Strategy and communications are what the marketing lead decides. The tenth movement is about getting work out of the people who aren't in the room: external designers, editors, copywriters and developers. A self-sufficient brief, a board with WIP limits, and deadline control before the deadline, not on the day of it.
Brief, board, critical chain and five breakdowns. Enough to see the approach
Brief templates, three filled examples, the board's design, deadline-control rules
brief_templates.md, example_briefs.md, creative_board.md, deadline_control.md, asana_import.csv, n8n_creative_intake.json
Production volume, budget, channels — from movement VIII's report; creative legal limits — from the Bible
A network of four aesthetic-cosmetology studios in Kyiv. The creative team is external contractors: a designer, a motion editor, a copywriter, a developer. In-house — an SMM manager (shooting at the studios, organic) and the marketing lead (sets the brief, accepts the work). The contractor wasn't at the stand-up and doesn't know that "art. 21" is about medical-service advertising. So the whole load is on the brief.
14 blocks. Until all are filled — the task doesn't go into work. The contractor takes it on knowing the scope; the client can't later say "I meant something else." Templates — brief_templates.md, three filled examples — example_briefs.md.
| Block | What's in it | Why |
|---|---|---|
| Context | which campaign, what's already running, what didn't work — with movement VIII's numbers | the contractor understands what they're plugging into |
| Task in one sentence | "N statics for cold Meta on the injectables line, autumn peak" | protection against scope creep |
| Audience and insight | the segment from the Bible, its fear/objection, what should click | without this the creative is "pretty but off" |
| Key message | the one thought a person takes away — not a slogan | every element of the layout works toward it |
| Forbidden elements | a compliance checklist (art. 21 + Meta), verbatim | the contractor physically doesn't know these rules |
| Acceptance criteria | what it's accepted on: message, compliance ✔, spec, number of variants | "like it / don't like it" → into objective points |
| Deadlines and checkpoints | not a date, but: draft / revisions / final | control before the deadline |
| Revision rounds | 2 by default, then a new task | protection for both sides |
Full list: title · context · task · audience · message · required elements · forbidden · format and spec · references · acceptance criteria · deadlines · revision rounds · budget · where to deliver.
1 · baseline
Message the contractor: "Need creatives for botox, I'll send references, deadline this week."
2 · what's wrong
"Creatives" — how many, what format, for which audience, with what message? "This week" — a draft or a final? The contractor guesses — and makes the wrong thing. The first delivery goes back for rework, the deadline shifts, both sides are annoyed. And crucially — the work can't be accepted objectively, because there were no acceptance criteria. A "like it / don't like it" dispute has no resolution.
3 · senior level
A 14-block brief as a Definition of Ready: the task doesn't move to "In progress" until everything is filled, including the compliance checklist and acceptance criteria. Context — with real movement VIII numbers ("in Q2 the 'price list' static gave a CPL of ₴910, switched off — don't repeat"). Audience insight — from the Bible, not "women 25–45." One brief form for all task types, plus type-specific fields (for video — the hook verbatim; for development — tech requirements and the movement XII QA checklist).
4 · outcome
The contractor starts with full understanding and a scope estimate. Acceptance is by a list, not a mood. The average number of revision rounds drops — and this is measured (target ≤ 1.3; higher means the briefs aren't specific enough).
A kanban (Asana / Trello). 10 columns, a card with fields, WIP limits. A demo sprint for import — asana_import.csv. Design — creative_board.md.
The card name — LUME_meta_inject_static_S38_v2 — matches the UTM-naming logic from movement VIII, so a creative on the board links to its row in the report: which layout gave which hook rate and CPL.
1 · baseline
One task per contractor: "make the creatives for September," we accept it all as a batch at the end of the month.
2 · what's wrong
A batch at the end of the month means learning about problems a month later. Nothing can be tested earlier. If five of ten concepts miss — rework the whole batch at the last minute. And the contractor allocates effort themselves: there's "busyness," but the specific needed task doesn't move.
