# Understanding Analytics in Shift-Left API

> How Analytics differs from Reporting: trends over time rather than one run. Covers project metrics, time ranges, and what each view answers.

Source: https://totalshiftleft.ai/help-center/product-documentation/understanding-analytics

## Overview

**Analytics & Dashboards** turns your execution history into trends. It's reached from the top navigation and requires a **Pro**, **Trial**, or **Enterprise** plan.

The distinction that matters:

| Area | Question it answers | Scope |
|------|--------------------|-------|
| **Reporting** | "What happened in *this* run?" | One execution |
| **Analytics** | "What's been happening over *time*?" | Many executions |

If you're diagnosing a failure that just occurred, you want [Test reporting](/help-center/product-documentation/test-reporting), not Analytics. If you're asking whether quality is improving, whether a test is flaky, or which endpoint keeps breaking, you want Analytics.

Analytics is project-scoped — you'll be prompted to **Select a Project** first.

## Time ranges

Every view is bounded by a range: **Last 7 days**, **Last 30 days**, **Last 90 days**, or **Last year**.

Pick the range to match the question. A regression introduced yesterday is invisible at **Last year** — the trend line barely moves. A gradual performance decline is invisible at **Last 7 days** — it looks like noise. Most "the dashboard doesn't show the problem" confusion is a range mismatch.

## What's measured

| Metric | What it tells you |
|--------|-------------------|
| **Total Tests** | How many tests ran in the range. |
| **Success Rate** | The proportion that passed. |
| **Coverage %** | How much of your API surface is exercised. |
| **Avg Response Time** | Typical response time across runs. |
| **Failed Tests** | The count of failures. |
| **Test Status Breakdown** | Passed / Failed / Error / Not run over time. |
| **Response Time Distribution** | The spread of response times, not just the average. |
| **Top Failing Tests** | Which tests fail most often. |

**Top Failing Tests** is the most actionable panel in Analytics. A test that fails intermittently across many runs is either flaky or catching a real intermittent bug — and either way it's costing you trust in the whole suite. Neither is visible from a single report; you can only see it across time.

**Response Time Distribution** is worth more attention than **Avg Response Time**. An average hides the tail, and the tail is where users notice. A stable average with a widening distribution means some requests are getting much slower while most stay fine.

## Reading Success Rate honestly

Success rate is the number most often quoted and most easily misread. Two cautions:

- **Check it against Test Status Breakdown.** A rate dragged down by **Error** means runs didn't complete — that's an environment problem, not an API quality problem, and "fixing tests" won't move it.
- **Check it against Coverage %.** A high success rate on low coverage means you're reliably testing a small slice. It is not evidence the API is healthy.

The pairing to watch is coverage rising while success rate dips slightly — that usually means you started testing things that were always broken, which is progress, not regression.

## The views

| View | What it's for |
|------|---------------|
| **Project Overview** | The project's headline numbers. |
| **Execution Summary** | Run outcomes across the range. |
| **Performance Summary** / **Performance Metrics** | Response times and their distribution. |
| **Endpoint Analytics** | Which endpoints are healthy and which aren't. |
| **Environment Comparison** | The same tests across environments. |
| **Trends Analysis** | Movement over time. |

**Environment Comparison** answers a specific and common question: is this failure real, or is staging just broken? The same tests passing in one environment and failing in another points at the environment, not the contract.

For building views around a role or an audience, see [Custom dashboard personas](/help-center/product-documentation/custom-dashboard-personas) and [Analytics dashboards](/help-center/product-documentation/analytics-dashboards). For one project in depth, see [Project analytics](/help-center/product-documentation/project-analytics).

## Troubleshooting

- **"No analytics data available"** — no runs in the selected range. Widen the range, or check the project has actually run anything.
- **"No endpoint analytics data available"** — no per-endpoint data in range.
- **"No environment comparison data available"** — the tests haven't run in more than one environment, so there's nothing to compare.
- **Analytics isn't in the navigation** — it's gated on Pro, Trial, or Enterprise.
- **Numbers disagree with a report you're looking at** — check the range and project. Analytics aggregates many runs; a report is one.
- **Success rate dropped with no code change** — check **Test Status Breakdown** for a spike in **Error**. That's usually environment or connectivity, not quality.

## Related articles

- [Analytics dashboards](/help-center/product-documentation/analytics-dashboards)
- [Project analytics](/help-center/product-documentation/project-analytics)
- [Custom dashboard personas](/help-center/product-documentation/custom-dashboard-personas)
- [Test reporting](/help-center/product-documentation/test-reporting)
- [Sharing and embedding dashboards](/help-center/product-documentation/sharing-and-embedding-dashboards)