Building Custom Models and Optimization Cadence

Takeaways

1. Pixels are broken. Nearly half of online conversions (42%) go untracked due to blockers, privacy settings, and attribution window limitations.
2. Sales aren’t just on your site. 40% of eCommerce sales happen off-site (Amazon, Walmart, marketplaces), invisible to pixel tracking.
3. Ad spend leakage. 17% of spend goes into channels like audio and TV that pixels can’t measure, leaving major blind spots.
4. Last click is lazy. It oversimplifies the customer journey, rewarding the final step and ignoring the true path to purchase.
5. Duplication is rampant. Multiple platforms often claim credit for the same conversion, inflating perceived ROI.
6. MMM is back. Once considered slow and outdated, modern MMM now runs weekly, giving marketers near real-time budget insights.
7. Unified framework matters. Combining MMM, incrementality testing, and MTA gives the clearest picture of channel performance.
8. Custom > generic. AI-powered custom models now outperform generic off-the-shelf ones—at similar cost—thanks to automation.
9. 36% of ad dollars are wasted. Bad data and poor measurement account for over a third of spend, creating a massive opportunity for optimization.
10. Experimentation wins. The best way to break through opinion vs. opinion debates is to run structured tests and let the data speak.

Chapters

Intro & Sponsor Shoutout (00:01–00:44) Sean introduces the theme: why marketers have been “flying blind” and how better measurement restores vision.

Meet Zeke Camusio (00:44–02:10) Zeke shares his background as an entrepreneur and why he founded Data Speaks after seeing marketers’ frustration with measurement.

The Problem with Pixels (02:10–03:47) Why current attribution methods fall short: off-site sales, blocked conversions, arbitrary attribution windows, and duplication.

The Measurement Framework (04:19–05:38) Zeke explains the three pillars: MMM, incrementality testing, and MTA—and why they work best together.

MMM Misconceptions (05:38–06:25) Why MMM is misunderstood, how it’s evolved, and why it’s valuable in a real-time, digital-first environment.

The Trouble with Last Click (06:25–08:30) Why simplistic attribution models distort ROI and why causal measurement matters more.

Applying MMM & Incrementality (08:30–12:23) How Data Speaks uses experiments and triangulation to reveal true performance and reallocate budgets.

Onboarding & Custom Models (14:16–17:30) What implementation looks like, why labeling data is critical, and how custom models beat generic ones.

Culture Shift: Gut vs. Data (18:25–19:34) The organizational challenge of moving from instinct-driven marketing to data-driven experimentation.

AI, Trust, and the Future (21:39–24:41) Why brands need independent measurement in the age of AI-driven media buying—and how AI enhances, not replaces, human judgment.

Case Study: From $2M to $22M (24:41–25:51) How reallocating spend and applying the framework drove double-digit ROAS gains and 10x revenue growth.

Competitor Landscape & Differentiation (27:45–30:50) Where Data Speaks sits relative to Recast, Measured, Triple Whale, and others—bringing consolidation under one roof.

Mindset Shift: Do You Even Know You Have a Problem? (30:50–32:06) Why most marketers don’t realize their measurement stack is broken until they see the alternative.

Lightning Round (32:06–34:32) Quick-fire Q&A: myths, frameworks, tools, decompression habits, and Zeke’s rugby background.

Closing Thoughts & Call to Action (34:32–end) Where to find Zeke, upcoming product enhancements, and a final reminder: let the data speak.