Analyze web traffic

Skill altertable-ai/skills/skills/analyze-web-traffic

Analyzes website traffic (pageviews, sessions, referrers, landing pages). Use for site visits, traffic sources, bounce, or visitor behavior.From its SKILL.md

Install

npx -y skills add altertable-ai/skills --skill analyze-web-traffic

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

7 stars7 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.6 KB, as published. Nobody here has run it

Analyze Web Traffic

Quick Start

To analyze web traffic:

Query pageview and session data via the Altertable MCP server
Compare the current period against a previous period (WoW, MoM, or YoY)
Segment by traffic source, device, geography, or page
Surface anomalies, trends, and actionable findings

When to Use This Skill

User asks about website traffic, pageviews, or sessions
Investigating traffic spikes, drops, or trends
Comparing traffic across time periods or segments
Evaluating traffic source mix or user engagement

Analysis Workflow

Step 1: Determine Scope and Time Frame

Ask the user (or infer from context):

Period: What date range to analyze (default to last 7 days if unspecified)
Comparison: What to compare against (previous period of equal length)
Focus: Overall traffic, a specific source, a specific page, or a segment

Step 2: Query Traffic Data

Use Altertable MCP tools to fetch web analytics data:

get_catalog to confirm the Product Analytics tables and columns available in this environment
query_lakehouse for custom web traffic analysis
render_insight when the user wants a chart preview

Common Product Analytics tables include product_analytics.analytics.web_sessions and product_analytics.analytics.web_pageviews. Query them to compute:

Volume: pageviews, sessions, unique visitors
Engagement: bounce rate, pages per session, avg session duration
Acquisition: traffic by source/medium, referrer breakdown

Always pull both the current period and the comparison period so you can compute deltas.

Step 3: Segment the Data

Break down by at least one dimension to find where changes originate:

Traffic source: organic, paid, direct, social, referral, email
Device type: desktop, mobile, tablet
Geography: country, region
Page or section: top pages, landing pages, exit pages

When a top-level metric moves, drill into segments to isolate which segment drove the change.

Step 4: Identify Patterns and Anomalies

Look for:

Trends: sustained directional movement across multiple periods
Anomalies: single-period spikes or drops that break the pattern
Shifts in mix: a source growing as a share even if total traffic is flat

Quantify every observation with absolute numbers and percentage change.

Step 5: Summarize and Recommend

Present findings with:

The metric, its value, and the delta vs. the comparison period
Which segment is responsible for the change
A hypothesis for why (site changes, seasonality, campaigns)
A suggested next step or action when applicable

Time Period Comparison Guide

Comparison	When to Use
WoW (week-over-week)	Short-term monitoring, recent changes
MoM (month-over-month)	Growth tracking, campaign evaluation
YoY (year-over-year)	Seasonal businesses, long-term trends

Always compare equal-length periods. When comparing WoW, align on the same day of week. When comparing MoM, account for differing month lengths.

Segmentation Priorities

When the user does not specify a segment, default to this order:

Traffic source -- most common driver of traffic changes
Device type -- surfaces mobile vs. desktop divergence
Top pages -- pinpoints content driving volume shifts

Only add geography or other dimensions if the first pass does not explain the change.

Common Pitfalls

Reporting totals without comparison: always include a delta to a prior period so the user can gauge significance
Ignoring seasonality: a WoW drop on a holiday week is expected, not alarming -- flag it rather than over-interpreting
Mixing up pageviews and sessions: these measure different things; present both when discussing volume
Not drilling into segments: a flat total can hide offsetting gains and losses across sources or pages
Presenting numbers without context: raw counts are meaningless without comparison, percentage change, or benchmarks
Forgetting to check for tracking issues: sudden drops to zero or impossible spikes often indicate instrumentation problems, not real traffic changes

Reference Files

What ships with it: 3 files

14.9 KB alongside SKILL.md

references/

Gives 0 of the 12 instructions most research analysis skills give

Counted across 1,213 of the 2,113 authors here whose files we hold, read 2026-09-06

Cite sources for every important claimin 47 of 1213, across 38 files
Separate facts from inferences and recommendationsin 21 of 1213, across 12 files
Write findings to a markdown filein 19 of 1213
Label every insight with a confidence levelin 18 of 1213, across 8 files
Read product marketing context before asking questionsin 18 of 1213, across 8 files
Rank themes by frequency and intensityin 16 of 1213, across 6 files
Establish research mode before proceedingin 16 of 1213, across 6 files
Segment survey responses by customer tier or tenurein 16 of 1213, across 6 files
Categorize support tickets before analyzingin 16 of 1213, across 6 files
Weight research sources from the last twelve monthsin 16 of 1213, across 6 files
Use at least five data points per segmentin 15 of 1213, across 5 files
Extract verbatim quotes for all research findingsin 15 of 1213, across 5 files

Said here and by no other author read

Query pageview and session data
Segment data by source, device, geography, or page
Pull both current and comparison periods
Compute deltas for all metrics
Drill into segments to isolate changes
Quantify observations with absolute numbers and percentages

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once.