Ecommerce Replatforming Checklist

Replatforming — moving the store to new cart software — is the general migration discipline with ecommerce's multipliers attached: thousands of catalogue URLs whose patterns the new platform wants to change, revenue metered daily so every settling-week has an invoice, and the store-specific systems (feeds, apps, reviews, redirects-at-scale) that each need their own moving plan. The general checklist governs; this is the store-specific overlay — the items replatforming projects forget until week two of the traffic dip.

Pre-migration: the store-specific inventory

1. The URL-pattern negotiation, first and hardest: platforms impose URL opinions (forced /products/ prefixes, category-path handling, parameter formats) — map old patterns to new before contract signature where possible, because "can we keep our URLs?" answered late is the difference between a redirect-light migration and a 40,000-row map. Where patterns must change: exact-counterpart rules (pattern-based, per the parameter-ghost lesson), with the equity findings' single-hop discipline.
2. The catalogue-truth export: every URL with traffic/rankings/links (the standard ledger) plus the store layers — review corpora (portable? indexable on the new platform, per the rendering rule?), the accumulated redirect estate (it migrates too — old-old URLs must still resolve), and the markup/feed configuration being rebuilt.
3. The app-stack reckoning: replatforming resets the app sediment — rebuild the stack from the audit's survivors, not from habit; it's the one silver lining migrations hand you free.
4. Staging verification, store edition: the locked environment checked for template parity (titles, canonicals, schema per state — in-stock/out/sale/variant), facet policy re-implemented (new platforms mint new parameter formats — the old robots rules won't match them), and pagination mechanics rendered crawlable.

Cutover and the store-specific monitoring

The general launch hour plus: Merchant Center feed re-pointed and re-validated same-day (feed disapprovals during a migration compound the organic dip with a Shopping outage), payment/checkout regression-tested before anything else (SEO waits; revenue doesn't), and the monitoring dashboard split by template — categories, products, content tracked separately, because store migrations leak by template class and the aggregate hides which. The case study's edge-hunting routine applies doubled: 404s-with-referrers daily (old product URL variants are the store's parameter ghosts), crawl-diffs weekly, and the consistency triangle (page↔markup↔feed) re-audited at week two when template fixes have stopped churning.

The expectations conversation

Set before signature, in writing: the settling dip is real (the clean case runs weeks, per the timeline), it lands partially on revenue, and it's the price of whatever the new platform buys — which means replatforming needs a reason worth the dip (platform end-of-life, genuine capability walls, cost structures) and never happens in Q4. The stacking rule at store stakes: platform this release, taxonomy changes the next, redesign the one after — each dip diagnosable, per the don't-stack doctrine the case files keep re-proving.

Frequently asked questions

Our new platform can't keep our URL structure at all. Deal-breaker?

Not with a complete pattern-mapped 301 layer — the equity transfers when mapping is exact and single-hop; the cost is the settling window plus permanent redirect maintenance. It is a real line-item in platform comparison: the platform that preserves URLs starts the race weeks ahead.

Do we migrate the blog/content layer simultaneously?

If it lives on the same domain (as the subdirectory rule wants), it migrates in the same map with the same care — guide URLs carry the earned links whose routing funds the catalogue. The forgotten-blog-URLs migration leak is a classic; the content inventory belongs in the master ledger.

How long until we know the migration "worked"?

Template-level recovery read at weeks 4–8 against the exported baselines (the 97%-then-104% shape is the healthy pattern); full verdict after one seasonal cycle on stores with peaks. And the post-migration quarter is the moment to resume the offence — the authority programme that makes the platform question matter less every year (our quarter).