Craig 🪄

I’m convinced the only training Meta’s support gets is in ragebaiting

@ami_ads structure

this one tiny switch saved us $8,114,896 in 2025 alone. this year it will probably be even more we see it every day with the affiliates that we work with offers are good creatives are ready to convert funnels are designed to do their thing but still they are not able to get their campaigns profitable 90% of the affiliates neglect the importance of their FB structure. relying on structures that doesnt allow them to scale aggressively structures that have low reputation on meta which leaves them with 40% higher cpms then their competitors structures that in case they face bans, they need to start over completely and lose all their pixel data we have been a/b testing ad account structures since 2025 and the difference it makes for affiliates is insane. looking at 2025 alone we managed to get an average of 41% lower cpms, which allowed us to scale more aggressively then our competitors simply due to the fact we were able to outbid them on the fb auction if you are interested in copying the exact setup that we have used to spend over $21,184,999 comment “structure” and ill be able to introduce 10 serious spenders on our structure. (must like + follow)

Now this is the shit youtube.com/watch?v=1xx4eY…

@vlelyavin @claudeai I’m so happy in this timeline being a couple months behind is late. Always something cool happening

@claudeai openclaw users have had this for months but it's cool to see claude code catching up lol ngl the "start a task and check from your phone" workflow is genuinely addictive once you try it

New in Claude Code: Remote Control. Kick off a task in your terminal and pick it up from your phone while you take a walk or join a meeting. Claude keeps running on your machine, and you can control the session from the Claude app or claude.ai/code

Every company is just a tech company now

@helloitsolly @SenjaHQ Your solution looks really cool

PSA: You do not own your TrustPilot reviews! If you add them to your site without paying they will send you a cease and desist You are sending your most powerful marketing asset - customer stories - to someone else This is why we built @SenjaHQ to help you collect your own customer stories and use them to sell more

day 3, maybe onto something?

I made an app and am getting $6 paid signups on day 0, brb changing my bio

@jforjacob Totally agree. Also meta performance is always noticeably worse when testing for me, even if the test is a “winner”, algo just doesn’t know what to expect

Something to think about when testing new pages that I see nobody talk about Anyone who runs more than one page from their ad account will know that you can run the exact same ads at the exact same budget to two different pages and one will have higher, CPM’s and CPC’s When you split test pages using split testing tools this isn’t accounted for Ok great, you had a 10% lift in rev per visitor but then when you roll it out across your ad account your CPM goes up 20% This is just one of many reasoning believe split testing pages from the ad account is the best way to do things

I wish Google’s ai winning streak extended to their analytics and maybe not needing a rocket surgery degree to set up GTM

@danieldalen Any in the UK? Easiest country imo

Got a few overseas clients that owe me around $250,000 USD. Anyone got any good debt collectors or any other tips?

> ad delivery disruptions > start running 5 orders a minute without changing anything

Why are email pop up apps over $100 a month

Has anyone noticed that Shop Pay appears to leak the user’s last checkout (or attempted) in the referral url?

@jameshyujinlee @RomanEcom Awesome! I def will! :)

@ecomhonesty @RomanEcom Yes sir! I live in KL! :) HMU whenever you around! Dinners on me!

You can get all of this in HK for a fraction of the price of NYC or LDN I’d add the following to the list: - World class international schools for your kids - Driver if you need one - living in a place that’s safe - living in a place with free / very affordable healthcare for all - living in a place with real laws around crime and drugs

@karri_tweets Idea: increase effort to create the ai response slightly. Clicking to reply and sometimes the response is too fast to press S to skip, especially when wanting to use a macro instead. idk how impactful it’d be but

LLM costs were last month 40% of Commslayer's MRR, even after pretty heavy optimisations (can't wait for self hosting to be viable) software is a low gross margin business now, good luck to incumbents trying to protect the 85% margins

@val_shah @taboola And you have the call with the team to be allowed onto the “premium” publishers?

How to scale your ads on @taboola from $300/day to $30,000/day (without burning cash). Most marketers fail at scaling because they rely on "hype." Real pros rely on data physics. Here is the 5-step framework to unlimited scale 🧵👇 1. The "Mandatory Minimum" ($300/Day) Stop launching campaigns at $50/day. You are feeding the algorithm noise, not signal. You need at least $300/day to exit the learning phase. • Below this? You compete for low-quality inventory. • At this level? The AI gets the data density it needs to optimize. 2. The Scale-Up Indicator When should you increase the budget? ❌ When you "feel good" about the creative. ✅ When you hit your Target CPA. If it works at $500/day, the algorithm can usually replicate that success at $5,000/day. Scale is just letting the model do its job on a bigger playground. 3. Bidding Strategy: The Phase Shift Your bid strategy must match your growth phase. • Scaling Phase (Early/Mid): Use Max Conversions. Let the AI explore aggressively and find pockets of opportunity. • Stabilization Phase (High Spend): Switch to Target CPA. Lock in profitability once you have enough volume. 4. Respect the "Saturation Plateau" Scale is not linear. In the US, saturation often hits between $20k - $30k/day. In smaller markets (like Spain/France), it might be $5k - $10k/day. When CPA rises and returns diminish, do NOT push harder. Stabilize the budget to find your "Maximum Efficient Daily Spend" (MEDS). 5. The "Parachute" Rule Automation is great, but always keep one manual safety net. Set a rule: If CPA triples -> Pause Campaign. This isn't for optimization; it's for survival. It saves you from broken tracking, site outages, or runaway spend. TL;DR Summary for Scaling: 1. Start at $300/day minimum. 2. Scale only when CPA is stable. 3. Use Max Conv to grow, Target CPA to maintain. 4. Stop pushing when you hit market saturation. 5. Use a "Parachute" rule to kill bad days instantly. taboola.com/marketing-hub/…

@jameshyujinlee @RomanEcom KL? Planning on visiting later this year!

@RomanEcom That's why I've decided to stay in Malaysia. It's been 7 years and can't imagine paying to pay double or triple for basic necessities.

Lately I've been trying blind testing. No A/B. I just make the change, wait a few weeks, export the data, and ask Claude & Gemini to guess when the change was made. If they guess correctly, it was probably impactful. If not, it probably wasn’t.

@nadooshjas in Paris I pay 8 euro for a glass of water

In Paris you can get a coffee and croissant for €2, in London it’ll cost you your first born child and your soul