Launch is treated as the finish line, which is why so many redesigns quietly underperform. The month after go live is when problems surface, when search engines re evaluate everything, and when you find out whether the project actually worked. It is also when teams panic at normal fluctuations and start changing things, which makes diagnosis impossible. Here is what to do, in order.
Week one: find what is broken
Something is always broken. The job is finding it before customers do.
Watch 404s daily in Search Console and your server logs. Real traffic hits URLs your redirect map missed, and the fix is minutes if you are looking. Confirm every form still delivers to the right inbox, and send a test through each one, since a form that appears to work while silently discarding messages is the most expensive bug on this list. Check that analytics is recording on every template and that goals fire, because a week of missing data cannot be recovered.
Then check the things that only exist in production: email deliverability from the new server, third party scripts, payment or booking flows, and any security policy that behaves differently on the live domain.
Ask the people who talk to customers whether anything looks odd. Sales and support hear about problems days before they reach a dashboard.
Expect a dip, and know what is normal
Even a technically flawless migration usually sees ranking movement for two to six weeks while search engines recrawl, reindex, and re evaluate the new pages. Traffic can wobble by double digits in either direction during that window.
That is normal. What is not normal is a sharp sustained drop that does not begin to recover after a few weeks, or pages disappearing from the index entirely. Those point at specific causes worth checking immediately: a stray noindex, a robots.txt blocking crawlers, broken canonical tags, redirect chains, or content that was cut without anyone noticing it carried the rankings. The redesign SEO checklist is the diagnostic list.
Track index coverage weekly rather than obsessing over daily positions. The shape of the curve matters far more than any single day.
Resist changing things for three weeks
The strongest discipline in this period is doing nothing hasty.
If you change the headline in week one, adjust the navigation in week two, and add a popup in week three, you will never know what caused any of the movement you observe. Let the site settle, gather real data, and keep a running list of everything you want to change instead of shipping it immediately.
The exceptions are genuine defects: broken pages, wrong information, anything failing on mobile. Fix those the day you find them. Preferences wait for the list.
What to actually measure
Judge the redesign against the baseline you recorded before launch, and use the same window and the same traffic sources so the comparison means something.
Conversion rate matters more than traffic, because a redesign can legitimately reduce traffic (deleted thin pages) while increasing leads. Look at conversion by device especially, since mobile is where redesigns usually gain or lose the most. Check the paths people take now against the paths they took before. Ask sales whether lead quality changed, which is the number no dashboard reports and the one that decides whether the project succeeded.
If you did not baseline before launch, this is the moment you find out how much that cost, and it is worth reading about before your next project.
Week three and four: watch real behaviour
Once the numbers have settled, look at how people actually use the site. Session recordings and scroll depth tell you which sections nobody reaches. Site search queries tell you what people expected to find and could not.
Ten recordings will show you hesitations no metric reports, and five short conversations with recent customers will tell you what nearly stopped them. That is a week of cheap research that turns the post launch list from guesses into priorities.
Then start improving, deliberately
A launched site is a starting position, not a finished object. The best results we see come from teams who treat the first month as measurement and the following months as steady improvement: one page at a time, changed for a reason, measured afterwards, and tested properly when the traffic supports it.
Set a review date in the calendar for 90 days out, when seasonality has evened out and the verdict is fair. And decide who owns the site now, because a site with no owner degrades quietly no matter how good it was on launch day. For many teams that is a small ongoing arrangement rather than another project.
Just launched and unsure whether what you are seeing is normal? Send the numbers and we will tell you: hello@beconfidency.agency.
If you want a team that stays past launch day, that is exactly what our web design service is for.