S1 · Platform overview and the hit model

Session 1: Platform overview and the hit model

What is Adobe Analytics?

Adobe Analytics is a web analytics platform that captures user behavior and aggregates it into reports. Think of it like:

Bad analogy: "It's like Google Analytics" Good analogy: "It's a database that collects hits (user actions), stores them with dimensions and metrics, and lets you query them in reports"

The platform layers

Layer 1: Data collection (your site)

Your website runs JavaScript (Web SDK or AppMeasurement). When users interact with your site, JavaScript fires tracking calls.

// User lands on page
s.pageName = "Home";
s.events = "event1"; // page view
s.t(); // sends the hit

Layer 2: Edge Network (Adobe's servers)

The hit travels to Adobe's Edge Network in milliseconds. The Edge validates it, adds metadata, routes to the correct report suite.

Layer 3: Processing (transformations)

Before data lands in the report suite, processing rules and VISTA can modify or delete it.

Layer 4: Report suite (database)

Data lands in your report suite. It's now queryable in Analysis Workspace, Data Warehouse, APIs.

Layer 5: Reporting (Analysis Workspace)

You build reports and dashboards by asking questions of the report suite.

The hit model (core concept)

Every action in Adobe Analytics is a hit.

A hit is one HTTP request carrying:

• What happened (page load, button click, form submit)
• When it happened (timestamp)
• Who did it (visitor ID/ECID)
• Context (which page, what device, what traffic source)
• Variables you defined (props, eVars, events)

Examples:

Hit 1: Alice loads homepage from Google (page view)
Hit 2: Alice clicks "Site Search" link (link click)
Hit 3: Alice types "death benefits" in search (search event)
Hit 4: Alice clicks search result (link click)
Hit 5: Alice is back on results page (page view)

Hits are atomic

You cannot have half a hit. Each hit is independent:

Hit 1: page_name="Home", eVar1="organic_search"
Hit 2: page_name="Results", eVar1="organic_search" (remembered)
Hit 3: page_name="Article", eVar1="organic_search" (remembered)

eVar1 value from Hit 1 is remembered in Hits 2 and 3 *within the same visit.* But each hit is still atomic — it either fires or it doesn't.

Scoping: hits, visits, visitors

Hits are grouped into visits (sessions). Visits are grouped into visitors (people).

Implication for your data:

• Prop value only exists on the hit where it's set (hit-scoped)
• eVar value persists for the visit (visit-scoped) or across all visits (visitor-scoped)
• Visitor count = unique ECIDs in the period

The anatomy of a hit

When you fire a hit, it contains:

All of this goes to Adobe in one request. One hit = one database row.

Example Super: The hit model in action

Alice visits Example Super (superannuation fund website):

Visit 1 (Monday 2:00pm)
  Hit 1: Page view "Home" from Google
    - Visitor ID: ecid_alice123
    - Page: Home
    - Referrer: google.com
    - eVar1: "organic_search"

  Hit 2: Page view "Search Results"
    - Visitor ID: ecid_alice123
    - Page: Search Results
    - eVar1: "organic_search" (remembered from Hit 1)

  Hit 3: Search event "death benefits"
    - Visitor ID: ecid_alice123
    - event1 fires (search)
    - eVar2: "death benefits"

[15 minutes pass]

  Hit 4: Page view "Article - Death Benefits"
    - Visitor ID: ecid_alice123
    - eVar1: "organic_search" (still remembered)
    - eVar2: "death benefits" (still remembered)

[Alice closes browser, comes back Wednesday]

Visit 2 (Wednesday 10:00am)
  Hit 5: Page view "Home" (direct traffic)
    - Visitor ID: ecid_alice123 (same person)
    - eVar1: empty (reset, new visit)
    - eVar2: empty (reset, new visit)

Key takeaways

1. One hit = one action = one database row
2. Every hit has a timestamp and a visitor ID
3. Props are hit-level, eVars are visit/visitor-level
4. Hits are grouped into visits (30 min), visits into visitors (60 days)
5. Your job is to design what data goes into each hit

---

Next: L1 · Read a beacon in the AEP Debugger