Overview
Health scoring combines configured metrics and thresholds with customer evidence and AI assessment. This page explains how to interpret the settings and their limits.Thresholds and Buckets
Metrics with threshold controls use three values to define scoring bands. These guide the configured rubric. Qualitative category assessments also consider relationships between the metrics and supporting evidence; their metric rows are not always independent threshold calculations. Example - “Days Since Last Call” with thresholds [14, 30, 60] and direction “lower is better”:Direction
Each metric specifies whether higher or lower values indicate better health:Available Metrics by Category
Revenue
- MRR (Monthly Recurring Revenue)
- Outstanding invoice balance
- Days until renewal
Product Usage
Available metrics come from your configured usage sources, such as a warehouse, PostHog, or the Usage API. They depend on the data you send to Quivly and which metrics you include in scoring. Each usage metric supports a scoring basis (Trend + Volume, Absolute Value, Growth Percentage, or Absolute Difference) and a time period (30d to 1y). If every enabled usage metric uses Absolute Value, the category is calculated from stored, permitted observations using each metric’s aggregation, thresholds, and lookback period. All windows end at the same assessment cutoff. The calculation preserves metric weights and source exclusions. Missing observations or an unreadable source must not be treated as zero. Other bases use AI assessment of the configured rubric and available evidence.Engagement
- Calls (volume over a configurable time period)
- Days since last call
- Call sentiment
- Slack conversations (volume over a configurable time period)
- Days since last Slack message
- Slack sentiment
Support
- Open ticket count
- Ticket volume (configurable time period)
- First response time
- Resolution time
Market Signals
Quivly assesses available developments against the scoring instructions. Positive or negative signals are interpreted in customer context rather than applied as a fixed adjustment. Configure the lookback period (30d or 90d).How Scores Combine
- Each enabled category is assessed against its configuration and available evidence. Qualitative metric rows explain that category’s assessment; inferred metrics should be read with their evidence and interpretation.
- For the all-Absolute-Value usage rubric, metric scores are combined using their metric weights to produce the category score.
- Category scores are multiplied by their category weights and summed to produce the overall health score.
