CHURN IS DEAD

Your Renewal Forecast Needs Receipts

8 min read · Commercial

A renewal forecast can look precise and still be impossible to defend.

The account is 80 percent likely to renew. The CSM is confident. The health score is green. The renewal date is correct.

Then somebody asks a more useful question: what has the customer actually done or said that makes this forecast credible?

The room starts searching through call notes.

Confidence is not evidence. A percentage does not tell you whether the economic buyer supports the renewal, procurement has a path, the customer can explain the value, or a known risk has an owner. It compresses different kinds of uncertainty into one number and hides the reasoning that produced it.

Customer Success needs a forecast that can survive inspection.

Call it the Renewal Evidence Packet: a compact record of the current evidence, the contradictions, and the next customer decision. It does not promise perfect prediction. It makes the forecast explainable, reviewable, and easier to change when the facts change.

Forecast states are useful. They are not proof

Forecast categories create a shared language. Salesforce, for example, uses standard categories including Pipeline, Best Case, Commit, Omitted, and Closed. Its official guidance explains that opportunity stages can map to those categories and that users may override the default category.

That structure is operationally useful. It lets a team aggregate a book of business and discuss different levels of confidence.

But a category is still a classification. It does not, by itself, show why a renewal belongs there.

GitLab's public renewals handbook shows what a more detailed operating record can contain. Its renewal fields include next steps, notes, risk category, forecast health, available-to-renew revenue, forecast amount, active mitigation, and best-case, most-likely, and commit scenarios. Its process also distinguishes actionable churn or contraction risk from outcomes believed to be certain.

This is not a universal template. It is evidence that a real renewal operation needs more than a colour or percentage. Forecast, risk, intervention, ownership, and next action are separate pieces of information.

The Renewal Evidence Packet brings those pieces into one reviewable record.

The six evidence domains

This framework is a Churn Is Dead proposal, not an industry benchmark. Adapt the fields to your segment, contract model, and sales process.

A forecast is review-ready when the packet addresses six domains.

1. Contract reality

Record the renewal date, notice period, available-to-renew scope, current commercial configuration, and any known procurement or legal steps.

Do not confuse a correct date in the CRM with a confirmed buying process. A contract may auto-renew while the customer intends to reduce scope. A commercial conversation may be positive while a notice deadline is approaching. The packet should show both the contractual mechanism and the expected customer action.

2. Customer authority

Name the people involved in the decision and the evidence of their roles.

Who owns the budget? Who can approve the renewal? Who can block it? Who will operate the product after renewal? A friendly champion is valuable, but enthusiasm from one contact is not the same as authority across the decision.

Record the most recent direct interaction with each critical role. If the economic buyer has not engaged, write unverified. Do not convert absence into optimism.

3. Value proof

State the outcome the customer expected, the evidence that the outcome occurred, and whether the customer has acknowledged it.

Usage can support the case, but usage is not automatically value. A value claim becomes stronger when it connects product behaviour to an outcome the customer recognises. Keep the evidence specific: an approved success-plan outcome, a customer-confirmed operational change, or a business review record that the customer accepted.

If the value case exists only in your internal deck, classify it as an internal hypothesis.

4. Decision path

Record the remaining steps between today and the commercial decision.

That might include a value review, security approval, budget submission, quote, procurement review, legal review, signature, or purchase order. Give each step an owner and expected date. Mark whether the date came from the customer, your team, or an assumption.

This turns “renewal in progress” into a sequence that can be tested.

5. Risk and response

Describe the risk as a specific uncertain event, not a mood.

“Low engagement” is an observation. “The new sponsor may remove the product from next quarter's budget because the original use case has no executive owner” is a decision risk.

For each material risk, record its source, potential effect, response, owner, and next evidence date. GitLab's public health guidance illustrates the need for this separation: it treats lack of information, loss of a sponsor, poor adoption, and known retention threats as different conditions, and asks account owners to log progress and next steps more frequently for at-risk customers.

6. Contrary evidence

Give the forecast a place for facts that weaken it.

Competitive evaluation, seat reductions, a silent budget owner, unresolved support history, a delayed security review, or an unconfirmed success metric should not disappear because the overall sentiment is green.

The strongest packet is not the one with the most positive evidence. It is the one that makes disagreement visible.

Evidence needs a quality label

Not all evidence deserves the same weight. Give each item one of four labels:

Customer-confirmed: stated or approved by someone with relevant decision authority.
Observed: behaviour or system data directly recorded by your organisation.
Internally asserted: a judgement from the account team that the customer has not confirmed.
Missing: required information that has not been obtained.

Then add a source date and an expiry date.

Expiry is deliberately contextual. A procurement path may remain stable for months. An executive's intent can change after a reorganisation. Product usage may need a shorter review cycle. Set the interval your business can defend, then make stale evidence visible.

This borrows a principle from assurance, not a rule from government. HM Treasury's Orange Book defines assurance as confidence derived from objective information about activities, controls, compliance, and the credibility of information used for decisions. It also warns that confidence falls when the integrity of the information or underlying process is uncertain.

A renewal forecast deserves the same discipline: confidence should rise because the evidence is current and inspectable, not because the field is green.

Replace percentages with promotion rules

Keep your CRM categories if the wider business relies on them. Change the rule for entering each state.

Here is a starting model:

Unverified: the basic contract facts exist, but customer authority, value, or decision path is missing.
Evidence building: the decision roles and value case are being tested; material unknowns remain.
Credible path: the customer has confirmed the decision path and value case, with no unresolved contradiction that could materially change scope or timing.
Commit: the authorised commercial path is active, required approvals are understood, and remaining steps have owners and dates.
Closed: the commercial outcome is recorded.

These are proposals, not benchmark definitions. Tighten them for your business. The important move is to require evidence for promotion and to demote the forecast when evidence expires or a contradiction appears.

An account does not earn Commit because the CSM feels better this week. It earns Commit because the packet can show what changed.

Run the review as an audit, not a performance

Choose a small set of upcoming or high-impact renewals. The owner presents only three things:

1. the current forecast state;

2. the newest confirming or contrary evidence;

3. the next decision or evidence required.

Reviewers do not ask for a longer narrative. They test the packet:

Which item proves the customer has authority and intent?
What evidence could overturn this forecast?
Which item is stale?
Is the next step owned by the customer or only by us?
What must become true before this account is promoted?

End with one recorded outcome: retain the state, promote it, demote it, or mark it unverified. Record the reason and the next review date.

The review is working when two people can inspect the same packet and understand why the account sits in its current state, even if they disagree with the judgement.

Start with one account

Do not rebuild the CRM first.

Take one renewal currently labelled Commit. Build the six-domain packet using the evidence you already have. Mark every item customer-confirmed, observed, internally asserted, or missing. Add contrary evidence. Expire anything too old to guide the next decision.

Then ask whether the account still belongs in Commit.

The purpose is not to make the forecast pessimistic. It is to make it credible.

Download the Renewal Evidence Packet

Sources and methodology

This issue proposes a Churn Is Dead operating model. It does not claim that the six domains, evidence labels, state definitions, or expiry intervals are universal benchmarks. The model draws on the category structure in Salesforce's official forecasting guidance, the separation of forecast, risk, mitigation, ownership, and next steps in GitLab's public renewals methodology, the treatment and review of uncertainty in GitLab's customer health guidance, and the relationship between objective information and decision confidence in HM Treasury's Orange Book. Sources were reviewed on 24 August 2026.

By Kuber Sethi · All issues · Subscribe