Valentin Radu

⛏️ CEO & Founder @ Omniconvert
🎬 Instructor at CVO Academy
🎙️ Keynote Speaker
📕 Book Author: The CLV Revolution

[Valentin Radu @ValentinRadu]

Claude killed the entire Shopify ad workflow.
And a need for a full-time CFO
Or merchandise expert. Claim $300 in credits here:
https://t.co/dgUtsewGdA

[Valentin Radu @ValentinRadu]

@markobalazic Main image first. It's likely what shoppers see first on Meta's product page too, and it leads your PDP, so a change there is easiest to read. Swap the whole set and you learn something moved, not what. Which wins more often as your main image: clean product shot or in use?

[Valentin Radu @ValentinRadu]

From today, Meta adds product browsing to existing ads unless you opted out.

A tap may now open a Meta product page before your own. That can change what your PDP tests and attribution see.

Hold out some ads and judge revenue per visitor, not Meta's conversion count.

[Valentin Radu @ValentinRadu]

@peeplaja Fiction is what you get from a workshop. Personas hold up when they start in the order data: group customers by first product and how fast they reorder, then interview the best group. Same slide format, very different people on it.

[Valentin Radu @ValentinRadu]

@harleyf Compounding shows up in testing too. AI shoppers in SimGym can flag where people get stuck before a change goes live, then a real test on revenue per visitor makes the call. I'm curious how often the simulated and live results end up agreeing.

[Valentin Radu @ValentinRadu]

@ericosiu Lost clients are the real scoreboard there. Output per head is easy to inflate. Renewals aren't. If AI made the work better, it shows up in who stays. I'd want every 'AI did this' claim reported next to client churn for the same months.

[Valentin Radu @ValentinRadu]

More creative won't fix Meta if it all says the same thing.

Of 122,574 ads we classified:
• 82.6% use the same 3 angles
• 50.8% of hooks open with a curiosity gap
• 1.6% use comparison

100 more lookalike ads buys nothing. Find the angle nobody in your category owns.

[Valentin Radu @ValentinRadu]

Hi all!
Is anyone working with DTC Shopify merchants running Meta ads for them?
I'd need a quick favor + have an opportunity, as well

[Valentin Radu @ValentinRadu]

@ecommilan Bravo!
I wrote a whole book about leveraging cohorts with RFM in ecomm

[Valentin Radu @ValentinRadu]

@OisinO Agreed. Features win the install, month-three retention decides whether they stay.

[Valentin Radu @ValentinRadu]

@OfficialOluseyi True, a bigger daily budget gets you to the conversions Meta needs faster. I'd just check the CPA holds once it's out, a quick exit at a CPA you can't afford doesn't help much.

[Valentin Radu @ValentinRadu]

@ecomnotes_ Green is green. At this spend I'd watch the 7-day number more than any single day, since one order flips the day either way.

[Valentin Radu @ValentinRadu]

@ecomedyy Margins holding while ROAS dips is the healthy version of scaling. Watch new-customer share as spend grows, that's usually what slips first.

[Valentin Radu @ValentinRadu]

@ecommozart Holding spend while you're profitable isn't losing momentum. Use the stable stretch to find the next angle; growth when Meta settles will be cheaper than growth forced now.

[Valentin Radu @ValentinRadu]

@GlobalSola Agree. Test the offer now too, not just the creative. A winning ad on an offer that leaves no margin at Black Friday CPMs is still a losing week.

[Valentin Radu @ValentinRadu]

@murodovsultan Agree with where this lands. If Meta generates the creative, the edge moves to what it can't generate: product, margins, retention and the customer insight you feed it.

[Valentin Radu @ValentinRadu]

@linafahizul This. One check before the next brief: CVR by landing page, new vs returning. If new visitors convert at half the rate of returning ones, the fix is the page or offer, not the hook.

[Valentin Radu @ValentinRadu]

@lazeskid Can work if you set the guardrails first: break-even CPA from your margin, a budget cap, and a weekly check of which ads brought new customers. AI optimizes for what you tell it, not for profit.

[Valentin Radu @ValentinRadu]

@CEO_Vlad The 'Your Customers' Words' folder does most of the work. Reviews and support tickets give you the angles, objections and exact phrasing before any avatar or format choice. That review-to-ad step is what we just launched in Commander.

[Valentin Radu @ValentinRadu]

@anzedetn Agree. Writing the hypothesis before launch is what makes a loss useful: 'the claim isn't believable' is a decision, 'this ad lost' is not.

[Valentin Radu @ValentinRadu]

@alex_momeni Point 1 is the one I'd fix first. When the page repeats the ad's promise, colder traffic doesn't have to re-learn why they clicked. A page per ad is part of Omniconvert Commander, just launched: https://t.co/dgUtsexe38