Podcast Summary

Sub Club by RevenueCat

Episode: The Post-Attribution Playbook for Growth — Eric Seufert, Mobile Dev Memo

Released: September 3, 2025 | Hosts: David Barnard, Jacob Eiting | Guest: Eric Seufert

Overview

In this episode, David Barnard chats with Eric Seufert—media strategist, quantitative marketer, and author of Mobile Dev Memo—about the realities of AI in app marketing, the true value and risks of paid growth channel concentration, the paralyses caused by “measurement dysfunction”, and why most common creative and attribution strategies are rapidly losing relevance. Eric draws from his deep operational, entrepreneurial, and advisory experience, contrasting the hype around “AI everything” in marketing with sobering, practical insights that can transform how leading app businesses invest for growth after attribution has become unreliable.

Key Discussion Points and Insights

1. AI in App Growth: What’s Real, What’s Hype

[02:11–18:47]

• Eric emphasizes the need to define “AI” precisely: most platform optimizations are powered by mature machine learning, not by generative AI.
• Real Transformation: Companies that use AI most effectively bake it into every function, not just use it for ad creative or superficial tasks.
• Example: At “Fabulous,” Eric led a bottoms-up approach for AI, letting each team define their own roadmap for AI transformation within guardrails.
• Superficial Uses: Generating hundreds of creative ad variants with GenAI yields minimal incremental value beyond covering basic creative needs.
• What works today: Automating repetitive human tasks (like competitive creative prospecting) and extracting insights from large datasets; not full creative/decision automation.
• Limitations: Generative AI is still unreliable for anything requiring numerical precision, sequence, or high-stakes decisions (hallucinations can undermine results).

Notable Quote:

“If you want to actually embrace this in a transformative way for your company...make sure every single functional team feels enabled to do this and they also feel like they have the agency to define what implementation looks like.”
— Eric Seufert [05:01]

2. Creative Testing and the Futility of Post-Hoc Rationalization

[12:27–18:48]

• The obsession with why certain ads win is misplaced; in an era of black-box algorithms, you rarely have the context to understand what worked.
• Massive creative testing (“feeding the beast”) is more effective than trying to intuit or post-rationalize why an ad succeeded.
• The key process: Automate the production and testing of wild, even random, creative variants, and measure win rates—don’t over-interpret causality.
• “If you could mass experiment like that, you could find [winners]. But why they win is just random. You shouldn’t try to interpret it.” — Eric [13:37]

3. Where AI-Driven Advertising is Headed

[18:10–28:43]

• Both Meta (Facebook) and Google will soon offer near fully-automated, brand-specific, auto-generated creative and on-the-fly landing page optimization based on individual user profiles and past performance.
• Adoption of such tools will be gated by marketer and product manager comfort/trust, not technical limits.
• “Signal engineering” (designing in-product touchpoints to communicate intent and value to ad platforms) will become a major lever, possibly even automated by the platforms.

Notable Quote:

“My theory is that Facebook could deploy a change tomorrow [auto-generating creative at the brand level]… They are slow rolling this because they know that people need to be onboarded and comfortable with it.”
— Eric Seufert [18:48]

4. Risks and Realities of Channel Diversification

[34:28–38:24]

• The urge to diversify away from Meta/Google for its own sake is typically inefficient and costly due to additional creative, data, and operational overhead.
• True need for diversification only exists if channels are at saturation or if an additional channel can produce interaction effects (e.g. boosting performance elsewhere).
• Structural risks (like a new competitor entering) usually impact all channels, undermining the idea that channel diversification alone reduces risk.

Notable Quote:

“Diversifying for the sake of diversifying is often a mistake. You diversify when you’ve reached saturation, or you think there’s an interaction effect.”
— Eric Seufert [36:38]

5. Measurement Dysfunction: The Biggest Pitfall for Growth Teams

[38:33–41:52]

• The most common—and most damaging—mistake is “measurement disorganization”: conflicting measurement tools, misaligned KPIs across teams, lack of clarity about what success looks like.
• Fixing this is people work, not just tech work; requires cross-team coordination and hard conversations to create unified, credible measurement principles.

Notable Quote:

“Your measurement system is the heartbeat of the company...it needs to be credible, but also in a way that everyone understands and appreciates. That serves their different use cases.”
— Eric Seufert [39:13]

6. Balancing Speed versus Accuracy in Early Campaign Optimization

[42:57–45:22]

• Avoid false precision by modeling LTV too early; instead, build repeated short-term ROAS/retention measurements for new cohorts, slowly expanding as you collect more data.
• Kill loser creatives quickly; winners are revealed over time.

