The writing operating model — from founder draft to catalogue scale.

The first fifty listings on a young brand are usually written by one person — the founder, the marketing lead, the agency principal. The voice is consistent because it lives in one head, the keyword choices are consistent because the same person did the research, and the asset hand-offs are consistent because there are no hand-offs. The system holds. It also doesn't scale. The next fifty listings break it: a second writer joins, a third category opens, the brand starts running in three markets, and within a quarter the catalogue starts to read like four different brands talking to each other. The fix is to write down the operating model before that happens — not after.

The four roles

Four roles cover the writing function at scale. They can collapse into one person on a small team or expand into a dozen on a vendor; what matters is that the responsibilities are named, not who holds them.

1. The brand voice owner. Maintains the brand voice document, decides voice questions when a writer is unsure, and signs off on the first listing in any new category or market. One person, never two.
2. The keyword owner. Maintains the keyword harvest from Module 6, runs the search-term-report and Brand Analytics pulls on cadence, and produces the per-SKU keyword brief that the writer works from. The role that keeps the catalogue searchable as Amazon's term landscape drifts.
3. The writer. Drafts the title, bullets, description, A+ copy and alt-text from the keyword brief plus the brand voice document. Self-runs the six-check editorial pass from Episode 17.
4. The second reviewer. Runs the same six-check pass blind, arbitrates against the writer, and signs off for moderation submission. Never the same person as the writer on the same listing.

The three documents that make hand-offs survive

Three living documents. Without them, every hand-off becomes a verbal context-transfer that loses 30% of the brand intent each step.

1. The brand voice document. Two to four pages, never more. Voice principles (we are direct, not casual; we are specific, not general; we are warm, not cute), a vocabulary list of preferred and forbidden words, the brand name's capitalisation and spelling in every context, the variant naming convention, and ten before/after examples of voice in practice. Every writer reads it on day one; every writer re-reads it on day thirty.
2. The per-SKU keyword brief. One page per ASIN. The hero keyword (mandatory title placement), the secondary keyword cluster (bullet placement), the long-tail cluster (description and A+ alt-text), the keywords explicitly to avoid (competitor brand names, off-brand attributes), and the indexing notes (what's already indexed, what's missing). Produced by the keyword owner, consumed by the writer.
3. The per-SKU asset pipeline. The current state of every asset that has to ship with the listing — main image, secondary images, video, A+ image set, Brand Story mapping, translations per market. Each asset has an owner, a status (drafted / in review / approved / live) and a target ship date. Without this document, half-finished listings ship with placeholder images and the half-finished asset becomes the permanent one.

The cadence that keeps it honest

The operating model is not a one-time setup — it's a cadence. Three rhythms keep a 500-SKU catalogue from drifting.

• Weekly — the new-and-changed sweep. Every ASIN drafted, edited or republished in the last seven days is reviewed against the brand voice document and the six-check editorial pass. Catches drift while it's still cheap to fix.
• Monthly — the keyword refresh. The keyword owner re-pulls the search-term report, the Brand Analytics top-search-terms, and the competitor reverse-ASIN harvest. New keywords with material volume get queued for inclusion in the next per-SKU brief; dead keywords get retired from the cluster.
• Quarterly — the evergreen audit. Re-read every listing in the catalogue against the current brand voice document and the current keyword brief, not just the recently-changed ones. The audit usually catches 10–15% of the catalogue carrying voice or keyword choices from a previous era of the brand.

The hand-off discipline that fails first

Across audited brands, the hand-off that fails earliest is the keyword brief → writer step. The keyword owner ships a brief that's thorough but unprioritised; the writer treats every keyword as equally important; the title becomes a comma-separated keyword list and the moderation queue rejects it. The fix is to make the brief explicitly tiered:

• Tier 1 — the hero keyword. Exactly one. Mandatory title placement, ideally inside the first 80 bytes.
• Tier 2 — secondary cluster. Three to five keywords. Distributed across bullets and the description.
• Tier 3 — long-tail cluster. Ten to twenty keywords. Distributed across A+ alt-text and the backend search-terms field (Module 7).
• Tier 4 — avoid list. Competitor brand names, retired attribute words, claims the legal team has flagged.

Without tiers, the writer has to do keyword research while writing — which is exactly the role split the operating model exists to prevent.

Where AI assistance fits the operating model

AI changes the unit economics of three of the four roles, not all four. The brand voice owner role stays human — voice is a judgement role, and a model that drifts the brand voice in any direction destroys the catalogue's coherence. The keyword owner role is amplified by AI for harvest and clustering, but the prioritisation tiers stay human. The writer role is the one most changed: a model that has the brand voice document and the per-SKU keyword brief in context can draft a credible first version in under a minute. The second reviewer role stays unambiguously human — review is the role that catches what the model missed, and a model reviewing a model finds nothing.

What this closes out

With the operating model in place, Module 8 has covered the full writing surface: titles, bullets, descriptions, the upload paths into them, A+ Content in standard and premium tiers, Brand Story, the moderation playbook, the editorial review pass, and the operating model that holds it all together as the catalogue grows. The next episode in the Listing Guide picks up from a different angle entirely — how the listing performs after it goes live, and how the brand reads the signal back from Amazon's surfaces.