create* pipeline. That is channel, outbound, or [account planning](../06-account-field-and-partner/account-planning.md)—not a call. * Legal reporting (public company guidance) must set the process. Get qualified owners; this is not securities advice. ## A few useful terms | Word | Meaning here | | -------------- | --------------------------------------------------------------------------- | | **Hygiene** | Stages, amounts, close dates, and next steps match reality | | **Stage** | A state with **entry/exit criteria**, conversion math, and required actions | | **Category** | A published bucket (for example commit / best case / pipeline / omit) | | **Call** | The number a person will stand behind in this period | | **Coverage** | Pipeline relative to the target (and the **gap** plus the plan to close it) | | **Top-down** | High volume, high velocity: rates and cohorts more than deal archaeology | | **Bottoms-up** | Long cycle: inspect the deals | ## Keep this in mind **Always have a reason. Deliver bad news early. Ask for help. Document the process. Align with leadership and evolve it.** A forecast without a reason is a wish. Bad news on Friday of week 12 is a reputation event. Data is a path to the answer. It is rarely the answer. ## How to do it ### Step 1: Understand how deals really progress Inspect, in this order: * **Buyers.** Process and messaging that match where they actually are. * **Journey.** The problem, their familiarity, accelerators and decelerators on *this* deal. * **Sales process.** If it is not consistent, you cannot measure it. Consistency is the foundation of repeatability. * **Metrics.** Velocity versus the plan’s assumptions. Turn a qualitative journey into something you can compare. * **Hygiene.** Does the pipeline reflect what you need to hit the number? Where is the bottleneck? RevOps is not the enemy. * **The forecast itself.** Do not reinvent a private model every week. Protect your reputation. Block time for hygiene or it will lose to “real selling.” Calendar tools are optional; the block is not. ### Step 2: Check the buyer experience and funnel records Forecast quality is downstream of deal quality: * **Discovery / demo:** account mapping before the call; maturity of the buyer; a real [account plan](../06-account-field-and-partner/account-planning.md) when the account is complex. * **Consideration:** mutual action plans; async explanation; a standard evaluation path; value-add touches you can count. * **Negotiation:** known redline minimums; business terms agreed with the business contact; legal as a partner, not a surprise. * **Handoff:** async context (pain, solution, goals, dependencies) so [customer success](../08-lifecycle-and-customer-marketing/customer-success.md) does not start from zero. * **Lost:** feedback, a cadence, a memory—so “lost” is data, not a sulk. A methodology exists to help sellers understand needs, prioritize, repeat, and give the customer a better path. Bake it into the process. A poster in onboarding is not a methodology. ### Step 3: Use consistent pipeline definitions Document conversion by stage. Compare people and periods. Define **opportunity health** with your manager and RevOps: a short score you can fill without a novel. Then look at where an AE can actually improve—not a pep talk. ### Step 4: Choose a method that fits your sales process * **Top-down** when volume and velocity dominate. * **Bottoms-up** when cycles are long (the class used \~90+ days as a rule of thumb for enterprise-shaped deals). You can mix: bottoms-up on the named few, rates on the rest. What you cannot do is swap models mid-quarter without telling the people who call. Rewind once a year: how was the **annual plan** built? A forecast that ignores the plan’s assumptions will always look “wrong.” ### Step 5: Design and review the forecast process A better forecast process: 1. Clear sales process 2. Goals, metrics, and **shared words** 3. Roles and responsibilities 4. Deadlines 5. Train the team (not one-and-done) 6. Automate and simplify; single source of truth 7. Rigor and accountability 8. Review results over time and adjust For each **stage**: number, name, entry/exit and conversion, actions and deliverables. Keep **categories** and the **call** simple. Define, publish, train, integrate into the CRM, review qualitatively, hold the line. Cadence (small team: collapse floors, do not skip the job): 1. Rep / manager 1:1 2. Leadership meeting 3. Manager forecast call 4. VP call if you have one 5. Head-of-revenue call Enablement: explain why the process exists, train, reinforce, share outcomes. Redundant meetings are how people stop telling the truth. ### Step 6: Keep the review focused Five lenses, then stop adding rituals: * **Cadence** with clear expectations, the right people, no duplicate forums * **Establish / enable / empower:** process, metrics, terms, R\&R, training, compliance * **Technology:** one source of truth; less manual rekey; guardrails * **Transparency:** marketing, sales, CS, implementation, finance—not a sales-only cave * **Questions that matter:** pipeline acceleration, deal execution, risk, strategy, predictability, outcomes ### Step 7: Use coverage gaps to guide action Do we have enough