7. Missed Growth Opportunities: Measurement Paralysis

[45:35–46:51]

• The constraint on most companies’ growth is lack of trusted, actionable measurement; this stymies efforts to branch into new channels or tactics.
• Robust incrementality-focused measurement is rare but essential for scaling into new opportunities.

8. Overinvestment in GenAI Creative Tools & Attribution Hacks

[50:14–51:10]

• Most teams are pouring excessive resources into GenAI-driven creative tools—increasing quantity, not (meaningfully) increasing effective concepts.
• Likewise, building hacky attribution solutions (especially around Apple’s SKAdNetwork/ATT) is a poor use of resources given diminishing returns and likely policy clampdowns.

Notable Quote:

“Just cranking out, we went from 50 variants a week to 200 or 2000, but they’re all the same concept. There’s no value add there.”
— Eric Seufert [50:58]

9. Philosophical Perspective: Advertising, Data, and D2C/App Ecosystem Growth

[28:44–33:15]

• Hyper-personalized, data-driven advertising uniquely enabled the rise of D2C and, by extension, many profitable “app-first” businesses that wouldn’t have survived in a traditional, broad-based media context.
• Banning personalized ads means “throwing the baby out with the bathwater”; identify and surgically fix problems, don’t destroy foundational enabling tech.
• CLick-through rates remain low, which is evidence of friction (not over-manipulation) in targeted digital ads.

Notable Quotes & Memorable Moments (with Timestamps)

• On the role of AI in creative/marketing:
  “If you want campaigns to be optimized in real time...you’re going to rely on Facebook doing that, or Google. That’s a platform imperative. Leave that heavy lifting in the platforms. You can’t do that, so what can you do?”
  — Eric Seufert [05:18]

• On creative concepting vs. variant generation:
  “What you really care about is the concept, and actually, it's coming up with concepts that you yourself couldn't come up with. The AI is not doing that much more to influence the performance [beyond that].”
  — Eric Seufert [12:56]

• On the future of creative + attribution:
  “Within 18 months, we’re at brand-specific, full-fidelity video ads being generated without prompt, based on past performance.”
  — Eric Seufert [18:48]

• On healthy channel strategy:
  “Diversifying for the sake of diversifying is often a mistake...max out the biggest channel to ROAS threshold, then move to the next.”
  — Eric Seufert [36:38]

• On measurement dysfunction:
  “The most common thing I see is just a very chaotic approach to measurement.”
  — Eric Seufert [38:33]

• On GenAI creative tools:
  “I think you could use off-the-shelf stuff and get almost all the value you’re going to get.”
  — Eric Seufert [50:14]

Timestamps for Important Segments

• [02:11–08:10]: Defining what “AI” actually means for app businesses and marketing.
• [12:27–16:55]: The diminishing returns of creative variant generation and the “randomness” of algorithmic ad success.
• [18:10–22:19]: The trajectory toward fully-automated, individualized ad creative/landing pages powered by platforms.
• [34:28–37:06]: Should you diversify away from Google/Meta? Operational and opportunity cost perspectives.
• [38:33–41:52]: The paralyzing effect of measurement dysfunction inside orgs.
• [42:57–45:35]: Early campaign optimization—how to balance speed vs. “accuracy”.
• [50:14–51:10]: Over-investment in GenAI-driven creative tools and attribution hacks—why these will matter less moving forward.

Takeaways for App Growth Leaders

• Automate where it grants leverage; don’t get stuck increasing variant counts for their own sake.
• Concentrate on robust, universally trusted measurement and incrementality frameworks above channel proliferation.
• Let platforms do the heavy ML lifting—don’t waste resources trying to out-optimize them on their home turf.
• Creative concepting and process innovation outperforms post-hoc rationalization of why any particular ad worked.
• Beware the “data-driven” fallacy: deterministic attribution is getting harder, not easier—accept uncertainty and build adaptive decision-making processes.

Final Thoughts

This conversation provides an unvarnished, deeply practical map for app founders and marketers navigating the post-attribution era: automate judiciously, unite your teams around reliable measurement frameworks, don’t reflexively diversify channels, and stay skeptical of GenAI hype—but remain optimistic about the long-term transformation these tools can unlock for both startups and scale-ups.

For more insights like these, find Eric at Mobile Dev Memo and connect with the Sub Club community on chat.subclub.com.