The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
The account was tracking WhatsApp link clicks as conversions, hiding poor lead quality behind a fake £1.45 cost per lead. I paused Performance Max, rebuilt Search into one streamlined campaign, and replaced the fake tracking with enhanced conversions on real form submissions and Calendly bookings. On a £16.66 daily budget, 4 of the last 5 leads have been qualified — verified directly with the client — with cost per lead (£96.78) now being actively brought down.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Trending
Claude
Claude has entered the design space. How are you using Claude Design?
Contra University
Learn from expert creatives how to earn more using next-gen AI tools.
creativeaiflow
Creative AI workflows are evolving. What tools do you use, and what are their strengths and weaknesses?
freelancerlife
Freelancer life is wins, pivots, and everything in between. What’s yours right now?
Related posts
🚀 Officially Google Ads Search Certified!
Thrilled to share that I’ve successfully earned my Google Ads Search Certification! 🎯
When the Irish College of Traditional and Integrative Medicine came to me for a full Webflow rebuild, the first thing I did wasn't open Figma.
It was open GA4.
Their existing site had real problems and the data was telling the story. Low retention, drop-offs at key pages, forms that weren't converting. Before touching a single design element, we needed to understand why.
Once we had that picture, everything else followed. Copy restructured around what visitors actually needed to see. Forms rebuilt for clarity. Layout decisions made with evidence, not guesswork.
The result was a site that works harder for them and their students.
That's what I mean when I say web design is strategy first. The visuals are just how the strategy shows up.
See the full project here.
Okay so I need to talk about how I'm making UGC ads now, because it's honestly kind of ridiculous how easy this has gotten.
If you've ever tried to get UGC ads made, you know the pain. Booking actors, briefing agencies, waiting days, paying way more than you wanted to. I used to think that was just... the cost of doing business. Turns out it's not anymore.
Here's what I've been doing instead, the whole thing runs inside one Claude conversation. No creative tools, agency, or ten tabs open at once. Just two connections: Meta and Arcads.
Let me break down how it works.
Step 1: Research first, always
Before generating a single ad, I use Claude and other Premium tools to pull my top performing ads from the last 30 days, then search the web for what's currently working best in my product category. What creative formats are getting impressions, what hooks are driving engagement, and what patterns keep showing up in winning ads.
And this is the step almost everyone skips. It's exactly why so many AI-generated ads flop, people jump straight to generation with zero signal on what actually converts. Research first means your first test isn't a guess. It's informed.
Step 2: Turn research into scripts
Once I've got the research, I hand it to Claude with a specific brief: act as a senior performance marketer, here's my product, here's who it's for, here's the platform. Then I ask for scripts built around real pain points, hook, problem, solution, proof, CTA, all timed out. And I get Claude to score each one and only keep the ones that actually hit.
This part used to take me hours of back and forth with a copywriter. Now it's minutes.
Step 3: Generate the actual creative
So this is where Arcads comes in, it's basically got every major AI video/image model in one place, so I'm not paying for five different subscriptions. I generate an actor image first (with my product reference attached so nothing gets weird or off-brand), approve it, then turn it into a video with the script from step 2 dropped straight in.
Small tip that saved me a lot of wasted spend: always fix issues at the image stage, before it becomes a video. It's way cheaper, and way faster.
Step 4: Launch without leaving the chat
Once the videos are ready, I'll upload them straight to my Meta account through the same conversation. Then I just tell Claude what I want, campaign objective, budget, targeting, in plain language. No Ads Manager or manual setup.
And here's the part that actually changes how you test: after a few days, I just ask Claude which ad is performing best and to show me why. It pulls the live data and explains it in plain English. No exporting reports, no spreadsheets.
Step 5: Double down
Once I know the winner, I ask for more variations of it, same actor, new scripts and the loop just keeps going. Every single round gets a little smarter than the last.
The honest truth is this doesn't just save time, it changes the economics entirely. You're not paying agency rates, you're not waiting on turnaround times, and you're testing with actual data. If you're not at a scale where an agency makes sense yet, this is genuinely one of the best ways to get high output without needing a big team or a big budget.
Don't forget to hit follow, I'm breaking down a new ad strategy or workflow every week, and you don't want this to be the last one you catch. 👀
