# ZDX score factor explainer

> Paste the metrics ZDX exposes and read back what each one is: probe family, documented meaning, the score against the documented Poor band, and the honesty notes on what Zscaler does not publish.

- Tool: https://ronutz.com/en/tools/zdx-score-factor-explainer
- Family: Networking

---

# ZDX score factor explainer

Paste the metrics Zscaler Digital Experience exposes - the composite score, the Web Probe metrics, the CloudPath metrics - and read back what each one is, per the vendor's own documentation: which probe family measures it, what it means, and how the documented score semantics apply.

## Honesty scope, stated up front

Zscaler documents the ZDX score's factors, its probing cadence, its lowest-of-the-hour rule, and the Poor band - and does **not** publish the exact composite formula. This tool therefore **computes no score from metrics**. It classifies a given score against the documented Poor band (0-33), explains each metric's documented meaning and probe-family attribution, and states plainly where the documentation stops. For raw metrics, no published threshold makes a value good or bad by itself - ZDX alert criteria are administrator-defined per rule - so the tool explains the metric and declines to grade it. Same no-invention rule as the rest of this site's tools.

## Input grammar

One metric per line, `metric = value`. Lines starting with `#` are comments.

- `score` - the composite ZDX score, a 1-100 value
- `pft`, `dns`, `srt` - Web Probe times in milliseconds (Page Fetch Time, DNS Time, Server Response Time)
- `availability` - Web Probe availability, percent 0-100
- `path-latency`, `path-loss` - CloudPath metrics (milliseconds, percent)

## What you get back

Each metric renders as a reading card: its probe-family badge (Web Probe / CloudPath / composite), its documented meaning - Page Fetch Time requesting only the top-level page document, Server Response Time as the time to first byte - and, for the score, classification against the documented Poor band with the automatic root-cause-analysis note. When both probe families are present, the diagnostic split renders: bad web metrics over a clean path point at the application; a bad path indicts the network. The standing notes carry the documented semantics every reading rides on: five-minute probes, the lowest score of the hour as the hourly score, group scores averaging each user's lowest, the roughly twenty-minute telemetry delay, and the small approximate-aggregation variance between API and dashboard.

Everything runs locally in your browser; nothing you paste leaves the page.

## Sources

- Zscaler Reference Architecture: Zscaler Digital Experience (ZDX) - probe metrics, cadence, lowest-of-hour and group-averaging semantics
- Zscaler Help: Evaluating User Details - the documented Poor band (0-33) and its automatic root-cause analysis
- Zscaler Help: Understanding the ZDX API - the telemetry delay and aggregate-variance calibrations

## Standards and references

- [Zscaler Reference Architecture: Zscaler Digital Experience (ZDX)](https://help.zscaler.com/downloads/zdx/reference-architecture/zscaler-digital-experience-zdx/zscaler-digital-experience-zdx-reference-architecture.pdf) - Web Probe metrics (Page Fetch Time requesting only the top-level page document, DNS Time, Server Response Time, Availability); probes every 5 minutes for most plans; the lowest score of the hour as the displayed hourly score; group scores averaging each user's lowest over the timeframe
- [Zscaler Help: Evaluating User Details (ZDX)](https://help.zscaler.com/zdx/evaluating-user-details) - the documented Poor band (scores 0-33) and the automatic root-cause analysis run on the most recent Poor score
- [Zscaler Help: Understanding the ZDX API](https://help.zscaler.com/zdx/understanding-zdx-api) - the ~20-minute estimated telemetry reporting delay and the ~2% approximate-aggregation variance between API and dashboard values

## Related reading

- [The ZDX Score: What the Probes Measure and How the Number Is Made](https://ronutz.com/en/learn/zdx-score-anatomy-and-probes.md): Zscaler Digital Experience turns user experience into a 1-100 score, and the anatomy is precise: Web Probes measuring Page Fetch Time, DNS Time, Server Response Time, and Availability; CloudPath probes measuring the network path hop by hop; probes every five minutes; the lowest score of the hour becoming the hour's score; group scores averaging each user's worst. What each metric means, the Poor band's automatic root-cause analysis, and what Zscaler does not publish about the formula.