New-business pipeline gets a probability attached to every dollar. A deal is commit, best case, or pipeline, and everyone in the forecast call agrees on what those words mean. Revenue at risk on the existing book rarely gets the same treatment. It shows up as a shrugged percentage in a QBR deck, or a list of “accounts to watch” with no number attached at all.
The Forecast Has Two Halves, and Only One Is Disciplined
Most CROs at non-SaaS service companies — logistics, IT managed services, professional services, manufacturing and distribution, financial services and insurance — run a tight new-business forecast. Stage-based probabilities, weekly deal inspection, a defined bar for what counts as commit.
The renewal and expansion side of the same book gets none of that structure. Gartner’s State of Sales Operations research found that fewer than half of sales leaders and sellers have high confidence in their organization’s forecasting accuracy, and traced much of that gap to inconsistent data discipline rather than bad judgment (Gartner). That same discipline gap is worse on the retention side, because most revenue teams have never built the underlying data model to support it.
That is the actual problem behind “revenue at risk is a gut feel.” It is not that CROs lack conviction about which accounts are shaky. It is that there is no structured, repeatable way to turn that conviction into a probability-weighted number the way pipeline stages do for new business.
Why a Single Revenue-at-Risk Number Misleads
A blended revenue-at-risk figure — “12% of the book is at risk this quarter” — collapses two very different situations into one line. An account with a disengaged economic buyer and a lapsed contract renewal date is a different kind of risk than an account with a single delayed delivery milestone and an otherwise engaged buying committee. Both can land in the same at-risk bucket if the only input is a support ticket count or a CSM’s flag.
McKinsey’s research on B2B growth performance found a sharp split between market leaders and laggards: 60 percent of self-identified leaders reported double-digit revenue growth, against just 21 percent of laggards, and leaders were far more likely to report improved sales effectiveness across the board (McKinsey). The same research points to inconsistent information across teams and inability to reach a knowledgeable contact as leading reasons B2B buyers switch suppliers — both are signals that live in delivery and engagement data, not in a single retention percentage.
The fix is not a better guess. It is treating revenue at risk the way new-business pipeline is already treated: as a set of dollar amounts, each with its own probability, rolled up into commit, best-case, and at-risk categories instead of one blended figure.
Borrowing Forecast Discipline for the Back Half of the Book
What Makes Pipeline Forecasting Work
Stage-based forecasting works because every deal carries structured inputs — stage, close date, engagement level, competitive status — that roll up into a probability. Nobody argues about whether a deal is “commit” from memory; the CRM enforces the criteria.
Applying the Same Logic to Renewal Risk
The same structure works for the existing book once each account carries its own structured inputs. That means scoring every account across five signal categories — Satisfaction, Engagement, Commercial, Delivery, and Expansion — rather than relying on whichever category happens to generate the loudest complaint. An account with strong Commercial and Engagement scores but a slipping Delivery score is not the same risk profile as one where all five categories are trending down, even if both show up as “red” in a simple traffic-light view.
Multi-source account scoring makes this possible by pulling from internal systems — CRM, delivery and ticketing tools, finance — alongside external and qualitative signals like survey responses and QBR notes, instead of relying on one data source to carry the whole judgment.
From Score to Dollar Figure: A Worked Structure
Once every account has a current five-signal score, the roll-up looks like this:
- Committed revenue: accounts with strong scores across Satisfaction, Engagement, and Commercial, contract terms in good standing, and no open CAPA (corrective and preventive action) plan.
- At-risk, plan in motion: accounts with one or more declining signal categories that already have a structured recovery playbook underway, with a named owner and a review date.
- At-risk, uncovered: accounts with declining signals and no active recovery plan — the category that should draw the most immediate attention, because it is the one still driven by chance rather than process.
This is the same three-tier logic a forecast call already applies to new pipeline. The difference is that the probability driving each tier comes from a structured account health score instead of a rep’s stage judgment, which makes it auditable and reproducible across a portfolio a single CRO could never review account by account.
Where the Model Breaks Down Without the Right Inputs
A probability-weighted revenue-at-risk model is only as good as the data feeding it, and non-SaaS service companies have a structural disadvantage here: there is no product-usage telemetry to lean on the way a SaaS vendor would. The signal instead has to come from Salesforce data synced through native integration and custom objects, delivery and ticketing systems connected through tools like AWS AppFlow, and structured qualitative input from account managers and QBRs.
Stakeholder mapping matters at this stage too. Bain’s research on customer loyalty found that a five-percentage-point increase in retention can increase profits by 25 to 95 percent, depending on the industry (Bain) — and a companion analysis in Harvard Business Review makes the same case for treating retained revenue as a compounding asset worth measuring with real rigor, not an afterthought behind new logos (Harvard Business Review). A revenue-at-risk model that only tracks one contact per account will miss the stakeholder churn that precedes almost every surprise loss, because the person who disengages first is rarely the one filling out the satisfaction survey.
Operationalizing the Model
None of this holds up as a one-time exercise. The categories need a cadence:
- A monthly revenue-at-risk roll-up, reviewed the same way a forecast call reviews new pipeline, broken into the three tiers above.
- CAPA recovery playbooks assigned automatically to any account that drops into the uncovered tier, so accounts do not sit unmanaged between QBRs.
- QBRs built around the account’s own five-signal trend line, not a generic template, so the review connects directly to the number on the revenue-at-risk report.
- A customer portal that gives the account’s own stakeholders visibility into delivery and engagement status, closing some of the information-consistency gap that McKinsey’s research flags as a top reason buyers switch.
EvaluationsHub is also building Eva AI auto-trigger (coming soon), designed to move an account into a recovery playbook automatically the moment its signal trend crosses a defined threshold, rather than waiting for the next scheduled review to notice.
The Payoff Is a Forecast Leadership Can Actually Defend
A CRO who can walk into a board meeting with committed, in-motion, and uncovered revenue-at-risk figures — each backed by a specific account list and a specific playbook status — is answering a fundamentally different question than one presenting a single blended percentage. The first is a forecast. The second is a guess with a decimal point.
Getting there requires the same ingredients that make new-business forecasting trustworthy: structured inputs, a consistent scoring model across the whole portfolio, and a system that surfaces the accounts that need a recovery plan before they need a save. Multi-source account scoring is the foundation; CAPA recovery playbooks turn a declining score into action; native Salesforce integration keeps it where revenue teams already work; and the full platform ties all five signal categories to one number leadership can trust.
See how a probability-weighted revenue-at-risk model would look against your own portfolio. Book a demo to walk through it with your account data, or start exploring in the free sandbox — no card required.