What Is a Client Health Score and How to Calculate It
Your top client hasn't responded in 47 days. Your spreadsheet says everything is fine. A client health score would have told you the truth six weeks ago.

01. Why Client Health Scores Matter
Agency revenue typically concentrates among a few top clients — often 50-70% from three or four relationships. When one of those relationships ends unexpectedly, the cash flow impact is immediate and the recovery time is measured in months, not weeks. The warning signs for client departure are almost always present before the client leaves. Fewer meetings. Slower response times. A reduction in project requests. Invoice payments arriving later than usual. None of these signals are visible in a spreadsheet that tracks revenue but not relationship activity.
"We lost a $4,200/month client without any warning. When I looked back at the email thread, the last message from them was 47 days before they sent the notice. We just weren't tracking the silence."
— Agency founder, 8-person creative studio
02. Health Score vs NPS: The Key Distinction
Net Promoter Score measures declared sentiment — what clients say about your business when asked directly. It is a periodic survey that captures how a client feels at a specific moment in time. A client health score measures observed behavior — what is actually happening in the relationship based on data already in your CRM. Task activity, communication recency, follow-up status, billing health. These are not what clients say; they are what clients do. Both are useful for different purposes. NPS is appropriate for periodic audits of stated satisfaction — once or twice a year. Health scores are appropriate for daily operational monitoring — catching the relationship declining before the client articulates their dissatisfaction. By the time a client gives you a low NPS score, they are usually close to leaving. By the time their health score has been declining for six weeks, you still have time to intervene.
03. The Four-Component Formula
The formula has four components that combine to produce a score between 0 and 100: **Component 1 — Base Score (50):** Every client relationship starts at 50 — the neutral midpoint. Adjustments from the other three components move the score above or below this baseline. **Component 2 — Activity Signals:**
- Active tasks assigned: +20
- Active project in progress: +10
- No active tasks or projects: −15
04. Real-World Score Examples
Four clients showing how the formula produces different scores based on actual relationship activity:
- Brightworks Co. — Active tasks (+20), active project (+10), contact this week (+15), follow-up scheduled (+5): Score 100. Active, well-maintained relationship.
- Folio Studio — Active tasks (+20), contact within 30 days (+5), no follow-up scheduled (0): Score 75. Healthy but could benefit from a scheduled next step.
- Redwood Agency — No tasks or projects (0), contact 31-60 days ago (−10), no follow-up (0): Score 40. At Risk — work has slowed and communication is fading.
- Crest & Pine — No tasks or projects (−15), contact beyond 60 days (−25), no follow-up (0): Score 10. Dormant — highest priority for immediate outreach.
05. The Three Client Health Buckets
Scores group naturally into three buckets that suggest different responses: **Healthy (50-100):** Active work, regular communication, relationship momentum. Action: schedule check-ins every 2-4 weeks, propose additional scope, request referrals. These clients are advocates — invest in deepening the relationship. **At Risk (20-49):** Cooling relationships where work is slowing and communication gaps are opening. Action: unscheduled outreach call this week. Do not wait for the next scheduled check-in. This bucket generates most agency churn — clients here rarely proactively communicate that they are drifting. **Dormant (0-19):** No active work and no meaningful contact for 60+ days. Action: immediate personal outreach — this is highest-priority. Some dormant clients are simply between projects; others have made a quiet decision to move on. You cannot know which without contacting them.
06. Enriching the Formula With Additional Signals
The four-component formula works as a starting point. Additional signals improve accuracy for agencies with the data to track them:
- Invoice payment status: On-time payments (+5), overdue 30 days (−10), overdue 60+ days (−20). Slow payment is often an early signal of relationship problems before they are articulated.
- Briefing response rate: How quickly clients respond to briefing requests. Fast responses indicate active engagement; slow or absent responses indicate declining attention.
- Meeting frequency trend: Is the meeting cadence stable, increasing, or decreasing? A client who used to meet weekly now meeting monthly without a stated reason is a signal.
- New project initiation frequency: Clients who proactively bring new work tend to stay. Clients who have stopped initiating new projects are more likely to be evaluating alternatives.
07. Tracking Trends Over Time
A single score matters less than the direction of change over time. A score of 60 rising from 45 over four weeks indicates a recovering relationship. The same score of 60 falling from 80 indicates decline. Two patterns warrant immediate outreach regardless of the current score: Sustained decline: any client whose score has declined for three or more consecutive weeks, even if still in the Healthy range. A score falling from 85 to 78 to 71 to 64 in four weeks is a trend, not normal variation. Sharp drop: a decline of 20 or more points in a single week. This indicates a significant change in relationship activity — a project ending, a communication lapse, or something else that happened and requires follow-up.
08. Three Levels of Automation
Agencies implement health scores at three levels of sophistication:
- Level 1 — Spreadsheet: Manual entry in Google Sheets. Feasible for fewer than 8 clients; the manual input discipline breaks as team and client base grow. Score accuracy depends on someone updating it every week, which rarely happens consistently.
- Level 2 — Manual CRM field: A custom field requiring weekly human entry. Better than spreadsheets because the score lives alongside client data, but the accuracy problem remains. Inconsistent without enforced process.
- Level 3 — Automated calculation: Formula runs continuously based on live CRM data. Scores recalculate when tasks are updated, contacts logged, or follow-ups scheduled. No manual input. This is the only approach that scales reliably beyond 10-15 clients.
09. Implementation in Melororium
Melororium calculates client health scores automatically from existing CRM data — no manual scoring, no weekly spreadsheet discipline required. The dashboard shows three stat cards: Healthy clients (count + revenue total), At Risk clients (count + revenue at risk), and Dormant clients (count + dormant revenue). The revenue totals make the business impact of relationship health immediately visible — not just how many clients are at risk, but how much revenue is at risk. The client table includes individual health scores with one-line explanations of the primary driver: 'At Risk — no contact in 38 days' or 'Healthy — active project, recent follow-up.' Each explanation points to the specific action that would improve the score. Scores recalculate in real time: logging a client contact immediately updates their communication recency. Assigning a new task immediately updates their activity signal. The score you see reflects the current state of the relationship, not last week's data.
See which clients are healthy, at risk, and dormant — with the revenue totals for each.
Melororium's CRM calculates health scores automatically and shows the business impact in one dashboard.