pipeline to hit the target? * Coverage to target * Out-quarter pipeline * Created vs converted this period * The **gap** * The **strategy** to close the gap Performance snapshot: forecast accuracy, variances, coverage, gap to close, attainment to date. ## Worked example (illustrative) Four AEs. \~45-day cycle. Hybrid: bottoms-up on deals > \$40k, rates on the rest. | Field | Fill | | ------------ | -------------------------------------------------------------------------------------------- | | Stages | 5 stages, each with entry/exit and a required next step dated \< 14 days | | Categories | Commit / best case / pipeline / omit. Omit means “not this period,” not “delete.” | | Call | Manager call = sum of commit they will defend in the 1:1, not CRM rollup of hope | | Hygiene | Friday 45 minutes: close date, amount, next step, category. Stale > 14 days drops category. | | Coverage | Target 3× on this motion; below 2.5× requires a named create plan, not a speech | | Cadence | Weekly AE 1:1 · Monday leadership 25 min · no second “forecast standup” | | Health score | Mutual plan exists? Economic buyer engaged? Next step owned? (yes/no, not a 40-factor model) | ## Copy: forecast design (fill) * Motion and why top-down, bottoms-up, or hybrid: * Stage list with entry/exit: * Categories and what the **call** means: * Hygiene rule (stale definition): * Cadence (who, when, duration): * Coverage target and gap-response: * Opportunity health fields: * Single source of truth: * Who may change a category after the call: ## Before you start * [ ] Stages have entry/exit, not only names. * [ ] Categories and the call are published and trained. * [ ] Hygiene has a clock; stale deals cannot sit in commit. * [ ] The model matches cycle time and volume. * [ ] Cadence has no duplicate forecast meeting. * [ ] Coverage questions include a gap strategy. * [ ] Bad news has a path that is earlier than the last week. ## Metrics | Metric | Diagnostic use | | ------------------------------------- | ------------------------------------------------------- | | Forecast accuracy / variance | How predictions compare with the outcomes that followed | | Coverage to target and gap | Whether the quarter is a math problem | | Stage conversion vs plan | Process vs heroics | | Hygiene (stale commit) | Trust in the CRM | | Time from known slip to surfaced slip | Reputation and the golden rules | Do not count forecast meetings held, or a decorated inspection tree, as success. ## Common mistakes * Calling a number with no reason. * Sitting on a slip until it is undeniable. * Stages without exit criteria. * Commit as a parking lot. * Enterprise bottoms-up on a self-serve motion. * Two sources of truth (spreadsheet vs CRM). * Hygiene as a once-a-quarter cleanup. * Treating pipeline coverage as a vanity multiple with no create plan. ## What to read next Pay still sits in [sales compensation](https://b2-b-playbook.mintlify.app/playbooks/09-operations-pipeline-and-measurement/sales-compensation). Whether next year is even possible is [GTM planning](https://b2-b-playbook.mintlify.app/playbooks/09-operations-pipeline-and-measurement/gtm-planning). The record underneath the call is [CRM data model](https://b2-b-playbook.mintlify.app/playbooks/09-operations-pipeline-and-measurement/crm-data-model). How meetings split so the call is not also pipe-gen and coaching is [sales operating cadence](https://b2-b-playbook.mintlify.app/playbooks/09-operations-pipeline-and-measurement/sales-operating-cadence). Deal quality sits in [account planning](../06-account-field-and-partner/account-planning.md) and the [pitch](../02-product-marketing/sales-enablement.md). After close, hygiene continues in [customer success](../08-lifecycle-and-customer-marketing/customer-success.md). Creating pipeline is not a forecast meeting: [channel strategy](../04-channels-and-distribution/channel-strategy.md) and [outbound](../05-outbound-and-prospecting/). ## Sources and evidence boundary This is an owner-maintained operating synthesis. It is not financial guidance, not a public-company forecasting control, and not a CRM implementation guide. The chain (hygiene → forecast → number → reputation), the curiosity order (buyers, journey, process, metrics, hygiene, forecast), experience optimization across discovery through lost, making conversion quantitative, golden rules, top-down vs bottoms-up, stage/category/call design, operating cadence, rigor lenses, and coverage questions are distilled from an Enterprise GTM School class on building a forecast model (pipeline and forecasting; instructor-attributed deck, class labeled #10, September 2023 cohort). That deck is a **method prompt**, not a source to copy. Personal bio, bit.ly samples, and tool shout-outs in the slides are not imported as vendor requirements. The deck points at Atrium’s *Sales Manager’s Guide to Forecasting* and at Gong mutual-action-plan examples as further reading; this repository does not abridge those. *** Copyright © 2026 Ivan Xu. All rights reserved. See the [copyright and reuse terms](https://github.com/weilun88313/B2B-Playbook/blob/main/LICENSE). Canonical source: [github.com/weilun88313/B2B-Playbook](https://github.com/weilun88313/B2B-Playbook)