Traffic Anomaly Explainer

What it does

Cross-references the date of the anomaly against publish dates, marketing campaign launches, referral source changes, and any known site or plugin changes to build a ranked list of plausible causes rather than a single guess. Distinguishes a real anomaly from normal day-of-week or seasonal variance before investigating further.

Install

This skill isn't packaged in an installable registry. It's a plain SKILL.md file you can copy directly. Copy it from the SKILL.md tab and save it at the path for your assistant:

Use it in your AI assistant

Claude Code

.claude/skills/traffic-anomaly-explainer/ (or ~/.claude/skills/traffic-anomaly-explainer/ for all projects)

GitHub Copilot (VS Code)

.github/skills/traffic-anomaly-explainer/

OpenAI Codex

.codex/skills/traffic-anomaly-explainer/

Cursor

.cursor/skills/traffic-anomaly-explainer/ (or ~/.cursor/skills/traffic-anomaly-explainer/ globally)

Compatibility

Requires a connected analytics source, such as Google Analytics or Jetpack Stats, for traffic data; works on any WordPress site with that integration in place.

The complete SKILL.md for this skill:

---
name: traffic-anomaly-explainer
description: Use when a user notices a sudden traffic spike or drop and wants a likely explanation, or asks "why did traffic jump" or "why did traffic drop".
---

# Traffic Anomaly Explainer

## When to use
Traffic moved sharply in one direction and the user wants a data-backed explanation rather than a guess, before deciding whether to act on it.

## Procedure
1. Confirm the anomaly is real by comparing against the same weekday in prior weeks, not just the day before, to rule out normal variance.
2. Check what published or changed on the site around that date: new posts, plugin or theme updates, price or product changes.
3. Check referral and channel breakdown for the anomaly window, a spike concentrated in one source points to a specific cause.
4. Check for anything external and known, a mention elsewhere or a social post going viral, if referral data shows an unfamiliar spike source.
5. Present a ranked list of plausible causes with the evidence for each, not a single definitive claim.

## Guardrails
- Don't state a cause as certain when the evidence is only correlational, phrase it as most likely with the supporting signal named.
- If the drop coincides with a tracking or tagging change, flag that a measurement problem is possible before assuming real traffic loss.