3 · senior level
Decomposition: one card = one deliverable artefact (a static pack for one line, one Reels, one landing page). A WIP limit of 2 "in progress" per contractor: a third task isn't picked up until one has gone to acceptance. This optimises the throughput time of a specific task, not busyness. A separate limit on "acceptance + revisions" at the marketing lead (≤ 5) — if it's more, the lead has become the bottleneck, a signal to delegate part of the acceptance to SMM.
4 · outcome
Tasks pass the board faster and more predictably. Cycle time is measured by type (benchmark: statics 3–4 days, video 5–7, landing 8–10) and is visible at the sprint retro.
The problem isn't that the contractor slips a deadline. The problem is that you find out on the day of the deadline, when the campaign launch (movement XI) is the day after tomorrow. Rules — deadline_control.md.
1 rough concept / storyboard / text structure / section prototype. Catch "we're going the wrong way" while rework is cheap.
All layouts rough / rough cut / full text. Catch scope and quality before the finish.
Done to the Definition of Done. Acceptance against the brief's criteria.
a missed checkpoint = a red flag immediately, silence counts as a slipthe autumn-peak campaign launch. Every task is estimated aggressively; half of the total padding is collected into a shared buffer before the start
The naive way — add +2 days to every task. Then the performer spends the whole reserve (work fills the time allotted), and there's no buffer left for a real force-majeure.
Critical chain: every task is estimated with no personal padding, and half of the total reserve is collected into one buffer at the end. Buffer spent — visible at the stand-up ("ate 2 of 4 days at the design stage"). While the buffer > 0 — the launch date is safe. Gone negative — escalation: cut scope, add a contractor, or move the date. That's a decision made a week ahead, not the night before the start.
1 · baseline
Set the deadline, the day before it ask "is it all ready?"
2 · what's wrong
Asking the day before saves nothing: if it's not ready — a full creative can't be made in a day, and the campaign is already in the calendar. Plus the contractor keeps answering "I'll make it" till the last — not out of malice, out of optimism. Word of a slip arrives exactly when you can no longer react to it.
3 · senior level
Three checkpoints instead of one: a slip is visible at 30% of the timeline, when there's time. The buffer is held on the chain, not the task. Early signals are automated (n8n): a reminder the day before a checkpoint, an alert on a 1-day overrun, a "launch at risk" alert at a buffer < 1 day — straight to the founder. Dependency on a studio shoot is a separate risk with a plan B in the brief (an edit from archive).
4 · outcome
On-time rate is measured as the share of tasks delivered to the final without dipping into the buffer (target ≥ 80%). Buffer spend by campaign — a trend over sprints. The launch date stops being a lottery.
Endless iterations are the most common reason creative production doesn't fit the timeline or the budget, and the reason contractors don't like working with beauty.
| Revision type | Whose fault | How it's counted |
|---|---|---|
| Doesn't match the brief | the contractor didn't do what was asked | counts against the 2 rounds, no extra pay |
| Clarification within the brief | normal process | round 1, round 2 |
| Client changes the inputs | changed their mind on offer / audience | a new task, a new brief, payment |
1 · baseline
We revise until everyone's happy. "We're all about quality."
2 · what's wrong
"Until everyone's happy" has no limit: everyone on the team adds their opinion one at a time, revisions come as a stream of messages, the contractor works for free after the third iteration and next time bakes it into the price or leaves. The task timeline becomes unpredictable, which means the launch becomes unpredictable.
3 · senior level
2 revision rounds fixed in the brief and built into the timeline between checkpoint 2 and the final. The revision list is one document with concrete points, collected by the marketing lead and handed over whole, not 15 messages from three people. Split by cause: an under-delivery — free, against the rounds; a client input change — honestly a new paid task. This protects both the budget and the relationship with the contractor.
4 · outcome
The contractor knows the limits and prices adequately. The client is disciplined about collecting revisions. If there are regularly more than two rounds — what gets fixed isn't the revision process, it's the input brief quality.
Someone who hasn't read the brief looks at the creative and names the main thought. Matches the brief — this point passes.
Line by line: no doctors on camera, no "most effective," no before/after with a flaw, there's a "Sponsored" mark, there's the MoH licence. One unmet point — not accepted.
