Skip to content
Centriu
Back to Resources
Playbook · 22 min read

Reliable sales forecast in 90 days

Why do 78% of sales forecasts miss by more than 30%? Because they're built on gut feel, not method. This playbook shows how to build a forecast your partner trusts — in three 30-day sprints.

The cost of a bad forecast

+34%
average error in forecasts built on gut feel
3,2x
more likely to hire wrong when the forecast deviates >25%
±8%
typical error in agencies with a mature forecasting method

30dSprint 1 — Define the pipeline in objective stages

The first source of forecast error is poorly defined stages. "In negotiation" means different things to every rep. Without an objective criterion, the forecast is fiction.

Use the MEDDPICC framework (or a simplified version)

StageObjective criterionProbability
New leadEntered the funnel, not yet qualified5%
QualifiedConfirmed problem + budget + authority20%
Meeting scheduledCalendar confirmed with the decision-maker (not the assistant)35%
Proposal sentDocument delivered + positive verbal alignment55%
NegotiationDiscussing final terms, no longer "whether" it will close75%
Verbal commitmentClient said "let's move forward" — only the contract is left90%
Golden ruleIf the rep can't tell you why the lead is in a given stage (with evidence: email, meeting notes, proposal link), the lead doesn't belong there. Move it back a stage.

Sprint 1 tasks

Audit every deal currently open in the CRM
Re-categorize each one by the objective criteria above
Train the team (1h) on the new criteria
Lock manual editing of probability (it comes from the stage)
Set a weekly pipeline review cadence (60 min, Tuesday)

60dSprint 2 — Calculate weighted forecast & measure real conversion

A clean pipeline is the foundation. A reliable forecast needs math.

The weighted forecast formula

// For each open deal:
weighted_value = valor_deal × probabilidade_estagio
// Forecast for the period:
forecast = Σ weighted_value (deals closing <= period)

Practical example — quarterly forecast

DealValueStageProb.Weighted
Client AR$ 120kVerbal90%R$ 108k
Client BR$ 80kNegotiation75%R$ 60k
Client CR$ 200kProposal55%R$ 110k
Client DR$ 60kMeeting35%R$ 21k
Client ER$ 40kQualified20%R$ 8k
Quarterly weighted forecastR$ 307k

But careful: adjust by your funnel's REAL conversion

The probability in the table above is a reference. The truth is in your last 12 months. Calculate:

conversão_real_por_estágio = deals_fechados_que_passaram_por_X / total_que_passaram_por_X

If the table says "Proposal = 55%" but in your history only 38% of deals that reached proposal closed, your real probability is 38%. Recalibrate.

90dSprint 3 — Review ritual and early risk detection

A living forecast, not a snapshot. Without a review ritual, it rots in 2 weeks.

Canonical cadence

Daily (5 min)
Sales rep
Update the stage of deals that moved yesterday
Weekly (60 min)
Team + leader
Pipeline review: every deal above R$ 30k, real status
Biweekly (45 min)
Leader + partner
Reconcile weighted forecast vs. quarterly target
Monthly (90 min)
Leadership
Lookback: forecast vs. actuals, adjust probabilities
Quarterly (3h)
Leadership + finance
Full calibration: real conversion, average deal size, sales cycle

Early warning signs (enable in the dashboard)

Deal stalled > 14 days
No new activity (email, call, doc). A sign it has cooled off.
Skips a stage
Went from "qualified" straight to "negotiation" with no proposal. Toxic forecast.
Probability rose without an event
The rep "feels" it will close. Without a concrete event, knock 20% off.
Decision-maker didn't answer the last call
Last proposal without feedback for 7+ days. Re-engage or write off.

⚠The 7 most expensive mistakes beginners make

❌ Subjective probability per deal
✅ Lock probability to the stage. No manual editing.
❌ Forecast done by the CEO alone
✅ The forecast must be built by the reps and validated by the leader.
❌ Ignoring deals that come back
✅ A returning lead has a different profile. Create a "Reactivated" stage with its own conversion.
❌ Forgetting churn
✅ Net revenue forecast = new MRR − projected churn.
❌ Not measuring velocity
✅ Velocity = (deals × value × conversion) / avg_cycle. It's the funnel's health metric.
❌ Mixing high and low ticket
✅ Separate pipeline for deals >R$ 100k. Different behavior, different cycle.
❌ Not closing the loop with finance
✅ The forecast must feed into cash flow, not live isolated in the CRM.

How Centriu supports this playbook

A trustworthy forecast only exists when sales, delivery and finance data sit in the same place and are refreshed on the same cadence. The systems that help with that:

Orbit

Sales and revenue — every opportunity with an owner, a deadline and a next step, in a funnel with stages and a follow-up cadence.

Gauge

Managerial finance, from the reconciled statement to margin per client.

Helix

Risk intelligence — alerts with evidence, severity and a stated confidence level. No action goes out without human approval.

Helix

The account portfolio as living organisms: health per account, a portfolio-level reading, and action executed only after a person approves it.

Want to run a lookback of your last quarter?

In 1h we rebuild your forecast with the method and show you where the error lives.

Você também pode gostar

Prompt

Diagnóstico de CRM e RevOps

Diagnóstico executivo de CRM/RevOps: gargalos, problemas de dados/processo/adoção, oportunidades de automação, riscos, quick wins e roadmaps de 30/90 dias.

Abrir Template

Planilha LTV / CAC

Calcule unit economics: LTV, CAC e payback.

Abrir Coleção

Coleção: head comercial e pipeline

Estruturar pipeline e previsão de receita.

Abrir Prompt

Jornada do Cliente em Estrutura de CRM

Mapeia a jornada como funil contínuo de CRM: etapas, status, campos obrigatórios, gatilhos de movimentação, automações, responsáveis, indicadores e riscos por etapa.

Abrir