Overview

Regression analysis turns your run history into trends so you can tell whether quality is improving or slipping. It reports whether a project or test has regressed, how confident that verdict is, whether performance has degraded, and how success, failure, duration, and response time are trending per day. Explore the capability at regression analysis.

Regression analysis draws on the same run history as the Trends Analysis section of a project dashboard and the Pass Rate Trend Analysis chart. It reads your recorded runs — no configuration and no AI key are required.

Analytics is a paid capability. On the Free edition the Analytics area is gated; Professional, Trial, and Enterprise licenses unlock it.

Before you begin

You need a project with run history. Regression signals are statistical — a project or test with only one or two runs shows "No regression data available" until more runs accumulate.
Navigation path: top navigation bar → Analytics → Project Analytics tab → choose a project.
Scores are computed over a time range (the default is the last 30 days). Widen the range to smooth out noise; narrow it to focus on a recent release.

Step 1 — Open Project Analytics

In the top navigation bar, click Analytics.
The Analytics area opens with two tabs: Dashboards and Project Analytics. Click Project Analytics.
Pick a project from the list. The project's analytics dashboard opens with collapsible sections: Project Overview, Performance Metrics, Trends Analysis, Top Failing Tests, Environment Comparison, and Endpoint Analytics.

Step 2 — Read the regression verdict

The regression panel summarizes whether quality has slipped over the time range.

Field	What it does
Regression Status	A chip that reads Regression Detected (red) or No Regression (green).
Confidence Level	A progress bar and percentage (0–100%) with a severity label beneath it: High (80%+), Medium (60–79%), Low (40–59%), or None (below 40%). Higher confidence is a stronger signal that the change is real, not noise.
Regression Reasons	A bulleted list explaining why a regression was flagged (for example, a drop in pass rate or a spike in failures). Shown only when reasons exist.

Step 3 — Check for performance regression

When response times or durations have degraded, a Performance Regression block appears with a warning callout describing the reason, plus up to two metrics:

Field	What it does
Response Time	The recent average response time in milliseconds, with the percentage increase over the baseline shown in red.
Duration	The recent average test duration in milliseconds, with the percentage increase shown in red.

If performance hasn't regressed, this block is hidden.

Step 4 — Read the per-day trends

The Trend Analysis grid shows the direction and rate of change for four metrics. Each has a trend arrow — up (green), down (red), or flat (—) — and a "% per day" figure:

Field	What it does
Success Rate	How the pass rate is trending per day. An up arrow is good here.
Failure Rate	How the failure rate is trending per day. A down arrow is good here.
Duration	How average test duration is trending per day.
Response Time	How average response time is trending per day.

Direction is derived from the metric's slope over the time range: a small positive slope shows as trending up, a small negative slope as trending down, and near-zero as flat.

Step 5 — Compare recent runs

The Recent Status History list compares the latest runs so you can see exactly when a regression started. Each row shows:

A status chip — PASSED (green), FAILED (red), or another state (amber).
The run date.
The run duration in milliseconds.

Scan top to bottom: a run of green that turns red pinpoints where quality changed, and the duration column reveals slowdowns even when the status stayed green.

Step 6 — Cross-check with the trend charts

Two related views corroborate the regression verdict:

The Trends Analysis section on the project dashboard breaks recent days into per-day trend rows with success-rate figures.
The Pass Rate Trend Analysis chart plots average pass rate over time. Filter it by Run Type (All Types / Manual / Scheduled), Environment (free text, e.g. staging, production), and Time Period (Last 7 days / Last 30 days / Last 90 days / Last year), then use Clear Filters to reset. Below the chart, tiles show Average Pass Rate, Highest Pass Rate, Lowest Pass Rate, and Data Points.

Troubleshooting / Notes

"No regression data available" — the project or test doesn't have enough runs in the selected time range yet. Run more tests or widen the range.
Confidence shows "None" — the data doesn't support a regression conclusion; this is the healthy state, not an error.
Analytics area is locked — Analytics is gated on the Free edition. Upgrade to Professional or Enterprise (Trial also unlocks it).
Trends look flat despite failures — a flat (—) trend means the slope is near zero over the range; a single bad run may not move a 30-day slope. Narrow the time range to surface recent shifts.
Pass Rate chart says "No data available for the selected filters" — loosen the Run Type / Environment / Time Period filters or clear them.

Regression Analysis