Funnel Analysis

Funnels let you measure how users progress through a sequence of steps - like a signup flow, checkout process, or onboarding sequence. See where users drop off and identify opportunities to improve conversion.

How funnels work

A funnel is an ordered sequence of events (pageviews or custom events). Moneat calculates how many visitors completed each step and what percentage dropped off between steps.

Step 1: Visit /pricing         →  1,200 visitors  (100%)
Step 2: Click "Start Trial"    →    840 visitors   (70%)
Step 3: Complete Signup        →    504 visitors   (42%)
Step 4: First Service Created  →    302 visitors   (25.2%)

Key concepts

• Strict ordering - Steps must occur in the defined order within a single session
• Session-scoped - All steps must happen within the same visitor session
• No time limit between steps - Steps can occur at any point during the session
• Pageviews and custom events - Mix both types freely in a funnel

Creating a funnel

Via the dashboard

1. Navigate to Analytics → Funnels tab
2. Click New Funnel
3. Add steps by selecting event names or page paths
4. Click Analyze to see results

Via the API

GET /v1/analytics/{projectId}/funnel?steps[]=pageview:/pricing&steps[]=Signup Start&steps[]=Signup Complete&period=30d

Each steps[] parameter defines one funnel step:

Reading funnel results

The funnel response contains:

{
  "steps": [
    {
      "name": "pageview:/pricing",
      "visitors": 1200,
      "conversionRate": 100.0,
      "dropoffRate": 0.0
    },
    {
      "name": "Start Trial",
      "visitors": 840,
      "conversionRate": 70.0,
      "dropoffRate": 30.0
    },
    {
      "name": "Signup Complete",
      "visitors": 504,
      "conversionRate": 42.0,
      "dropoffRate": 28.0
    }
  ],
  "overallConversionRate": 42.0
}

Funnel examples

Signup flow

Step 1: pageview:/pricing
Step 2: Start Signup
Step 3: Complete Signup

Measures: How many pricing page visitors convert to signups.

E-commerce checkout

Step 1: Add to Cart
Step 2: pageview:/checkout
Step 3: Enter Payment
Step 4: Purchase Complete

Measures: Cart-to-purchase conversion rate and where buyers abandon.

Onboarding completion

Step 1: Complete Signup
Step 2: Create Service
Step 3: Install SDK
Step 4: First Event Received

Measures: How effectively new users complete onboarding.

Content-to-conversion

Step 1: pageview:/blog/getting-started
Step 2: pageview:/pricing
Step 3: Start Trial

Measures: How blog content drives trial signups.

Following people across sign-in

Product events that your backend sends always carry a user_id, and the product funnels, saved funnels, feature flag comparisons, and retention grids count people by it (group_by=user_id). A person who starts signed out sends events under an anonymous user_id, then under their account's once they sign up or sign in. Moneat counts them as one person when a signed-in event names the anonymous user_id it replaces in the anonymous.id_hash property:

{
  "name": "signup.completed",
  "user_id": "acct_8f2c",
  "props": { "anonymous.id_hash": "anon_41d9" }
}

Every event sent under anon_41d9 then counts as acct_8f2c's. Grouping by session is unchanged.

• Linking events - Any event whose anonymous.id_hash is set and differs from its own user_id links the two. Sending it on every signed-in event is simplest; one is enough.
• Several devices - Each device's anonymous id links to the same account, so the person counts once.
• Shared devices - An anonymous id linked to several accounts belongs to the first account seen with it. Later accounts keep their own signed-in events, and no event counts twice.
• Lookback - Links are read from the queried range and the 30 days before it, so "first" means first in that span, and a person who signed in shortly before the range still resolves. Links made after the range never count, even when a retention period reads returns past it.
• Services - A user_id belongs to its service. The same account id in two services counts as two people when a query covers both.
• No chaining - An anonymous id resolves to the account that named it, never further.
• Feature flag comparisons - A flag evaluated for an anonymous id counts toward its account. A person assigned more than one variant counts in the latest one. Evaluations in the same millisecond go to the greatest targeting key, so a tie resolves the same way every time.

Combining with filters

Apply any filter alongside a funnel to segment results:

GET /v1/analytics/{projectId}/funnel
  ?steps[]=pageview:/pricing
  &steps[]=Signup Complete
  &period=30d
  &filters[]=country_code:is:US
  &filters[]=device_type:is:Desktop

This shows the funnel conversion rate for US desktop visitors only.

Best practices

• Start broad, then narrow - Begin with 2–3 steps and add detail once you understand the high-level flow
• Use meaningful step names - Name custom events clearly so funnels are self-documenting
• Compare segments - Run the same funnel with different filters to find which audience segments convert best
• Track both intent and completion - Pair "started" events with "completed" events (e.g., Start Checkout → Complete Purchase)
• Monitor over time - Check funnel conversion rates weekly to catch regressions after product changes