Targeting decisions used to determine campaign performance. In 2026, with Advantage+ shopping and audience consolidation, the primary lever is no longer who you show ads to — it is what you show them, how often you refresh it, and how disciplined you are about killing what stops working. The Creative Iteration Protocol is the system we run inside every active client account to keep creative output ahead of fatigue.
Why iteration beats optimization
A campaign with mediocre creative cannot be optimized into a great campaign. Meta's algorithm rewards engagement signals — thumbstop rate, video hold, click-through behaviour — that are produced almost entirely by the creative itself. When agencies talk about "scaling," they usually mean increasing budget against an existing creative. That works for two weeks, then ROAS collapses.
What actually scales fashion accounts past the €100K/month ceiling is a pipeline that produces fresh creative faster than the audience burns through it. The protocol below is how we keep that pipeline running on a 14-day cadence per account.
The 3-2-1 rule
Every iteration cycle inside a STRIDE account follows the 3-2-1 rule:
- Three distinct angles per cycle — different problems, identities, or desires being addressed.
- Two hooks per angle — variations on the first three seconds of the ad.
- One winner per cycle moves into evergreen rotation.
Six pieces of creative tested. One ascends. Five retired. That ratio is non-negotiable — most accounts ship too few angles per cycle and end up over-fitting on what worked last quarter.
The four phases
Phase 1 — Brief
Every creative starts with a brief that names: the desired action, the audience's current belief, the belief shift required, and the visual register. No brief, no production. This is where most agencies fail — they describe what they want the ad to look like instead of what it needs to do.
Phase 2 — Production
Six pieces shot or designed within seven calendar days. We work with a hybrid model: studio shoots for hero creative, UGC partners for social-native variants, and motion designers for the static-into-motion derivatives. The output gets reviewed against the brief — not against personal taste.
Phase 3 — Test
All six pieces launch into a single isolated test budget — typically 1× CPA per day per creative. Three days minimum, seven days maximum. The signal we care about is not just ROAS — it is the combination of CTR, video hold, and add-to-cart rate. Each signal tells a different story about which part of the funnel the creative is breaking.
Phase 4 — Retire
The retire step is where 90% of agencies fail. They keep tired creative running because it "still works" — meaning it produces revenue but at declining margin. By the time the account dashboard shows the decline, the algorithm has already deprioritized the asset and the audience has stopped responding. We retire creative based on leading indicators (frequency, CPM increase, thumbstop decline) — not lagging ones.
“The retire step is where 90% of agencies fail. They keep tired creative running because it "still works" — meaning it produces revenue but at declining margin.”
What this looks like inside a client account
On a typical fashion D2C account doing €150K/month in Meta revenue, the protocol produces roughly 12 new pieces of creative per month, 2 evergreen winners per month, and a creative library of 30–40 active assets at any time. The cycle runs continuously — there is no campaign launch, only campaign refresh.
The protocol is the same whether the brand is selling streetwear in Berlin or swimwear in Munich. The angles change. The framework does not.