Five signals, one number, and a portfolio that does not behave like one thing

Most account health scores are built the same way: pick five or six inputs, average them, and call the result a health score. It is a reasonable starting point. It is also where most revenue teams stop, and stopping there is the mistake.

A logistics account with a spotless on-time-delivery record and a quiet, unresponsive procurement contact is not the same risk profile as an IT managed services account with frequent tickets but a champion who answers every email. Averaging five signal categories — Satisfaction, Engagement, Commercial, Delivery, and Expansion — into one number treats both accounts as equally healthy or equally at risk. They are not. The weighting is where the score either earns its keep or quietly misleads the room.

Why a blended score hides the renewal that actually churns

Equal weighting is a default, not a decision. It assumes every signal category matters the same amount, to every account, at every stage of the relationship. In practice, the categories that predict churn shift depending on account type, contract stage, and industry — and a flat average smooths over exactly the divergence a revenue leader needs to see.

Health-scoring research on this problem is consistent: composite scores that blend signal categories without differentiated weighting tend to miss the accounts that churn quietly, because a strong showing in one category (say, Satisfaction) can offset a genuine deterioration in another (say, Commercial or Delivery) until the net number looks fine right up until the renewal conversation goes badly. The fix is not more data. It is weighting the data you already have according to what actually predicts renewal in your book of business, then testing that weighting against outcomes rather than assuming it.

What the weighting research says

Customer-success teams that have published their weighting logic converge on one principle: weight signals by predictive power, not by how easy they are to collect. One widely cited scoring approach weights product engagement and activity heavily in a composite score precisely because behavioral signals move before sentiment signals do — an account that goes quiet on Engagement is a leading indicator, while a dip in a satisfaction survey often arrives after the decision to leave has already been made, as Gainsight’s health-scoring research lays out.

That specific split does not transfer cleanly to non-SaaS service businesses, because there is no login activity to weight. But the underlying principle does transfer: whichever of your five categories moves earliest and most reliably ahead of a lost renewal should carry the most weight in the composite, and that category is rarely the same one twice across logistics, IT services, professional services, manufacturing, and financial services.

The stakes for getting this right are not abstract. Bain & Company’s research on retention economics found that increasing customer retention rates by just 5 percent can increase profits by 25 to 95 percent, depending on the industry. A weighting scheme that surfaces risk two quarters earlier is not a scoring refinement — it is the difference between a save play and a write-off.

Why the same five categories weight differently by industry

The five signal categories — Satisfaction, Engagement, Commercial, Delivery, Expansion — are constant. What changes is which one moves first when an account is heading toward churn or toward expansion.

Logistics

Delivery signals (on-time performance, claims, exception rates) tend to be the leading indicator. A logistics account rarely leaves over a single missed satisfaction survey; it leaves after a pattern of operational misses that Commercial and Engagement metrics won’t show until the RFP is already out. Delivery should carry the heaviest weight, with Engagement close behind as a check on whether operational issues are actually reaching the account owner. See how this plays out for logistics accounts.

IT managed services and MSPs

Here, Commercial signals (ticket-driven cost overruns, scope creep, contract utilization) often surface risk before Satisfaction does, because technical teams tend to under-report frustration until a renewal is imminent. Engagement — specifically, whether the relationship is single-threaded through one IT contact — deserves real weight too. More in our IT services playbook.

Professional services

Satisfaction and Engagement dominate, because the product is the relationship. A partner-level account with declining meeting cadence or a stakeholder who has gone quiet is a stronger churn signal than any commercial metric, since professional services renewals are frequently relationship decisions made before the fee conversation starts. Detail here: professional services.

Manufacturing and distribution

Delivery and Commercial signals lead, but Expansion signals (new plant lines, new SKUs, volume shifts) are unusually predictive of upside that gets missed when the score is weighted purely toward risk. This is a portfolio where the same weighting scheme should serve two purposes: flag delivery risk early and surface whitespace before a competitor’s account manager does. See manufacturing and distribution.

Financial services and insurance

Commercial and Satisfaction carry the most weight, but the lag between a warning signal and a visible churn event is longer and the volume tends to be block-based — a single relationship manager change can put an entire book of business at risk at once. Weighting needs to account for concentration risk, not just average sentiment. Detail in our financial services resources.

Turning weighting from a spreadsheet exercise into a workflow

None of this matters if the weighted score lives in a slide deck that gets updated once a quarter. Weighting only earns its value when it is calculated continuously across every signal source your teams already generate: CRM notes, delivery and ticketing data, survey responses, and commercial terms, combined through multi-source account scoring rather than manual assembly.

For teams running on Salesforce, the practical version of this is a weighted score living on the Account record itself, built from custom objects that pull in Delivery, Commercial, Engagement, Satisfaction, and Expansion data through native Salesforce integration, so account owners see the same weighted number CROs see in the portfolio roll-up. Data flowing in from adjacent systems — delivery platforms, ticketing tools, ERP — can be connected through AWS AppFlow rather than built as one-off integrations.

Once a weighted score flags an account moving the wrong direction, the response should be structured, not improvised. That means a defined CAPA recovery playbook triggered at a threshold, not a hallway conversation about whether someone should check in. And because weighting reveals risk earlier, it also creates room to have that conversation with the customer directly — sharing the relevant health signals through a customer portal rather than surprising them with an escalation.

Start with the weighting your portfolio actually needs

Equal weighting across five signal categories is a placeholder, not a strategy. The version that actually predicts renewals is calibrated to what moves first in your specific portfolio, tested against real outcomes, and revisited as your book of business changes. Getting there does not require a data science team on day one. It requires a platform that can pull Satisfaction, Engagement, Commercial, Delivery, and Expansion signals into one place and let you adjust the weighting as evidence accumulates — rather than a static formula nobody revisits after the first quarter.

If you want to see what a weighted, multi-source account health score looks like against your own portfolio, start a 30-day pilot — €30 a month for 10 accounts, fully refundable, cancel anytime, with setup done for you. Or create a free account to explore the scoring model yourself, no card required.