Marketing Attribution Is Broken: A CMO's Guide to What Actually Works

Why Every Attribution Model Is Wrong (And What to Do About It)

Here's the uncomfortable truth: no attribution model accurately measures marketing impact. Not last-click. Not first-click. Not linear. Not data-driven. Not MMM.

They're all approximations, and treating any of them as truth leads to bad capital allocation.

So what should CMOs actually do?

The PMC Measurement Framework

Instead of trying to attribute every dollar, we measure marketing effectiveness at three levels:

Level 1: Business-Level Metrics (The P&L View) - **MER (Marketing Efficiency Ratio):** Total revenue ÷ total marketing spend - **Incremental Revenue:** Revenue above your organic baseline - **CAC Payback Period:** Days until a new customer becomes profitable - **Contribution Margin After Marketing:** Revenue minus COGS minus marketing

These metrics don't require attribution. They measure the whole system.

Level 2: Channel-Level Directional Signals - **Platform-reported ROAS:** Useful as a directional signal, not a source of truth - **Blended channel CAC:** What each channel costs per new customer - **Incrementality tests:** Holdout tests that measure true lift - **Post-purchase surveys:** "How did you hear about us?" data

Level 3: Customer-Level Insights - **Cohort analysis:** How do customers acquired in Month X behave over time? - **Channel-of-origin CLTV:** Which channels produce the best long-term customers? - **Attribution-agnostic path analysis:** What touchpoints do your best customers share?

Why Last-Click Is Actively Harmful

Last-click attribution systematically: - Overcredits brand search, email, and retargeting (they're closers, not openers) - Undercredits social, content, and awareness channels (they're openers, not closers) - Ignores offline touchpoints entirely

The result: brands over-invest in bottom-funnel tactics and wonder why top-of-funnel growth slows.

The Incrementality Testing Framework

The gold standard for measurement is incrementality testing:

Geo-Holdout Tests - Turn off a channel in specific regions - Measure the revenue difference vs. control regions - Requires 4–6 weeks of clean data

Platform Holdout Tests - Facebook's Conversion Lift studies - Google's Brand Lift studies - Limited by platform cooperation

Budget Shift Tests - Increase spend 30% on one channel for 4 weeks - Measure MER change at the business level - The most practical test for $5M–$25M brands

What PMC Recommends for $5M–$25M Brands

Skip the $200K MMM study. Skip the multi-touch attribution platform. Instead:

1. 1.Track MER weekly. It's the single best metric for overall marketing health.
2. 2.Run one incrementality test per quarter. Start with your biggest channel.
3. 3.Survey post-purchase customers. Simple but powerful first-touch data.
4. 4.Analyze cohorts monthly. How do customers from different channels behave over 90/180/365 days?
5. 5.Report contribution margin, not ROAS. Your CFO will thank you.

The Attribution Trap

The biggest mistake CMOs make isn't choosing the wrong model. It's spending so much time on attribution that they neglect the things that actually drive growth: creative quality, offer strategy, product-market fit, and customer experience.

Perfect attribution with mediocre execution loses to imperfect attribution with excellent execution every single time.