Ryan

@benradack @binghott What's your workflow?

@binghott been playing with this. the speed from concept to live ad is genuinely different now.

Both Meta and Higgsfield MCP's launched this week. Meta -> Claude -> Higgsfield -> Claude -> Meta is now suuuuuper easy 😅😬 Great for both automations and for one-shotting new ads. I think advertisers are about to see a ton of volatility, especially for the brands who don't (or can't) keep up...

@PhilKiel What was the structural change?

We have a brand with spend +100% L7 vs same period last year and their CPMr is down 15%, new visitor percentage is up. This improvement is driven strictly by a new campaign structure and exclusions, no new creative has been launched. Importantly this is reversing the YTD trend which has seen CPMr, net new % going in the wrong direction. Sounds great right? The big win: they’re continuously beating their best weekly sales and profit performance outside of promo. We can also see the impact on their Google trends. But tell me again how media buying is dead 🤔 Q4 is going absolutely rip for these guys, we’ve essentially got 6 months to build audience buckets at a cheaper cost and with higher ad spend.

@jforjacob @nkecom What's your graveyard set up?

@nkecom I used to religiously but was able to break through spend plateaus moving to LC Only cost control I run now is on graveyard campaign

I had an ad that was spending $1000/day at a 0.8 ROAS. All the soft metrics said it was a shit ad. The CPMs on it were 2x higher than the account average. CPC & CTR metrics were shit too. If you had a “testing” campaign - this ad wouldn’t have made it out. But one day, after about 6 weeks of spend, the ad miraculously started printing at a 2.2+ ROAS overnight. It legitimately became one of our best ads. There is no explanation to why this happened & it would have been impossible for any human to predict…. but Meta knew. There’s a lesson in there.

@TJ_Bongiorno Heart & Soil's - Grass-Fed Whey Chocolate Sea Salt

What is the best tasting supplement powder you’ve ever had? Can be anything in powder form - electrolytes, pre workout, protein, whatever

@ronaldo_delduca @TJ_Bongiorno @johnhickey1970 @jforjacob @Ivan778899000 @chuckiegregory For the graveyard campaign you're just running HV?

@TJ_Bongiorno @RyanDTC @johnhickey1970 @jforjacob @Ivan778899000 @chuckiegregory Rawdogging it >

Hit me with the meta ad account structure that's currently working for you right now. Free UndrDog shirt to someone in comments.

@TJ_Bongiorno @johnhickey1970 @jforjacob @Ivan778899000 @chuckiegregory @ronaldo_delduca How strict are you BC?

@johnhickey1970 @jforjacob @Ivan778899000 @chuckiegregory We’re running this successfully right now too @ronaldo_delduca

@Aazarshad @alexgoughcooper How can you do this with Atria?

@alexgoughcooper Atria can also do this but the problem is still to make a thoughtful hypothesis and iteration.

Claude cowork can replace your entire ad research process. I asked it do creative strategy for Huel, and in 5 minutes it: > Analyzed Huel’s website > Looked at the ad library to see what’s already being tested > Looked at amazon reviews to see what people like/don’t like > Looked at competitor reviews to see what people like/don’t like > Scrolled TikTok and found viral TikToks we can recreate > Summarised everything into a 5-page document with fully fleshed out concept ideas that we can test today (and the copy was actually good) And all of this with me just watching it operate my screen. I’m in shock.

@rokhladnik @Meta Is LPV an event people optimize for?

Official Announcement 🚨 Herrmann & Hladnik Digital can confirm that the iOS PageView attribution issue that started on Jan 5th is now fixed. @Meta resolved the issue on Jan 13th, 2025 at 1 PM PST. No official dev comms — but attribution is back, and data loss is resolved. Cost per Landing Page View should start decreasing, and traffic quality should stabilize over the next few days. Big thanks to my partner in crime @herrmanndigital for all the effort 🤝

@herrmanndigital How are you optimizing them?

2025 was the year of learning and trying to grasp AI in marketing. 2026 is the year of taking those learnings and putting them in hyper drive for efficiency and better execution. My Google Gems are all optimized for each client and becoming our got-to assistant. It's unreal the power of Gemini has now and how I don't know what I'd do without them.

@alexgoughcooper Do you use AI to remake or real person?

> TikTok in our niche goes viral > Put the transcript through Claude and get it to recreate it with our product > We remake the video > It's now the top performer L30D Go and scroll the organic feed.

@oliverwhudson Is that screen shot from a deck you can share?

