Understand BlockNinja Analytics

Goal

Use BlockNinja Analytics to understand patterns in visits, engagement, performance, and errors without treating a single number as proof of success or identifying individual visitors.

Before you begin

• Use an Admin or Superadmin account with permission to view analytics.
• Write down the question, date range, and audience segment before opening the dashboard.
• Confirm the site's timezone and any campaign dates.
• Keep screenshots private when they contain hostnames, addresses, search terms, or low-volume location data.

Open Analytics and set scope

1. Open Analytics.
2. Choose a date range.
3. Decide whether Exclude bots should be active.
4. Use the same range and bot choice when comparing periods.
5. Choose Overview, Web Vitals, or Errors for the question.

Read the traffic metrics

If JS-confirmed appears with visitors, treat it as additional measurement context rather than a promise that all users were measured.

Read engagement metrics

Avg Time, Avg Scroll, Bounce Rate, and Avg Session Length describe different behaviours. Interpret them alongside the page's purpose. A short time can be successful for a contact-detail page; a high scroll percentage does not prove comprehension.

Use bot filters consistently

Use Exclude bots when reviewing likely human behaviour. Include bot traffic when investigating crawling, monitoring, or unusual load. Record the filter state in every comparison so unlike totals are not presented as a trend.

Read Pageviews Over Time

Look for sustained changes, expected campaign peaks, and sudden anomalies. Compare equivalent weekdays and similar date lengths. Annotate publishing, campaign, outage, and configuration events before attributing a change to content.

Review content and acquisition reports

Do not combine low-volume dimensions to infer a person's identity. Share aggregates only when the audience and privacy policy permit it.

Review Web Vitals

Open Web Vitals and look for repeated poor experiences across enough observations. Use the metric and affected page to reproduce the experience on an equivalent device and network. A dashboard signal starts an investigation; it does not replace a direct performance test.

Review HTTP errors

Use the HTTP Errors summary or View error log, then open Errors for detail. Prioritise errors on launch-critical pages, repeated failures, and recent regressions. Keep request addresses, user inputs, and environment details out of public screenshots.

Check tracking coverage

Review Tracking Coverage and Ignored IPs before interpreting a gap as missing audience activity. Coverage explains measurement boundaries. Ignored IPs can reduce internal traffic but should be administered as private operational data.

Turn a pattern into an action

1. State the observed pattern and filter scope.
2. Choose one plausible explanation.
3. Gather a second signal, such as a public-page replay, error log, or campaign record.
4. Make one safe content or configuration change.
5. Verify the public journey immediately.
6. Review the same metric after a suitable comparison period.

Avoid changing several variables and then crediting one of them.

Protect Reset Data

Reset Data is destructive. Do not select it for exploration, screenshots, troubleshooting, or documentation replay. If an authorised retention process requires a reset, confirm scope, backup or export requirements, approvals, and recovery expectations before the final confirmation.

Recover from a problem

• Totals look unexpectedly high: compare with Exclude bots on and inspect the time range.
• Totals look unexpectedly low: review Tracking Coverage, ignored traffic, publication dates, and the selected range.
• Two reports disagree: align timezone, range, bot filter, and metric definitions before escalating.
• An error exposes private data: restrict the evidence, follow the incident process, and do not copy it into the Wiki.
• Data was reset accidentally: stop further changes and escalate immediately; do not promise restoration unless a verified recovery path exists.

Completion check

The question, range, timezone, bot state, and metric are recorded; at least one second signal supports the interpretation; the resulting action was verified publicly; and shared evidence contains no private or identifying data.

Publication gate: keep this guide private until Overview, Web Vitals, Errors, filters, responsive layout, and the non-destructive Reset Data boundary have been independently replayed.