Analytics Platforms: GA4, Product Analytics & the Modern Stack

Analytics Platforms

Analytics platforms collect data about how people use your product or site and turn it into reports you can act on. The category spans classic web analytics (page views, traffic sources, conversions — GA4 is the dominant example), product analytics (event-level user behavior — Amplitude, Mixpanel, PostHog), and increasingly warehouse-native analytics built directly on your data warehouse. They overlap, but they answer different questions and model data differently.

The shift that defines modern analytics is from sessions and pageviews to events: instead of "how many pages did this visit touch," teams increasingly ask "what actions did this user take, and how do they flow through funnels and cohorts." Understanding that model — and which platform fits which question — is the foundation of analytics engineering.

TL;DR

Web analytics (GA4) measures traffic, sources, and conversions.
Product analytics (Amplitude/Mixpanel/PostHog) measures event-level user behavior — funnels, retention, cohorts.
Warehouse-native analytics builds reporting on your own data warehouse for full control.
Modern analytics is event-based, not pageview/session-based.

Quick Example

Modern analytics revolves around tracking events with properties:

The Categories

These aren't mutually exclusive — many teams run web analytics for marketing and product analytics for behavior, increasingly unified in the warehouse.

What They Measure

Acquisition — where users come from (channels, campaigns, referrers).
Behavior — what they do (events, funnels, paths).
Conversion — completion of key goals (signup, purchase).
Retention — whether users come back (cohorts, stickiness).
Attribution — which touchpoints get credit (see Attribution).

Sessions vs events

Classic web analytics groups activity into sessions (a visit). Product analytics centers on events (individual actions) tied to a persistent user, which models funnels, retention, and cross-session journeys far better. GA4 itself moved to an event-based model for this reason.

Choosing a Platform

Marketing/acquisition focus → web analytics (GA4 + cross-link to SEO Engineering).
Product/behavior focus → product analytics (funnels, retention, experiments).
Maximum control, custom metrics, one source of truth → warehouse-native (your warehouse + data engineering + BI).
Privacy-first / lightweight → cookieless analytics (Plausible, Fathom).

💡 Many teams use more than one. The unifying trend is sending events to your data warehouse so every tool reads from one consistent, governed dataset.

Best Practices

Define a tracking plan first — name events and properties consistently before instrumenting (see Event Tracking).
Centralize collection — send events once (often via a data layer / CDP) and fan out to tools.
Mind privacy and consent — analytics is subject to GDPR/CCPA; gate tracking on consent.
Pick platforms by the question — acquisition vs behavior vs custom modeling.
Land raw events in the warehouse as your durable source of truth.

Common Mistakes

No tracking plan

Treating pageviews as behavior

FAQ

What's the difference between web analytics and product analytics?

Web analytics (GA4, Plausible) focuses on traffic and acquisition — where visitors come from, which pages they see, and whether they convert — organized largely around sessions. Product analytics (Amplitude, Mixpanel, PostHog) focuses on behavior at the event level: what actions users take, how they move through funnels, and whether they retain, tied to persistent user identities. Marketing teams lean on web analytics; product teams lean on product analytics. Many organizations use both.

Is GA4 a web analytics or product analytics tool?

GA4 blurs the line. It replaced the old session/pageview-centric Universal Analytics with an event-based model, bringing it closer to product analytics — you can build funnels and explore events. But it remains strongest for acquisition, marketing channels, and conversions, and most teams that want deep funnel/retention/cohort analysis still pair it with a dedicated product analytics tool. Think of GA4 as event-based web analytics with some product-analytics capability.

Should I send analytics to my data warehouse?

Increasingly, yes — it's the foundation of the "modern data stack." Landing raw events in a warehouse (Snowflake, BigQuery) gives you one governed source of truth, full SQL access for custom metrics, and the ability to join behavioral data with everything else (billing, support, CRM). You can then feed BI tools and even other analytics platforms from it. Off-the-shelf tools are faster to start; the warehouse gives control and longevity. See Data Engineering.

How do privacy laws affect analytics?

Significantly. GDPR, CCPA, and similar laws regulate collecting and processing personal data, which most analytics does (identifiers, IP, behavior). You generally need a lawful basis and, in many regions, explicit consent before tracking — so analytics should be gated behind a consent banner, with options to anonymize IPs and respect "do not track." Privacy-first tools (Plausible, Fathom) avoid cookies and personal data to simplify compliance. Build consent and data-minimization in from the start, not as an afterthought.