Recovering Organic Traffic After a Store Migration
Migration traffic loss is usually gradual and mostly recoverable. The baselines to record beforehand, the four failures to check in week one, and the process signals worth watching.
A migration rarely destroys rankings on the day it launches. It destroys them over the following six weeks, as pages drop out of the index one by one and nobody connects the dots because the traffic chart looks “a bit down”.
Recovery is mostly about detecting the loss early and having the redirect data to fix it. Here is the sequence that works.
Week 0: what to record before you switch
You cannot measure loss without a baseline. Export these the day before launch:
- Full list of indexed URLs, with impressions and clicks per URL for the last 90 days
- Top 100 landing pages by organic sessions
- Which URLs have external backlinks — this is the recovery priority order
- Total indexed page count per language and per content type
Without the per-URL data, “traffic is down 30%” is unactionable. With it, you can name the pages that fell and check each one.
Week 1: the four failures that account for most of the loss
1. Redirects that point at the wrong page. The most common and most damaging. Everything redirected to the homepage passes no topical relevance and tells search engines the old page is gone. Every old URL must map to the closest equivalent page, not to home.
2. Redirect chains and loops. Old URL → new URL → another URL, or a loop that returns a 500. Chains lose signal and slow crawlers. Check every mapping resolves in one hop.
3. The old sitemap still being served. Very common: the old platform’s sitemap file remains, search engines keep discovering dead URLs, and crawl budget goes to 404s. Verify the sitemap contains only live URLs.
4. New URLs that canonical to the wrong place. Migration plugins sometimes leave canonicals pointing at old paths, or at a staging domain. Check ten product pages and ten category pages by hand.
The migration guide covers building the map itself. Recovery assumes the map existed in the first place — without it, this becomes an archaeology project.
Week 2: measure the right things
Rankings and traffic are lagging indicators. Watch process signals instead:
| Signal | Healthy | Concerning |
|---|---|---|
| 404 count in crawl | Falling steadily | Flat or rising |
| Indexed pages | Approaching the pre-migration count | Well below baseline |
| Pages with “Crawled – currently not indexed” | Small and stable | Growing |
| Redirect hits | High, then declining | High and flat |
Indexed page count is the single most useful number. If it recovers to baseline within two to three weeks, rankings usually follow. If it stalls at 60% of baseline, something is systematically blocking pages — often a noindex left on a template, or a robots rule inherited from staging.
Week 3–6: the recovery actions
Resubmit and refresh. Submit the new sitemap, then use URL inspection on the ten highest-value pages. Do not mass-submit hundreds of URLs; request indexing where it matters.
Rebuild internal links. Migrations routinely lose internal links that passed equity to product and category pages. Crawl the new site, compare the internal link graph to the old one, and restore the important paths. This is the most underrated recovery action and it costs nothing.
Update external links where you can. For your top referring domains, check whether they point at old URLs. If they 301, that is fine. If they point at a dead end, reach out and ask for the new URL — most sites will update a link when asked politely.
Publish something. A migration period with no new content gives search engines no reason to recrawl frequently. Two or three substantial pages during recovery month help more than a batch of thin ones.
What not to do
Do not change the design at the same time. A migration plus a redesign makes it impossible to attribute the loss. If both are necessary, do them months apart.
Do not add noindex “temporarily”. Temporary noindex tags outlive their purpose more often than not, and each one is a page removed from the index.
Do not migrate during peak season. A four-week recovery window in November is a four-week revenue loss.
When the loss is not going to recover
Two situations do not resolve on their own:
- Content that was genuinely thin. Some old pages ranked because the competition was worse. After a migration with better technical hygiene, they do not come back because they never deserved to
- Structural changes sold as a migration. If URLs moved from
/category/productto/productand category pages were consolidated, you are not recovering — you are rebuilding. Plan it as a content project
The honest summary: most migration traffic loss is recoverable within four to eight weeks, provided you have per-URL baselines and a complete redirect map. The loss becomes permanent when the migration had no map and no baseline, because then there is nothing to recover with. If you are planning a move and want the technical SEO side audited before launch, that sits under technical SEO.