What is web analytics?

Last updated: 2026-06-04

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.

Strong existing analytics services

There are already several strong analytics products on the market.

• Google Analytics: https://marketingplatform.google.com/about/analytics/
• Plausible Analytics: https://plausible.io/
• PostHog: https://posthog.com/
• Umami: https://umami.is/

These are real, credible choices. MoonRock is not trying to pretend they do not exist.

Why those tools are great

If you are technical, or you already have developers, product managers, growth people, or data analysts on your team, those tools can be excellent.

• Google Analytics is powerful for attribution, reporting, Google Ads integration, and deeper analysis workflows.
• Plausible is attractive when you want a simpler, privacy-friendly website analytics product with an easy dashboard.
• PostHog is strong when you want product analytics, session replay, funnels, feature flags, warehouse-style workflows, and a broad product engineering stack.
• Umami is attractive when you want an open-source, privacy-focused option with API access and self-hosting flexibility.

The tradeoff with traditional analytics tools

Those products are often best when someone on the team can:

• choose the right tool
• implement the tracker correctly
• define the right events
• understand the reports
• interpret the results and turn them into decisions

That is a perfectly good model for technical teams.

But if you are not technical, or if you do not want to personally deal with dashboards, event taxonomies, tagging plans, or query tools, even a good analytics product can still feel like one more system to operate.

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.

The idea is not that traditional analytics tools are bad. The idea is that many people do not want to be the one installing, configuring, querying, and interpreting them by hand.

MoonRock is for the case where the human wants the result, and the agent handles the operational work.

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

In other words:

• a technical team may choose Google Analytics, Plausible, PostHog, or Umami and operate analytics directly
• a non-technical user can ask their agent to use MoonRock Analytics and let the agent handle setup, tracking, and questions

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.