Search plugins, docs and guides… Browse plugins Browse plugins

Sell more

Protect and comply

Speed and delivery

Content and SEO

Launch and support

All 15 plugins Every plugin from $19 a year, or pay once from $49. Pricing Not sure which one? Search plugins, docs and answers

Does maintenance mode hurt SEO? Coming soon vs maintenance mode

When a holding page is harmless and when it costs you search traffic: 200 or 503, what Retry-After should say, and how to check what your site sends.

Updated 6 min read

On this page (7)
  1. The short answer
  2. Two situations, two status codes
  3. What Retry-After does
  4. How to check what your site is sending
  5. Mistakes that cost rankings
  6. Doing it with Coming Soon and Maintenance Mode
  7. Launch-day checklist

Not if the closure is short and the page answers with the right status code. Google's documentation treats a 503 as a temporary absence and gives a day or two as the limit; beyond that, pages can start dropping out of its index. A site that has never launched is a different case: it has nothing indexed to lose, and a normal coming soon page answering 200 is right.

The short answer#

  • A few hours, or a day or two: answer 503 Service Unavailable with a Retry-After header. Google's guide to temporarily pausing or disabling a website gives one to two days as the limit for this, and its page on reducing the crawl rate warns that a URL showing these codes for multiple days may be dropped from the index.
  • Longer than that: the same Google guide recommends not closing the whole site at all. Keep the pages up and limit what people can do, for example by switching off the cart and showing a notice. If you must close everything, it suggests an indexable placeholder home page answering 200, and warns that there is no fixed time for recovering from a complete removal.
  • A site that has not launched yet: a coming soon page that answers 200.

Two situations, two status codes#

A coming soon page stands in for a site that does not exist yet. Nothing ranks, so there is nothing to lose, and answering 200 lets people find and share the address while you build. Whether you also ask search engines not to index the holding page is your choice; just remember to remove that request on launch day.

Maintenance is a site that exists and is briefly away. Google's page on how HTTP status codes affect its crawlers says what a 503 does: its crawlers slow down for a while, whatever the page says is ignored, and pages already in the index are kept "but eventually dropped" if the errors carry on. The crawl-rate page adds that when many addresses answer 503, crawling slows across the whole host name, then picks up again once the errors drop.

A shop has a third option: the site stays up and only the buying stops, over a holiday for instance. Google's pausing guide puts this first. In its words, disabling the cart "is the simplest approach, and doesn't change anything for your site's visibility in Search". Add a notice saying when orders resume and keep any product structured data honest about availability.

What Retry-After does#

Retry-After tells whoever asked when to try again. RFC 9110 allows either a number of seconds or a date, and says that with a 503 it means how long the service is expected to be unavailable. Google's crawl-rate page shows both forms:

HTTP/1.1 503 Service Unavailable
Retry-After: 7200
HTTP/1.1 503 Service Unavailable
Retry-After: Wed, 21 Oct 2026 07:28:00 GMT

Give your honest estimate. If you expect to be back at three o'clock, say three o'clock (in GMT, as the date form requires). WordPress sends one with its own update message: Retry-After: 600, ten minutes, the same ten minutes after which it stops treating an unfinished update as maintenance. A long Retry-After does not buy a long closure, though. Google's warning about 503s lasting more than a day or two has no exception for it.

One file must stay open: robots.txt. Google's pausing guide says a 503 on robots.txt blocks all crawling, and a robots.txt that disallows everything may remove the site's pages from Search.

How to check what your site is sending#

Holding-page plugins often let signed-in administrators see the real site, so check as a visitor: from a terminal, or in a private window.

From a terminal, ask for the home page the way Google's own example does:

curl -I -X GET https://www.example.com/

The first line is the status. During maintenance you want 503, with a retry-after line among the headers; for a coming soon page, or after launch, 200. Do the same for an inner page, and for https://www.example.com/robots.txt, which should answer 200.

