SSubida Skills

Esc to close · results from all notes, skills and people

Marketing›Ad creative

Ad relevance and the relevance chain — first frame to landing page to homepage

Barry Hott's view that relevance (not Andromeda) decides Meta performance — micro-signals like accent, age or the hero image change who an ad is relevant to; relevance must carry from the top pixels of the ad through the video to the landing page and the homepage people escape to (MrBeast's build-from-the-thumbnail method); and with algorithmic feeds the cold-vs-warm relevance gap keeps widening.

Updated 2026-10-24·4 sources·ad-creativerelevancemeta-adslanding-pageshomepageandromedahooks

Ad relevance and the relevance chain

TL;DR

  • Feeds grow more personally relevant every day. The question each ad must answer is “is this for me, do I belong, is this who I want to be?” One inauthentic or off-target whiff and people leave.1
  • Relevance sits on a multidimensional spectrum: accent, age, skin colour, setting, interests, language. Micro changes shift who an ad reaches.2
  • Build a relevance chain: first frame (the top of the frame scrolls in first) → first 3 seconds → rest of the ad → landing page headline/image → homepage.34
  • Meta optimises on estimated action rate. Make destinations relevant to more of the people who could buy, and delivery improves.5
  • Blaming Andromeda is a cop-out. The “diverse creative” advice is what top advertisers have done since 2024; the problem is usually in your creative, pages or offer.67

Details

Micro-relevance examples. A deep Southern accent vs a thick Boston accent: the Southern one is relevant to far more Americans, Boston to a small north-east pocket.2 On pages, swapping the hero from an older woman to a younger man, changing skin colour or headcount, or adding/removing a person changes who converts. Barry has seen a female hero draw more male attention, so test rather than assume.8

The homepage leaks into ad performance. People escape landing pages to the homepage, Google or the .com no matter what you do. Ask: who is in the homepage hero, does it rotate, and is that person relevant to the audience your ads send?9 At a multi-SKU snack brand, ad and site performance rose and fell with whichever product the homepage banner featured that week. The swing tracked how delicious and broadly appealing the featured product looked.10

Build from the first moment. MrBeast’s team builds each video from the title and thumbnail. Do the same: design the first frame and its top strip, then make sure the relevance carries through to the click and the page.4 It does not have to be one-to-one: an ad can open on one niche problem and broaden so the landing page headline still fits that visitor. If a super-niche one-to-one chain scales, use it.11

Thumbstop is now “relevance building”. Users skip fewer posts than ever because the algorithm already knows them. Ads that aren’t immediately relevant get no delivery or very expensive delivery. Warm audiences already find your brand relevant, so the gap between cold and warm relevance keeps widening — build creative for the cold side.12

Expectation shift. The move from following-based feeds to For You feeds made expected relevance “hockey-stick”. Brands that were great at polished viral content had to change pacing and hooks.13

Disagreements / caveats

  • Barry’s claims come from audits and his own accounts, with no controlled data. He also sells consulting and an agency (AdCrate).
  • Other notes stress a strict one-to-one ad → page message match (Message match between ad and landing page, Landing pages for Meta ads — congruency, listicles and sales pages). Barry accepts that, but says a chain that broadens can work just as well.11
  • A related debate from the founders’ roundtable (speakers are hard to separate in the transcript, likely Ovadia Lebaton vs Barry): one side says Meta’s image processing picks a seed audience at upload, so “how you get off the line matters” and identical ads can stall in one ad set. The other side says real human reactions are what decide scale. Both views are opinion, not data.14

See also

Footnotes

  1. Ad Relevance: What Everyone Still Gets Wrong @ 00:00 — Barry Hott ↩

  2. Ad Relevance: What Everyone Still Gets Wrong @ 00:52 — Barry Hott ↩ ↩2

  3. Ad Relevance: What Everyone Still Gets Wrong @ 05:04 — Barry Hott ↩

  4. Ad Relevance: What Everyone Still Gets Wrong @ 06:09 — Barry Hott ↩ ↩2

  5. Ad Relevance: What Everyone Still Gets Wrong @ 05:04 — Barry Hott ↩

  6. Ad Relevance: What Everyone Still Gets Wrong @ 08:05 — Barry Hott ↩

  7. Ad Relevance: What Everyone Still Gets Wrong @ 09:43 — Barry Hott ↩

  8. Test Landing Pages on Meta THE RIGHT WAY (2026) @ 19:24 — Barry Hott ↩

  9. Ad Relevance: What Everyone Still Gets Wrong @ 02:23 — Barry Hott ↩

  10. Test Landing Pages on Meta THE RIGHT WAY (2026) @ 24:57 — Barry Hott ↩

  11. Ad Relevance: What Everyone Still Gets Wrong @ 06:55 — Barry Hott ↩ ↩2

  12. AI Ads: No Going Back @ 37:38 — Barry Hott ↩

  13. I spoke to 3 Founders about MAKING ADS (feat. Dara Denney) @ 56:51 — Barry Hott ↩

  14. I spoke to 3 Founders about MAKING ADS (feat. Dara Denney) @ 59:55 — Barry Hott, Ovadia Lebaton, Dara Denney ↩