Perpetual Traffic — Episode Summary

Why Pausing Meta Ads Based on CPA Is a BIG Mistake!

Podcast: Perpetual Traffic
Host: Ralph Burns (Tier 11)
Date: April 3, 2026

Episode Overview

This episode delivers a critical update for digital marketers, brand managers, and paid traffic pros using Meta Ads. Ralph Burns and his team at Tier 11 dive deep into why cutting ads based solely on in-app Cost Per Acquisition (CPA) is outdated — especially under Meta’s evolving Andromeda algorithm and the new GEM (Generative Ads Recommendation Model) update. The show dismantles popular ad management myths and shares advanced strategies, backed by data from over $70 million in spend, for scaling campaigns in today’s hyper-complex ad environment.

Key Discussion Points & Insights

1. Why "Cut Losers, Let Winners Run" Is Outdated

• The old rule of axing underperforming ads based on in-app CPA ignores how Meta’s sequencing actually works post-Andromeda and with GEM.
• Ads that appear to have poor CPA may significantly contribute to conversions elsewhere in the user journey (top, mid, or bottom funnel).

“Cut the losers, let the winners run was already dead in November of 2025 and has been proven wrong many times over since. Still pausing is still true. The new structure matters as well. Still true. Don’t pause ads based upon in-app CPA alone. Yes, look at your in-app CPA but also see how it’s affecting the entire ecosystem.”
— Ralph Burns (07:55)

2. Meta’s Andromeda & GEM: Smarter Sequencing and Attribution

• Andromeda’s hierarchical delivery means users don’t see ads in the same order; conversion actions occur at multiple points.
• GEM, rolled out globally in late 2025, acts as a ‘second brain,’ refining which ads are delivered based on creative granularity (e.g., hook rate, fatigue, audio signals).

“Andromeda now analyzes creative at a far more granular level. Hook quality in the first three seconds or hook rate is an extremely important secondary or really primary metric in many cases…”
— Ralph Burns (05:44)

3. Contribution vs Attribution: Changing How Ad Success Is Measured

• First-click and last-click attribution models are flawed alone; contribution models (using tools like Tier 11 Data Suite or Wicked Reports) assess what impact each creative has on the overall buyer journey.
• Even “loser” ads (by strict CPA) can be essential feeders for mid/bottom-funnel winners.

“The core insight still holds that CPA in app is a poor judge of individual ad performance inside Andromeda’s sequencing model... Not just attribution, but contribution. What does it contribute to? The bottom line.”
— Ralph Burns (04:36, 05:03)

4. Creative Diversification Is Key

• Instead of creative diversification as a standalone strategy, think of it as a system for covering all user journey segments.
• At least 25 unique creatives (especially UGC, founder videos, stories) should run simultaneously for proper audience exposure and algorithm training.
• Different creatives ‘warm up’ different users — UGC is the effective entry point; product-focused ads close.

“Top of funnel ads are usually UGC content... founder videos, stories about the company, something that introduces you to the brand... then middle of funnel ads do the heavy work.”
— Ralph Burns (06:29)

5. Real-world Ad Account Data: Spend, Not CPA, Drives Decisions

• Focusing on spend rather than CPA is a more accurate reflection of Meta’s algorithmic optimization.
• Example discussed: UGC ads generated lots of cold traffic and clicks (not always immediate conversions), while product ads with seemingly higher CPA drove crucial add-to-carts and sales later in the journey.

“Spend model is actually 100% accurate because it’s being spent on. If we look at the spend descending and you look at the CPA...that one’s better...but this is where CPA is a poor judge of performance.”
— John (11:57)

6. Avoid Early Pausing: Let Data Mature

• Too many brands pause ads after a few days of higher CPA, misunderstanding their long-term role in nurturing buyers across the funnel.
• Using advanced attribution tools, the team uncovered that some high-CPA ads were responsible for a major share of new add-to-carts and net new customers.

“In the first week I started hacking off ads that are having low CPA or high CPA... You cannot do it that way. If I didn’t have data suite, I would have killed my middle, which absolutely would have hurt...”
— John (17:59, 18:13)

7. The Feeder Strategy Now for Meta

• Formerly a Google technique, this involves running dedicated campaigns for specific SKUs to train Meta’s algorithm more granularly.
• These developments are so new they weren’t fully recognized even a few months prior — digital marketing is evolving rapidly.

Notable Quotes & Memorable Moments

| Timestamp | Speaker | Quote | |---------------|-------------|-----------| | 04:36 | Ralph Burns | “CPA in app is a poor judge of individual ad performance inside Andromeda’s sequencing model.” | | 05:44 | Ralph Burns | “Andromeda now analyzes creative at a far more granular level. Hook quality in the first three seconds… is extremely important.” | | 06:29 | Ralph Burns | “Top of funnel ads are usually UGC content...middle of funnel ads do the heavy work.” | | 11:57 | John | “Spend model is actually 100% accurate because it’s being spent on. If we look at the spend descending and you look at the CPA... but this is where CPA is a poor judge of performance.” | | 18:13 | John | “You cannot do it that way. If I didn’t have data suite, I would have killed my middle, which absolutely would have hurt.” | | 20:19 | Ralph Burns | “That’s your bottom of funnel ad right there.” |

Important Segments & Timestamps

• [01:48] — Ralph explains what’s changed since the original episode
• [05:00–06:50] — The impact of Andromeda and GEM on ad evaluation
• [11:35–13:30] — Debunking “cut the losers” logic and explaining new sequencing models
• [14:11–18:13] — User journey breakdown: UGC vs product ads; click-to-cart analysis
• [18:13–20:27] — The dangers of pausing too early and critical reflection on attribution data

Actionable Takeaways

• Do NOT judge ad performance solely on in-app CPA, especially post-Andromeda/GEM.
• Adopt a contribution mindset: Use cross-channel attribution to see the real value of each creative.
• Build creative variety: Develop diverse ad sets, especially UGC and sequenced storytelling, for full-funnel coverage.
• Let data mature: Avoid pausing ads early; monitor their downstream effects before acting.
• Utilize advanced attribution tools (e.g., Tier 11 Data Suite, Wicked Reports) for more granular performance insight.

Final Thoughts

Digital marketing is rapidly evolving, and so must your advertising strategy. Where once CPA dictated campaign management, today, understanding contribution, sequencing, and creative diversity is critical for scaling profitably on Meta. Emerging tools and smarter attribution reveal the real performance heroes in your ad account—don’t make the big mistake of pausing them prematurely.

More resources and reference episodes can be found at perpetualtraffic.com.