Discussion
  • Enterprise SEO

What go/no-go criteria would make you push back a migration launch date, and have you ever actually managed to delay one?

Published 1 reply

Discussion starter: an enterprise SEO question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.

Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.

1 reply

  • admin
    Admin
    Original poster

    I would push a launch date only for problems that are likely to cost meaningful organic revenue and are slow or impossible to reverse once Google has crawled them. Everything else goes on a dated post-launch fix list. That gives a short list of hard stops: an incomplete or untested redirect map for URLs that earn traffic or links, indexing blocks or wrong canonicals in the production configuration, key templates losing content or links in the rendered HTML, no working rollback, no way to measure the outcome, and a date inside peak trading.

    On whether delays can be won: they do get approved, but an argument framed as “SEO isn’t ready” is a weak one. A request has the best chance when the criteria were agreed before anyone was attached to the date, the risk is expressed as revenue exposed per week, and the request arrives with an alternative. That alternative might be a new date tied to a fix list, a phased launch, or a descoped launch that keeps the old URLs. The realistic win may be removing the riskiest part of the launch rather than moving the date.

    Reversibility is the real test

    During a migration, Google recrawls old URLs, follows redirects and moves signals to the new ones. Google’s site-move guidance says to expect temporary ranking fluctuation during the move, and that for a medium-sized site it can take a few weeks or more before Google gradually starts showing the new URLs instead of the old ones, and longer for larger sites.

    Mistakes made in that window get processed along with everything else. A stray noindex caught in the first hours costs little. A redirect rule that sends thousands of old product URLs to the home page, crawled for a fortnight, costs far more, because fixing the rule means waiting for Google to recrawl those URLs and process them a second time. A domain move is harder still to walk back: the Change of Address tool’s help page advises against chaining site moves, so a second move to undo the first is not a clean option.

    So the question for each open issue is not “is this perfect?” but “if we launch with this, how much traffic is exposed, and how quickly can we undo it after Google has seen it?”

    The hard stops I would agree before the date is set

    Each criterion needs a test and a threshold that someone other than the SEO can check. Criteria like “SEO has signed off” are the easiest to override.

    Criterion How to test it No-go if
    Redirect coverage Build the old URL list from pages with organic clicks in the last 16 months (API or bulk export on large sites), pages with external links, pages in current sitemaps and pages Googlebot requests in the logs. Run the full list through the production redirect rules on staging. Any URL in the top tier (for example, the URLs that together drive 80% of organic revenue or leads) lacks a tested, relevant destination. Chains, loops or blanket redirects to the home page remain.
    Redirect type Scripted checks of status codes and destinations Temporary or JavaScript redirects are used for permanent moves where server-side permanent redirects (301 or 308) are possible
    Indexing controls Crawl the production configuration for meta robots, X-Robots-Tag, robots.txt, canonicals and hreflang A staging noindex or robots.txt block is still in place, or canonicals point at staging hosts or old URLs
    Rendered parity on key templates Crawl staging with JavaScript rendering (behind authentication if needed) and compare main content, internal links, titles, headings and structured data with the old site A template that drives significant traffic loses its main content or its internal links in the rendered HTML
    Rollback A written plan naming who decides, what triggers it and how long it takes, rehearsed once Nobody can say how long a rollback would take, or the old platform is switched off at launch
    Measurement Analytics and conversion tags verified on every template; Search Console verified for any new host; log access arranged for launch week You could not tell within a day or two whether organic traffic and conversions are holding
    Timing and people Trading calendar, code freezes and leave Launch falls in or just before peak trading, or the engineers who can fix redirects are away in launch week

    Several of these come straight from Google’s own guide. It asks for server-side permanent redirects where possible, warns against redirecting many old URLs to one irrelevant destination such as the new home page, tells you to remove any noindex or robots.txt blocks that were only needed for the migration, and suggests timing the move for a period of lower traffic where you can.

    One testing detail catches teams out. Google’s URL-based testing tools can only test pages Google can fetch, and URL Inspection only covers properties you have verified, so a password-protected staging site needs a crawler that supports authentication. Google’s guide mentions URL Inspection for checking individual redirects, which in practice makes it a post-launch spot check unless your staging host is publicly reachable and verified.

