A Multi-Channel Funnel (MCF) shows how different traffic sources cooperate on the path to a conversion. Instead of crediting only the last click, MCF reconstructs the full conversion path across channels like organic, paid, email, social, referral, and direct. It’s the practical backbone of multi-touch attribution: you see assists, not just closers.
Why it matters
Single-touch models hide reality. MCF exposes supportive roles (assists), over-reliance on brand/direct, and cross-channel cannibalization. With a sensible lookback window and consistent UTM parameters, budgeting becomes evidence-based: scale what starts journeys, defend what closes them, fix channels that add noise.
Core concepts & quick math
- Assist (Assisted Conversion): a channel appeared on the path but wasn’t final; see assisted conversion.
- Last click: the closer; see last-click attribution.
- First click: the opener; see first-click attribution.
Assist-to-Last-Click Ratioassist_ratio(channel) = assisted_conversions(channel) / last_click_conversions(channel)
1 → great supporter; <1 → finisher.
Mini example
User path: Social → Email → Direct → Purchase
| Step | Channel | Role in MCF |
|---|---|---|
| 1 | Social | Assist |
| 2 | Assist | |
| 3 | Direct | Last click (conversion) |
Here, Social = 1 assist, Email = 1 assist, Direct = 1 last click. Content/SEO often rack up assists; brand/direct frequently close.
Implementation notes (analytics-savvy)
- Set the lookback window to your real buying cycle.
- Normalize channel groupings so “Paid Social” isn’t fragmented.
- Track at the user level (see unique visitor) to preserve continuity.
- Compare MCF views under different models: linear, time-decay, position-based (see attribution).
Tactically, rank channels by assists, last-clicks, and assist_ratio, then assign each a job: introducer, educator, or closer. That’s how you turn budgeting from myth into math.