By Roi Shivek
About this series
This two-part essay uses public statements by Gymshark employees and industry data to explore a constraint facing scaled DTC sporting-goods brands.
*Gymshark is not (yet) a Churney customer; quotes are cited for narrative and category insight, not endorsement.
In conversations with scaled ecommerce brands, we often see the same split. The business has fixed how it measures performance, while Meta, Google, and TikTok are still optimising their campaigns on broken signals. Usually that means day 1 conversions, and not customer value that shows up in repeat, loyalty, or returns weeks later.
That pattern is why Daniel Green’s Charlie Oscar essay, Life Beyond Ads Manager, caught our attention. Gymshark has been unusually open about how a scaled brand moves past Ads Manager, with enough detail to learn from, not just a headline. We use their public story as the thread through a two-part series.
FY25 brought £646m in sales and a thirteenth year of growth. Green, Head of Digital Marketing, has told that arc in detail. It started with a phase when spend followed a simple rule.
“There isn’t a budget. As long as the CPA is good, keep spending.” (Daniel Green, Head of Digital Marketing, Gymshark, Charlie Oscar, Life Beyond Ads Manager)
That works until it doesn’t. Green describes a period when the team treated Meta’s CPA in Ads Manager as the source of truth, which is reasonable when one channel dominates, and less reasonable when you’re running full-funnel, multi-channel marketing.
In the same essay, he names the failure modes that pushed Gymshark towards MMM and MTA:
“Marketing teams simply cannot rely on Ads Manager as their source of truth when running multi-channel, full-funnel marketing campaigns ... If accounts are using a 7-day conversion window, you may only be seeing half the story ... Spend becomes misplaced, and overall investment becomes less effective.” (Daniel Green, Charlie Oscar, Life Beyond Ads Manager)
Green’s public account is not a knock on Meta. It is the back half of the maturity arc sketched above: graduate from dashboard CPA, add MMM and MTA, then judge channels on incrementality rather than who shouted loudest in a seven-day window. As he puts it, MMM guides investment decisions; MTA helps implement them. That stack answers a question marketers care about: Where should we put money, and what actually drove incremental growth? It does not, however, answer a different question: What event is the platform’s algorithm optimising towards inside the ad account?
Meta, Google, and TikTok still optimise inside their own windows, on the events you feed them. These are often first-order purchase or CPA proxies, updated on a cadence that does not wait for repeat orders, loyalty behaviours, or returns to mature. The gap between knowing and teaching the platform is where this story turns.
Sporting goods and apparel sit in a category where value forms late: repeat purchase, loyalty engagement, and net revenue over time, not checkout alone.
Apparel benchmarks on Shopify-heavy samples show 365-day repeat rates in the 17-32% range. Activewear and community-driven brands often sit towards the higher end of apparel, but still well below consumables where replenishment drives 40%+ repeat. The timing gap is the issue: in the same benchmarks, only about 30% of second orders happen in the first 30 days (in a 365-day window); the rest land months later, while a meaningful share of platform learning still happens inside that window.
Scaled brands also track leading indicators that signal long-term value in first-party data (workouts completed, loyalty tiers, email and SMS engagement) but do not, by default, feed what Meta or Google bid algorithms learn on.
Two customers can look identical on day one: same CPA, same AOV, same campaign. One becomes a repeat buyer across three drop cycles. One never buys again. A standard purchase conversion treats them the same unless you teach the platform otherwise. Returns add a second correction on the same clock: industry data puts 19.3% of online sales returned in 2025, often weeks after conversion, and NRF’s 2025 returns research notes shoppers aged 18-30 averaged 7.7 online returns in the prior twelve months on ~30.4 total online purchases (~25% of orders). That matters for activewear, where audiences skew young and drop culture is part of the commercial model: the checkout win and the P&L truth can disagree.
Gymshark’s public narrative shows they understand this at the measurement layer. Green has described channels like TikTok looking “average” in Ads Manager while remaining a strategic growth opportunity: incremental impact that only becomes visible once you move beyond dashboard defaults (Charlie Oscar).
Noy Rotbart, CEO of Churney, sees this across scaled ecommerce: about 36 hours after a first purchase, the same buyer’s later repeat and engagement events mostly update measurement, not bidding; refunds rarely retrain the algorithm either, unless you send a stronger value signal up front.
The open question is the bid layer: not “do we know?” but “what are we asking Meta and Google to learn on?”
Here’s the distinction that matters for teams at Gymshark’s stage:
Green has been explicit that MMM “cannot be used to optimise campaigns weekly inside Ads Manager” (Charlie Oscar). MMM was never meant to be a daily bid engine.
But performance teams still live inside ad accounts every week. When every channel claims credit in overlapping windows, dashboard truth and bid events diverge. Green’s LinkedIn recap of moving beyond Ads Manager (“real incremental impact,” “reduce channel bias,” “smarter investment decisions”) describes where marketers make investment calls.
Our read from public sources: solving measurement for the marketing org does not automatically upgrade what Meta, Google, or TikTok optimise towards. The platform may still be learning on first-order purchase value while your team is making decisions on incrementality and multi-touch truth.
That pattern shows up beyond Gymshark. In fashion, operators have described purchase pixels celebrating checkout wins that finance teams know will shrink after returns (Google Think, Bestseller), a reason net-value and lifecycle signals are entering the conversation for scaled apparel brands.
Gymshark has not publicly described user-level pLTV or value-based bidding in the sources we reviewed. The next constraint for brands at this maturity level is often signal design, not another reporting tab.
Gymshark didn’t stop at measurement maturity.
Carly Natalizia’s promotion to Chief Commercial Officer frames the next chapter: becoming a “wholly omnichannel brand” without “losing that critical digital experience” (TheIndustry.fashion). Stores in London, Dubai, and New York. A loyalty program that rewards workouts and email signups, not just transactions. Drop culture, collabs, and promo windows that make short-term ROAS charts look excellent while real customer value is still forming.
Each of those moves is strategically right for a hundred-year brand. Each also adds noise to what a seven-day conversion window can see.
Part 2 of this series will follow that chapter: drop-week ROAS, new-customer vs repeat value, and what it takes to teach platforms long-term value once measurement is no longer the bottleneck.
This two-part essay uses public statements by Gymshark employees and industry data to explore a constraint facing scaled DTC sporting-goods brands. Gymshark is not yet a Churney customer; quotes are cited for narrative and category insight, not endorsement.
Your data warehouse has incredible value. Our causal AI helps unlock it.