Dimensions, weight, duration, number of variants, source files, file naming.
The "what this creative tests" field is filled — so in movement XI it launches as part of an A/B, not "just another layout."
1 · baseline
The marketing lead looks and says "yes" or "no, redo it, something's off."
2 · what's wrong
"Something's off" is not a revision task. The contractor doesn't know what to change and guesses. The judgement depends on the mood and on who else glanced over the shoulder. A good creative can fail because "it didn't land," and a weak one pass because "eh, fine." And there's no link between acceptance and the result in the ads.
3 · senior level
Acceptance is a checklist of the brief's criteria: the message reads, compliance line by line ✔, the spec is met, variants delivered, the hypothesis recorded. Each point is yes/no, not "how many out of ten." A subjective "I don't like it" is also allowed — but then the lead is obliged to translate it into a concrete point ("the model's face is too tense — a relaxed delivery is needed"), otherwise it's not grounds for a revision.
4 · outcome
Acceptance is reproducible: another person with the same checklist reaches the same decision. A creative that went "live" carries an ID and a hypothesis — and its result (hook rate, CTR, CPL from movement VIII) comes back into the next sprint's briefs. Production closes the loop with media buying.
Advertising aesthetic cosmetology in Ukraine — Advertising Law art. 21 + Meta policy. The contractor doesn't know these rules by default. The fine for a missing "Sponsored" mark is 5× the placement cost.
| Not allowed | Why |
|---|---|
| doctors / medical staff in the role of endorsing a service | art. 21 — a direct ban |
| fear of illness / ageing, playing on insecurity | art. 21 + contradicts the "trust" position |
| "the most effective," "safe with no consequences," "guaranteed result" | art. 21 |
| before / after where the "before" is a flaw / problem | Meta cuts it; art. 21 — "hint of a shortcoming" |
| promising a concrete result ("minus 10 years") | art. 21 + Meta personal attributes |
| placement without a "Sponsored" mark | fine of 5× the ad cost |
Required: a "Sponsored" mark, a note that LUMÉ is a licensed medical centre (MoH number), a calm clinical tone, an emphasis on the process and on the fact that the decision is made by the doctor at the consultation.
The same approach as in movements VIII–IX: every node does one clear thing. n8n_creative_intake.json — 17 nodes, import-ready.
| Node | What it does |
|---|---|
| Webhook · brief from a form | receives the filled brief |
| Definition of Ready check | all load-bearing fields filled? checkpoints ascending and before the deadline? type valid? — if not, the task doesn't go into work, a human-readable list of what's missing goes to the chat |
| Append to the board + notify | a card in creative_board, a ping to the contractor and the lead with checkpoint dates |
| Deadline scan (weekdays 10:00) | for each active task: checkpoint tomorrow → reminder; overdue → alert; computes buffer spend per chain; if there's nothing to report today — no message is sent |
| "Launch at risk" alert | chain buffer < 1 day → to the marketing lead and the founder |
| Sprint retro (Fri 18:00) | sent only if on-time < 80% or the average number of rounds > 1.3 — otherwise the sprint went as planned and no review is needed |
What needs setup: the brief form (Tally / Google Forms → webhook), the board as a Google Sheet creative_board or the Asana API, Telegram chat tokens. The logic is written and imports — as in movements VIII–IX.
Computed at the sprint retro. Each one leads to an action — otherwise we don't compute it.
Production volume is an estimate. ≈ 55–70 creatives/month is derived from movement VIII's budget, the refresh cycle and the channel count. The real figure depends on how many formats are tested and how active the organic is.
Cycle time and on-time are benchmarks, not measurements. The ranges are from agency and production practice, not from the real work of the LUMÉ team. On a live stream they're calibrated over 2–3 sprints.
The compliance checklist is from open sources. Advertising Law art. 21 and Meta policy are read, not checked by a specialist lawyer. For a real business the checklist is agreed with a lawyer and updated when Meta's policies change.
Rates and budgets in the briefs are placeholders. The examples show ₴____: concrete figures depend on contractors, which a model client doesn't have.
research/source_register.md block H