Fashion: seasonal turnover - launching and sunsetting collections without losing SEO equity
Fashion's second defining headache is temporal: the industry replaces large fractions of its catalog on a fixed drumbeat - spring/summer in, autumn/winter out, drops in between. Every turnover is simultaneously a launch operation, a retirement operation, and an SEO equity event. Stores that handle it as improvisation reset their search presence twice a year and wonder why rankings never compound. The fix is an operating model, and you've already met all its parts.
The structural principle: permanent frames, rotating contents
The Phase 4 seasonal rule, now as catalog architecture: URLs and category structures are permanent; products rotate through them. /collections/knitwear exists forever, accumulating authority across every season that passes through it; "AW26" is a tag, a custom label, a filterable attribute - never a URL. The same logic extends to the seasonal marketing frames: /collections/new-in is permanent (its contents turn over weekly), the seasonal campaign page ("The Autumn Edit") keeps one evergreen URL refreshed per season, and drop culture gets the same treatment - /collections/drop lives forever, drop names live in content. What this buys, compounding: a knitwear page with four years of accumulated standing enters each autumn pre-ranked, while the competitor who minted /collections/aw26-knits starts at zero. Again. The audit for existing stores: any collection URL containing a year or season code is equity scheduled for demolition - migrate the URL to permanence, 301 the old, and stop the cycle.
The launch half: arrive indexed, not just live
Seasonal launch as data operation, sequenced backward from Phase 4's lead-time rule (pages complete 2–3 months before demand peaks): products enter through the QA gate complete (Phase 10 - launch crunch is exactly when gates get bypassed and exactly when they matter; incomplete launch products spend their critical indexing window as thin pages), style-level enrichment done pre-publish (Phase 11's first lesson - the inheritance structure makes this a styles-count job, feasible pre-season), collections pre-built on rules so launch products self-place (Phase 1), internal links and content refreshed to point at the incoming season (the buying guides from Phase 4, updated - not rewritten), and feeds verified within days of publish (Phase 7's launch runbook: new items approved in Merchant Center before the campaign spend starts, not discovered disapproved during it).
The sunset half: retire like it's revenue, because it is
Season-end as the other runbook, applying Phase 5's lifecycle table at batch scale: carryover styles (continuing next season) keep pages live untouched - the majority case in most assortments, and the strongest argument against panic-unpublishing; replaced styles 301 to their successor (the style-code metafield makes successor-mapping a query, not a memory exercise); true retirements get the traffic check - pages with standing stay live as soft landings pointing to the closest current alternatives; the rest unpublish cleanly. Then the two verification passes that close the season: internal links (navigation, content, homepage) audited for pointers at sunsetted pages, and GSC's 404/coverage reports watched for two weeks for anything the runbook missed. Run twice a year, this whole model takes a day per season - and it's the difference between a store whose search presence saws up and down annually and one that enters every season stronger than the last. Fashion's rhythm doesn't have to fight SEO's memory; wired correctly, it feeds it.