The four-week cycle behind sustained creative throughput, broken down phase by phase.
People ask us how we produce volume without sacrificing quality. The candid answer is that volume and quality aren’t a trade-off when the underlying system is built right. They are products of the same disciplined process. The Maple Media Creative Lab is where that process lives.
The Lab isn’t a department. It’s a methodology, run by the same pod that owns strategy and media for the account. It operates on a four-week cycle with four phases: research, ideation, production, read.
Phase 1. Research, week one.
Week one is research, regardless of how many cycles we have already run for the account. Buyer language drifts. The Reddit threads active six months ago aren’t the same threads active now. Amazon reviews accumulate. Competitor angle libraries evolve. The brand that researched once at onboarding is the brand whose creative slowly stops resonating, anchored to an outdated picture of the buyer.
Research week pulls fresh inputs from five sources. Customer reviews on the brand and on competitors, including the long ones and the negative ones, which carry disproportionate signal. Reddit threads in adjacent and on-topic subreddits. YouTube comments on educational and review content in the category. Customer service transcripts and post-purchase surveys from the brand’s own customer base. Competitor ad libraries to identify which angles are getting heavy spend.
Output: a refreshed avatar map, an updated angle library, and a documented list of phrases and emotional triggers that have shifted since the last cycle.
Phase 2. Ideation, week two.
Ideation is structured, not freeform. The team takes the refreshed angle library and matches each high-priority angle to two or three hook formats from a documented bank of eighteen-plus formats. The matching isn’t arbitrary. Certain hook formats work better with certain angle types, and that pattern recognition is part of what compounds across cycles.
Output: a concept slate of forty to sixty potential ads, each documented with avatar, angle, hook format, and intended body structure. The slate gets pruned to the concepts most likely to produce learning, usually fifteen to twenty-five concepts into production.
The discipline here is not falling in love with concepts. The job of the slate isn’t to produce winners. It’s to produce learning. A concept that loses informatively is more valuable than a concept that wins by accident.
Phase 3. Production, week three.
Production is where the Maple Media Lab differs most from a traditional creative shop. Production is parallelized aggressively. The script library, hook variants, editor’s clip library, and AI-assisted production tools all feed into the same pipeline, so ten or fifteen concepts can move through production in the same week a traditional shop would produce two or three.
The trade-off is that some concepts in the production batch will be rougher than a polished hero piece. We accept that trade-off deliberately. A concept that is eighty percent of the polish at twenty percent of the production time generates four times the learning per dollar, and the learning is what compounds.
Output: fifteen to twenty-five ready-to-launch concepts, each tagged with avatar, angle, hook format, and hypothesis being tested.
Phase 4. Read, week four.
The read is where the Lab earns its name. The team reads results against the hypothesis stated at production. Which angle won. Which hook won. Which avatar responded. Which format outperformed. The read isn’t a slide deck. It’s a working document that updates the angle library, the hook bank, and the avatar map for the next cycle.
The read also drives the variant pipeline for the next month. Every winner from the cycle becomes the parent of a family of variations, and that family enters the next cycle as priority production. The losers get documented, and the documentation feeds the team’s collective judgment about what to test in future cycles.
Why the Lab compounds
The Lab compounds because every cycle improves the inputs to the next cycle. The avatar map sharpens. The angle library gets more validated. The hook bank accumulates formats with documented win rates. The team’s judgment about which concepts will work gets calibrated. After three or four cycles, the brand isn’t running disconnected campaigns. It is running a continuously improving system.
The volume isn’t the point. The compounding is the point. The fact that the Maple Media Lab also produces fifty-plus concepts per month is a side effect of the discipline, not the goal.
