A forecast is a decision system, not a target with a date
A missed forecast affects more than a sales meeting. Leadership may hire too early, delay a necessary investment, promise delivery capacity that is not needed or cut marketing because revenue expected this month moved into the next quarter. Repeated misses weaken trust even when the eventual sales arrive.
The owner decision is not how to force the team to commit harder. It is which evidence makes future revenue credible enough to plan against. A target states what the business wants. A pipeline records open opportunities. A forecast estimates what is likely to become revenue in a defined period. Treating those three numbers as interchangeable creates optimism disguised as analysis.
This problem is different from rebuilding an empty sales pipeline or creating a predictable B2B SaaS pipeline. Here, opportunities already exist; the problem is that their probability, value or timing cannot support a reliable leadership decision.
Five errors usually sit behind the miss
Salesforce distinguishes pipeline from forecast: the forecast is the subset expected to close in a particular period, based on deal volume, likelihood and when money will be received.
Forecasts usually fail when stage names mean different things to different people; close dates reflect the seller's hope rather than the buyer's process; stale opportunities remain open; headline contract values ignore scope, discount or payment timing; or the model assumes sales and delivery capacity that does not exist.
UK government financial-model guidance recommends transparent, defensible assumptions grounded in real-world data and tested with sensitivity analysis. Every important number should have an owner, source and reason it can change.
The Forecast Reliability Gate
Pass every opportunity through five owner-level checks before it contributes to the base forecast. The gate does not require a particular CRM. It requires shared definitions and evidence that can be inspected.
Is this a real commercial problem?
Confirm the buyer, need, suitability, authority path and reason to act. Interest alone is not a qualified opportunity.
What did the buyer do?
Define stages by observable progress: discovery completed, requirements confirmed, proposal reviewed or approval secured.
Why this period?
Anchor the close window to a buyer event, next action and normal cycle—not the end of the sales quarter.
What revenue can actually land?
Use the likely approved scope, price, probability and payment timing. Keep pipeline value separate from collected cash.
Can the business win and deliver?
Include sales coverage, onboarding, stock or service capacity when those constraints can move the revenue date.
When an opportunity fails a check, do not delete it automatically. Move it to the appropriate evidence state: excluded, lower case, later period or base case. The important change is that uncertainty becomes visible instead of remaining hidden inside one total.
Turn one optimistic total into three explainable cases
Hypothetical example—not a client result or benchmark
A service business reports £500,000 of open pipeline for the quarter. After applying consistent fit and stage evidence, £220,000 remains. Deal age and buyer-owned dates move £90,000 into a later period. The remaining £130,000 becomes the evidence pool—not the forecast.
| Case | Included revenue | Evidence boundary | Leadership response |
|---|---|---|---|
| Lower · £65,000 | Only approved or buyer-committed work | No dependency on unresolved scope | Protect cash and delay optional hiring |
| Base · £95,000 | Lower case plus opportunities with verified next events | Normal cycle and capacity assumptions | Plan resources against the most defensible case |
| Upper · £125,000 | Base plus two time-sensitive approvals | Named triggers must occur by a review date | Prepare capacity without treating it as committed |
The values only demonstrate the method. Replace them with reconciled CRM, finance and delivery evidence from the business.
Review the range against actual invoiced or collected outcomes, not only closed-won labels. Record whether variance came from qualification, stage progression, timing, price, delivery or data quality. That error record is the input to the next forecast.
Choose the repair from the pattern of error
| Observed pattern | Likely failure | Owner move | Do not do |
|---|---|---|---|
| Forecast is repeatedly high | Loose stages, stale deals or optimistic probabilities | Require buyer actions and recalibrate from completed cohorts | Apply one arbitrary haircut to every opportunity |
| Value is close; timing is wrong | Seller-owned close dates or payment lag | Track buyer milestones, cycle length and collection timing separately | Pull every deal into the current quarter |
| Marketing reports growth; forecast stays weak | More leads without fit or progression | Compare sources by qualified, accepted and won outcomes | Buy more of the cheapest lead |
| One large deal determines the quarter | Concentration hidden by the total | Show the base case with and without that opportunity | Present the upper case as the plan |
| Wins arrive but delivery moves revenue | Capacity omitted from the forecast | Connect sales commitment to onboarding and fulfilment dates | Count signed value as immediate cash |
Connect marketing evidence to the revenue forecast
A forecast problem can begin before an opportunity exists. If marketing attracts the wrong buyer or counts every enquiry equally, leadership may keep funding volume that rarely becomes forecastable revenue.
Google Analytics recommends distinct events for generated, qualified, disqualified, working, converted and unconverted leads. Google Ads likewise separates funnel stages and qualified from converted leads. The owner principle is simple: keep commercial stages distinct, then let the responsible specialist implement the measurement lawfully and accurately.
Reconcile suitable enquiries, opportunities, stage movement, expected value, buyer events, loss reasons and won value by source each week. Use the CRM readiness guide if records are unreliable. Review conversion tracking when marketing cannot see qualified outcomes, and landing pages when the promise attracts poor fit.
Review ThomPerformance's case evidence, evidence standards, growth services and direct operating model. The commercial question is which demand can become revenue—not which campaign created a form.
Sources and evidence notes
Underlying sources and current search results were reviewed on 7 November 2026. No search volume, universal accuracy threshold or client result is claimed. The Forecast Reliability Gate, three-case worked example and error-pattern decision matrix are original ThomPerformance analysis.
- Salesforce: Sales Forecasting—A Complete Guide to Creating Accurate Forecasts (13 October 2026)
- Salesforce Trailhead: Forecast with Precision
- UK Government: Financial Model Essentials (2 December 2025)
- Google Analytics Help: Recommended lead-generation events
- Google Ads Help: Offline conversion imports FAQs
Frequently asked questions
What is the difference between a sales pipeline and a sales forecast?
The pipeline contains open opportunities at every stage. The forecast is the smaller set expected to close in a defined period. A large pipeline can therefore produce a modest forecast.
How accurate should a sales forecast be?
There is no universal acceptable percentage. Track forecast against actual by period, segment and stage, then improve the assumptions that repeatedly create the error.
Should an owner use one forecast number?
Use a base case for operating decisions, plus lower and upper cases with explicit triggers. Each case needs deal-level evidence, timing assumptions and a defined response.
Can AI fix an inaccurate sales forecast?
AI can flag stale records, unusual deal ages and recurring variance. It cannot repair undefined stages, missing buyer evidence or invented close dates. Improve the commercial process first.
How should marketing data support the forecast?
Connect suitable enquiries and opportunities to their sources where lawful and practical. Return qualification, progression, loss and won value so leaders can locate the first forecast gap.
Make the uncertainty useful
A trustworthy forecast does not pretend uncertainty has disappeared. It separates raw pipeline from buyer evidence, makes timing and value assumptions inspectable, and gives leadership a lower, base and upper case with clear responses. Each miss then improves the next decision instead of becoming another argument about confidence.
Which part of your current forecast depends most on hope: buyer fit, stage evidence, timing, value or capacity?
