Webinar Summary
Ed Rudland from MyoMaster walked through two use cases.
The first use case: automated attribution and reporting. MyoMaster sells high-AOV products (saunas, ice baths, chillers) where the buying decision often stretches well past the standard 30-day attribution window, making it hard to tell which ads were actually driving sales. Ed built a Claude setup fed by meticulously labeled Meta and Google Ads data pulled in through Supermetrics, plus margins, target ROAS, and specific signals to watch for, such as not judging an ad until it has hit a set spend or click threshold. Reports run automatically every Monday and Thursday, catching creative fatigue and identifying high-quality traffic (email signups, returning visitors) days or weeks faster than before. That visibility let the team keep scaling ads that looked mediocre on ROAS alone but were actually feeding longer-consideration, high-intent buyers who purchase six weeks later.
The second use case: creative and media briefing at scale. Ed feeds Claude monthly forecasts, budgets, and the prior month's ad library context to generate a full briefing: how many statics, ambassador ads, and videos are needed each week, mapped against the trading calendar, personas, and pain points. What used to take him about two days now takes minutes, freeing the four-person team to output roughly 155 to 180 pieces of creative a month. The same engine extends to briefing MyoMaster's 30-plus brand ambassadors, matching detailed bios and content styles to campaign pain points, a task that used to be a heavy lift and now takes about 20 minutes.
A recurring theme was that context is the foundation. Ed's advice for getting trustworthy output: build detailed naming conventions so Claude knows exactly what's in every ad, feed in margins, targets, and known thresholds, and keep refining that context over time rather than relying on one-off prompts.
Human judgment still sits at the center. Claude can profile ambassadors and match them to briefs, but building the actual relationships with creators is still on the team. Ed also cautioned against reacting to every signal too fast: since ad algorithms don't reward constant changes, the model is there to help spot patterns early, but decisions on when to act are still made deliberately.
Key takeaway: the real gain isn't just faster reporting, it's what Ed called "complexity at scale." A four-person team can now run really complex, high-level analysis quickly enough to operate like a much bigger one, and the results (one of MyoMaster's best months on record) speak for themselves.
Ready to try Claude for your own reporting and campaign optimization?