I wanted to share a key insight from a creative workshop we ran with Meta this week. The focus was on scaling under Andromeda and how creative similarity now impacts reach, cost and fatigue. Meta now treats two ads as “identical” if they share the similar first 3 seconds.. even if the copy, message or CTA is different. What that means practically: - Same creator + same setting = same ad - Slight copy variations isn’t recognised as a new concept - Visual similarity overrides narrative difference The big shift: Meta now rewards creative built around personas, not platforms. Different motivators unlock new audiences 89% of the time. It confirms what we’ve seen inside accounts over the past 6 months: Persona-based ads perform better and last longer. Because they're not just re-skins, they're built to resonate with new types of people. So what does that mean for iteration? It’s not enough to just test copy and overlays anymore. We’ve started evolving our testing and creative strategy in a few key ways: ✅ Push bigger visual swings after a winner emerges – Once we’ve got a winning format, the instinct is to go incremental – But now we’re rethinking scenes, locations, setups entirely to escape similarity penalties ✅ Change the hook to change the audience – First 3 seconds = new ad in Meta’s eyes – Use this to tap into new motivators without remaking the whole asset ✅ Recycle winning ads with fresh b-roll – Swapping in new b-roll gives it a second life with fresh delivery ✅ Layer persona insights directly into creative briefs – We’re now briefing hooks and styles based on demographic and psychographic differences, not just product benefits If your ads are sitting left of centre on this chart, you're probably already feeling the ceiling. Meta doesn’t just want more ads. It wants more variation and it’s getting ruthless about what counts.

@itslukeaustin Can I get this in our Statlas?

What is the exact number of ads you need to create each month to achieve your forecast? Enter Creative Demand Model. These are the two brands that represent the lowest and highest Creative Scores we’ve ever seen... And these are the 5 metrics that inform Creative Score 👇

@TaylorHoliday we currently measure media spend against .com outcomes even though we also sell on Amazon. How can we account for both distribution channels when measuring media performance impact? Prime Day significantly exposed this blind spot.

@ritzfit @_redvelvet_80 How did you find out this was happening?

Someone ripped our homepage and is scamming people. FYI @_redvelvet_80 you may be getting impersonated too. See the footer in the scam website and the support email. It says support@primalqueens.us scam url: us.carnivoresnaxs.store/shop/collectio…

@DaveRekuc @binghott @SamMendelsohn4 @ron_ecomm What % of campaign spend goes to the engaged adset?

@binghott @SamMendelsohn4 @ron_ecomm 3 ways I've seen successful. 1. Bid caps instead of cost caps. Better for heterogeneous audiences. 2. ASC with capped % 3. Break them out into their own ad set

I just audited an account working with a well-known agency. Their Meta dashboard looked great. ROAS was solid. Agency was happy. But the business wasn't growing. Here's what I found: Spending ~25% of budget on existing customers Getting ~40% of sales FROM existing customers This is a massive red flag most marketers miss. Meta was being lazy and so is the agency. The existing customers are being targeted in the "prospecting". When you don't exclude recent visitors and existing customers from prospecting campaigns, you're telling Meta: "Go ahead, take the easy route." And Meta will absolutely take that invitation. Instead of finding NEW customers, it keeps targeting people who already know your brand. Your "prospecting" ads aren't actually prospecting and you never learn what works for cold audiences. Cost caps are based on easy warm conversions and never open up enough to target enough colder users. Growth stagnates. The fix is simple: Exclude recent website visitors and existing customers from prospecting campaigns. Yes, your ROAS might look "worse" initially. But here's what actually happens: ✅ You learn what creative works for cold audiences ✅ Meta finds genuinely new customers ✅ Your business grows instead of recycling ✅ You build a foundation for real scale Platform data always makes the platform look good. If your ROAS looks great but your business isn't growing, you're probably just getting efficient at selling to people who were already going to buy. That's just expensive remarketing disguised as prospecting. Stop letting Meta (and your agency) be lazy. Set boundaries.

@binghott @growthzacks @ron_ecomm How long then?

@growthzacks @ron_ecomm 180 day visitors is excessive

@ashvinmelwani How are you able to tie performance to this? Seeing the same notification in our settings

This is happening to a lot of brands it seems like. Received a ton of DMs. It's crazy that there is literally 0 support offered from Meta for this. Complete black box.... until its not.

@TaylorHoliday This is growth marketing

Brands: “we can’t run it, it looks too AI” Meanwhile Boom Boom on Tik Tok moving incredibly fast with no boundaries eating everyone’s lunch. Why do we feel the need to determine whether something looking ai is good or bad? Where did we develop our preconceived negative connotation?

@ron_ecomm @jforjacob It could, shipping expectations vary wildly across regions.

@jforjacob do you think it will work for a state/province level? this is a super interesting strategy

pull Shopify report for sessions by location > add checkout conversion rate metric > identify markets with high sessions but high rate of dropping off at checkout > check order tracking app and filter for that country > look up last mile tracking to see which courier is being used > create shipping profile in Shopify for that country > change shipping rate descriptions to mention the local courier e.g "Canada Post Tracked Shipping" > check back on checkout CVR in 7 days > come back to this tweet and thank me

Instead of building prompt libraries, build custom agents around the tasks those prompts are meant to accomplish. Invest in context libraries for those agents so they can execute better.