In a browser, open a private window, open the developer tools (F12, or Cmd + Option + I on a Mac) and choose Network. Reload. The first row is usually the page itself, its Status column shows the code, and clicking the row shows the response headers. Chrome's guide to the Network panel has pictures.

After launch, check that nothing still asks search engines to stay away. This prints the robots tag if the page has noindex, and nothing if it does not:

curl -s https://www.example.com/ | grep -i noindex

Mistakes that cost rankings#

A 503 left up for weeks. The status was made for short absences, and Google says plainly what happens to pages that keep answering with it.

A 200 "back soon" page during a short break on a site that already ranks. Google considers a 200 page's content for indexing, and when that content looks like an error message, Search Console reports it as a soft 404. Google's older post on planned site downtime recommends a 503 over a 200 error page, because a 503 tells crawlers the downtime is temporary.

Hiding the closure with noindex, a 404 or 410, or a robots.txt that disallows everything. Google's pausing guide warns that each of these can remove the site's addresses from Search.

Sending every address to a holding page with a permanent (301) redirect. Google's redirects documentation explains that a permanent redirect is a signal the target should replace the old address in results, which is the wrong message for a break.

Launching with Discourage search engines from indexing this site still ticked under Settings → Reading. WordPress then prints <meta name='robots' content='noindex, nofollow' /> on every page. The box is easy to tick while building and forget at launch.

Doing it with Coming Soon and Maintenance Mode#

Coming Soon and Maintenance Mode makes the choice of status code part of choosing the mode, as its documentation on coming soon, or maintenance describes:

  • Coming soon answers 200, for a site that does not exist yet.
  • Maintenance answers 503 Service Unavailable with a Retry-After. The number is worked out from your opening time when you have set one; without one it says an hour, and for an opening further away than a day it says a day.
  • Shop closed for orders needs WooCommerce. Every page stays open and indexable and only the buying pauses: the Add to cart button goes, carts and the checkout refuse, and a notice gives the day you take orders again (Shop closed for orders).

It opens on time: the switch-off time is compared on every request as well as scheduled, so the site opens on the minute even if WordPress's scheduled tasks have not run. Once the site is open, the Dashboard says so if Discourage search engines from indexing this site is still ticked, with a button that unticks it and empties the page caches it knows, and it names a robots.txt that asks every search engine to keep off every page. The admin, the login screen, robots.txt and your sitemap are never closed, and the closed page is sent with headers that tell caches not to store it.

What it does not do: it cannot tell you how long a closure is safe, which is Google's call and yours; it reports only what the WordPress setting and robots.txt ask for, never anything about search results; and a CDN or host cache that ignores those headers has to be emptied by hand after you switch.

Launch-day checklist#

  1. Switch the holding page off, or let its schedule do it, and look at the site in a private window.
  2. Under Settings → Reading, untick Discourage search engines from indexing this site and save.
  3. Empty your page cache and any CDN in front of the site.
  4. Run curl -I -X GET on the home page, two inner pages and robots.txt: all 200. The grep -i noindex check prints nothing.
  5. In Search Console, inspect the home page and request indexing, and submit your sitemap. WordPress's own sitemap is at /wp-sitemap.xml unless a plugin replaces it. Google's pausing guide suggests Search Console for a few pages and sitemaps for many.

The plugin for this

Coming Soon and Maintenance Mode

Put your site behind a coming soon or maintenance page and collect the email addresses of everyone waiting for it.

From $19 a year · every feature · 30-day refund

Close the site with the right status codeRead the docs

A coming soon page in the Split design. On the left, the headline Our new shop opens on 1 November, a short message, a countdown of 10 days, 0 hours, 43 minutes and 31 seconds with the line Opening November 1, 2026 9:00 am, and a sign-up form headed Be the first to know with an email field and a Tell me when it opens button. A photograph of two friends in knitwear fills the right half.

In the documentation

All guides