Ask a CRO at a B2B service company how much pipeline they have, and you get a number in seconds. Ask how much existing revenue is at risk this quarter, and you usually get a pause, a caveat, and an anecdote about one large account.
That asymmetry is expensive. In logistics, IT managed services, professional services, manufacturing and distribution, and financial services, the bulk of next year’s revenue is this year’s customers. Harvard Business Review puts the cost of acquiring a new customer at five to 25 times the cost of retaining an existing one, and cites Bain research showing that a five-point improvement in retention can lift profits by 25 to 95 percent. Yet most revenue teams instrument acquisition obsessively and measure retention risk by feel.
This post lays out a practical way to put a number on revenue at risk — one you can defend in a board meeting and act on in an account plan.
Why “revenue at risk” beats “churn rate” as a leading metric
Churn rate is a lagging indicator. By the time an account appears in it, the decision was made quarters ago. Bain has long argued that retaining customers is the real challenge, precisely because defection builds quietly while reported satisfaction looks fine.
Revenue at risk flips the lens. Instead of counting lost accounts after the fact, it asks: of the recurring and repeat revenue in the current portfolio, how much sits in accounts showing credible warning signs right now? That makes it a leading indicator a CRO can actually manage against — staff against, escalate against, and forecast against.
A working definition
Revenue at risk = the annualized revenue of every account whose health score sits below your defined risk threshold, weighted by severity.
Three components make that definition operational:
- Account-level annualized revenue. Contract value where contracts exist; trailing twelve-month revenue where the relationship is repeat-business rather than contractual, which is common in logistics and distribution.
- A health score built from more than one source. A single signal — usually a survey — produces false confidence. The score has to combine what customers say, what they do, and what your delivery data shows.
- A severity weighting. An account two points below threshold is not the same as an account in free fall. A simple three-tier weighting (watch, at risk, critical) keeps the headline number honest.
The five signal categories that feed the number
The hard part is not the arithmetic. It is that the inputs live in different systems: survey results in a spreadsheet, complaints in a ticketing tool, commercial history in the CRM, delivery performance in operational systems, and the account manager’s read in their head. McKinsey’s research on B2B buying shows customers now interact with suppliers across ten or more channels — which means the evidence of risk is more fragmented than ever.
A usable health score pulls those fragments into five categories:
- Satisfaction — structured feedback from the people who experience your service, not just the contact who signs the renewal.
- Engagement — are stakeholders showing up to reviews, responding, escalating early? Silence is a signal.
- Commercial — order frequency, scope changes, payment behaviour, pricing pressure.
- Delivery — SLA performance, complaint volume, rework, on-time rates. In service businesses this is where churn usually starts.
- Expansion — whether the account is growing with you. Flat or shrinking share of wallet is risk wearing a polite face.
This is exactly the model behind EvaluationsHub’s multi-source account scoring: each account gets one transparent score composed from those five signal categories, so “revenue at risk” stops being a debate and becomes a query.
From a number to a board-ready metric
Once every account carries a score, the portfolio rolls up cleanly. The reporting set most CROs land on looks like this:
- Total revenue at risk — the headline figure, in currency, not percentages. Boards respond to euros and dollars.
- Risk concentration — what share of at-risk revenue sits in the top ten accounts. Service portfolios are usually concentrated, and one critical strategic account can outweigh twenty small ones.
- Risk by signal category — if Delivery is driving the number, the fix belongs to operations; if Engagement is, it belongs to account management. The category split tells you who owns the recovery.
- Movement — accounts entering and leaving the at-risk pool each month. A flat headline number can hide a churning pool underneath.
Making the number actionable: recovery, reviews, relationships
A risk number nobody acts on is just decoration. Gartner’s guidance on retention stresses identifying at-risk accounts early enough to intervene — which means the measurement system needs a response system attached.
Three mechanisms close that loop:
Structured recovery. When an account crosses the risk threshold, it should trigger a defined corrective workflow, not an ad hoc email thread. EvaluationsHub’s CAPA recovery playbooks bring the corrective-and-preventive-action discipline familiar from quality management into account management: root cause, owner, actions, deadline, and a rescored outcome at the end.
Reviews with teeth. Quarterly business reviews work as a revenue tool when they open with the health score and close with committed actions. Review management built around the score turns QBRs from status theatre into the operating rhythm of the risk number.
Stakeholder coverage. Risk often hides in the contacts you do not talk to. Stakeholder mapping across the buying group — economic buyer, daily users, procurement — exposes single-threaded accounts, which are at-risk accounts that have not announced themselves yet.
Put the number where revenue decisions happen
The final failure mode is building a beautiful risk model that lives outside the tools your team works in. Account health belongs on the account record. EvaluationsHub ships a native Salesforce integration built on custom objects, so scores, signals, and recovery actions sit where AEs and account managers already operate — and an AWS AppFlow integration carries delivery and commercial data in from operational systems without manual exports.
Looking ahead, Eva AI auto-trigger — coming soon — will start recovery workflows automatically when signal patterns deteriorate, shortening the gap between detection and response even further.
Start with ten accounts
You do not need a transformation program to get a credible revenue-at-risk number. You need your most important accounts scored from real signals for one quarter.
That is what the EvaluationsHub 30-day pilot is for: €30 per month for ten accounts, done-for-you setup, fully refundable, cancel anytime — and customer surveys go out only with your explicit approval. Prefer to look around first? Create a free account, no card required, and see how multi-source scoring turns “I think that account is fine” into a number you can manage.