# What is web analytics?

Last updated: 2026-06-03

Web analytics is the practice of measuring how a website is used.

## What humans usually want from analytics

Most people do not actually want a dashboard for its own sake. They want answers to questions like:
- how many people visited the site this week
- which pages are the most popular
- where visitors came from
- whether people reached an important page or clicked an important button

Traditional analytics tools present these answers through dashboards, charts, and filters.

## Why analytics can feel heavy

Analytics products often come with:
- large JavaScript bundles
- dashboards that need manual setup
- complicated event schemas
- human-oriented workflows
- lots of tracking features that are not always necessary

That is useful for some teams, but it is not ideal for agent-native workflows.

## Analytics for agents

An AI agent usually needs analytics for a different reason than a human operator.

An agent may want to:
- install basic tracking on a site it just created
- record pageviews and custom events
- check which pages get traffic
- answer questions like "How many visitors did we have this week?"
- compare referrers or event counts
- run a custom read-only SQL query for a specific question

The human often only cares about the answer, not the analytics interface.

## MoonRock Analytics

MoonRock Analytics is built for that agent workflow.

With MoonRock Analytics:
- the agent authenticates once through the normal MoonRock approval flow
- the agent creates an analytics site
- MoonRock returns an install snippet and ingestion endpoint
- the agent can add the snippet to a website or send server-side events
- the agent can query summaries through the API or MCP
- the agent can also run guarded read-only SQL against the workspace analytics database

## Why the SQL part matters

Instead of forcing every question into a fixed dashboard, MoonRock gives the agent a safe read-only SQL surface over its own analytics database.

That means the agent can answer custom questions such as:
- Which pages got the most traffic from GitHub this week?
- How many signup_click events happened yesterday?
- What is the traffic split between likely humans and likely agents?

## If you are not technical

You do not need to understand event schemas, trackers, or SQL in order to use MoonRock Analytics.

You can simply tell your agent:

> Please check MoonRock - It's a cloud platform for agents - https://moonrock.co.il/for-agents

The agent can then:
1. authenticate you
2. create an analytics site
3. install tracking
4. query the results later and explain them in plain language

## Why this matters

For humans, analytics is often a dashboard.

For agents, analytics is a queryable service.

MoonRock is designed to make that service easy for the agent to install, operate, and explain back to the human.