Migration timelines are usually planned around the build and the launch date, with SEO work fitted in wherever it lands. That sequencing is backwards: the SEO work that matters most has to happen before the URL structure is finalised, which is often months before launch.
What the work involves is on SEO migration. This page covers when each part has to happen.
Before the build starts
This is the highest-value window and the one most often missed, because the site does not exist yet and SEO feels premature.
- URL inventory: one to three weeks depending on site size. Reconciling crawl, Search Console, analytics and backlink sources takes longer than the crawl itself.
- Performance baseline: days. Must be captured before anything changes or the comparison afterwards is meaningless.
- Structural input: while URL decisions are still open. Once the new structure is agreed, this window has closed.
Involvement here is worth more than anywhere else, because decisions can still change rather than merely be recorded. See pre-migration audit.
During the build
- Redirect mapping: the largest single task. One to four weeks depending on URL count, and longer where structures differ substantially.
- Exception handling: the URLs that do not map automatically. Frequently underestimated, and it grows with catalogue churn.
- Template review: as templates become available on staging.
Mapping cannot start until the new URL structure is known, and it cannot finish after launch. That constraint drives the whole schedule.
Before launch
- Staging validation: several days, and it needs to happen when staging genuinely reflects production.
- Redirect testing: against every URL pattern, not a sample.
- Pre-launch checklist: the final day or two, confirming directives, canonicals and sitemaps.
Compressing this is where migrations go wrong. A staging environment made available two days before launch does not allow meaningful validation.
After launch
- First 48 hours: redirects, directives and status codes. Immediate and intensive.
- First two weeks: indexation progress, crawl errors, content parity checks.
- Four to eight weeks: rankings and traffic against the stored baseline.
- Three months: the point at which recovery can be fairly assessed.
Judging a migration in week one is premature. Discovering a redirect fault in month two is negligent. See post-migration audit.
What compresses badly and what does not
- Compresses acceptably: template review, checklist execution.
- Compresses badly: URL inventory, which simply takes as long as the sources take to reconcile.
- Cannot be compressed: post-launch recrawling. Search engines take what they take, and no amount of budget accelerates it.
Frequently asked questions
How far in advance should SEO be involved?
Before the URL structure is decided, which for most projects means at kick-off rather than at the build stage. The cheapest hour in a migration is spent before anything is built.
How long before traffic recovers?
A shallow dip recovering over two to four weeks is normal for a mid-sized site. Large catalogues take longer because recrawling thousands of URLs is not immediate. A decline still deepening after three weeks is a fault, not a delay.
Can we migrate quickly if we have to?
Yes, but capture the URL inventory and baseline regardless, even without time to act on findings. Without them, post-launch diagnosis becomes guesswork.
When can we say it worked?
Around three months for most sites, longer for large catalogues. Earlier than that you are looking at recrawl progress rather than outcomes.
Consulting CTA
If you have a migration scheduled, book an SEO consultation early enough that the structural decisions are still open.