    What goes on the fix list instead

    These are real issues, but they are cheap to fix after launch and do little lasting damage in the meantime:

    • Performance regressions in lab tests, within an agreed tolerance. Field data for new URLs takes weeks to build, so set the tolerance before launch.
    • Missing or duplicated meta descriptions, alt text and structured data warnings that do not remove rich result eligibility on key templates.
    • Low-value URLs with no traffic and no external links that will return a clean 404 or 410 instead of a mapped redirect.
    • Internal links that still point at old URLs and pass through one redirect, provided they are updated within an agreed number of days.
    • For domain moves, the Change of Address submission itself. It needs you to own both properties with the same account, to use domain-level properties and to have a 301 from the old home page to the new one. Have those ready, but the submission happens after launch.

    “Rankings might fluctuate” is not a reason to delay. Google’s own guidance says they will, whatever the date.

    A note on launching during a core update

    No Google documentation I can point to says a core update rollout adds ranking risk to a migration, but it does wreck measurement: you cannot separate migration losses from update movement. Google’s advice for judging an update is to wait at least a full week after it completes before analysing your site. If the date can move by a week or two, move it. If it cannot, record the rollout dates and judge the migration on per-URL evidence (redirect responses, indexing of top URLs, crawl behaviour in the logs) rather than aggregate traffic.

    What makes a delay request land

    1. Agree the criteria before the date has an owner. A go/no-go list signed at kick-off by the person who owns the launch turns a delay into applying an agreed rule. The same list raised in launch week looks like the SEO team protecting itself.
    2. Translate the gap into exposure. “The redirect map is incomplete” is easy to wave through. “Pages producing a third of organic revenue have no tested redirect, which puts about this much revenue per week at risk until they are fixed and recrawled” is not. Use your own revenue data and give a range, not a point estimate.
    3. Bring alternatives, not just a no. A new date tied to a specific fix list. A phased launch, which Google’s guidance allows for large sites while recommending that small and medium sites move all URLs at once. Or separating the risks, for example launching the new platform on the existing URL structure, if it can serve those patterns, and changing URLs later.
    4. Ask the decision owner to accept the remaining risk in writing. If leadership keeps the date, document the open issues, the expected exposure and who fixes what, and ask the decision owner to sign it. Seeing the risk written down sometimes changes the decision. If it does not, it secures engineering time for launch week.
    5. Hold the formal go/no-go a week out and a final check the day before, on the production configuration, with named people who have the authority to stop it.

    Hypothetical example:

    A retailer plans a new platform and a new URL structure three weeks before its biggest trading period. The staging crawl finds that 7,500 of 32,000 URLs with organic clicks have no mapped redirect, including 40 category pages that carry a large share of organic revenue. Faceted URLs canonicalise to the staging host, and the rollback plan is “restore from backup, about two days”.

    The SEO lead brings three options to the go/no-go meeting: a late January date with a fix list; the original date, with the new platform serving the existing URL structure and the URL change moved to the quieter first quarter; or the original date as specified, with a signed risk acceptance and a reserved engineering squad. Leadership keeps the date and takes the second option. The date did not move, but the part of the launch carrying most of the risk did.

    If the date holds anyway

    • Freeze non-essential releases for launch week and reserve engineering time for redirect and template fixes.
    • On the day before, check robots.txt, meta robots, canonicals and a redirect sample on the production configuration.
    • On launch day, run the full old URL list through production and check every status code and destination. Spot-check key URLs with URL Inspection and submit the new sitemaps.
    • For the first two weeks, read server logs daily: Googlebot requests to old URLs and the responses they get, plus 404 and 5xx spikes. The Page indexing and Crawl stats reports lag, so logs come first.
    • Track organic landing sessions and conversions by template, and monitor the old site’s crawl and traffic as well as the new one’s, as Google’s guide suggests.
    • Keep redirects in place. The Change of Address help page gives at least 180 days, longer if the old URLs still get traffic; the broader site-move guide says generally at least a year. Plan for the year, and keep redirects for URLs that still earn links or visits.

    The rule I would take into the next project: agree measurable hard stops at kick-off, judge each open issue by how much traffic it exposes and how hard it is to reverse after crawling, and when the date cannot move, fight to move the risk out of the launch instead.

    Need help with this on your own site?

    Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.

    See SEO services

Add a reply

Replies are open to approved members. Log in or apply to join.