When your site goes down, customers usually notice before you've found the cause. What they see next, whether a bare error, silence or a clear note, decides whether they wait or buy elsewhere. Below are copy-paste templates for every stage of an outage, where to post each one, and the server setting that keeps a maintenance page from hurting your rankings.
What should a website down message to customers say?
A good website down message covers four things: what's affected, that you know about it and are working on it, what customers can do in the meantime, and when they'll hear from you next. Keep it short and non-technical. Don't promise a fix time you can't hit. Promise the time of your next update instead.
What is website downtime?
Website downtime is any period when your site is unreachable or too slow to use: a full outage, server errors, or pages so slow that customers can't finish a purchase. It's either unplanned (a hosting failure, a bad plugin update, an expired certificate, covered in what causes website downtime) or planned maintenance. Unplanned downtime needs an acknowledgment; planned downtime needs notice.
Which templates can you copy and paste?
Swap the bracketed parts for your own details. Include a time zone every time you mention a time.
1. Unplanned outage: first notice
We're aware that [our website / checkout] is currently unavailable, and our team is working on it now. [If you need help right away, email support@yourstore.com.] We'll post an update by [time, time zone].
Post this as soon as the problem is confirmed, before you know the cause.
2. Ongoing update
Update, [time]: [Our website] is still unavailable. We've traced the problem to [an issue with our hosting provider / a failed update] and are [restoring the site from a backup / working with our host on a fix]. Orders placed before [time] are not affected. Next update by [time, time zone].
Post at the promised time even if nothing has changed. "Still working on it" is a real update.
3. Planned maintenance: advance notice
Heads up: [our website] will be down for scheduled maintenance on [day, date] from [start] to [end] [time zone]. During that window you won't be able to [browse / place orders / log in to your account]. If you need anything before then, [place your order by X / email us]. Thanks for your patience.
4. Planned maintenance: the "down for maintenance" page
We're down for scheduled maintenance and expect to be back by [time, time zone]. Need help in the meantime? Email [support@yourstore.com]. Thanks for bearing with us.
5. Partial outage or degraded performance
Some customers are seeing [slow page loads / errors at checkout / missing images]. The rest of the site is working normally, and we're working on a fix. If checkout fails, [try again in a few minutes / use PayPal / call us at X]. Next update by [time, time zone].
6. Back online (the all-clear)
[Our website] is back up and running normally as of [time, time zone]. The problem was [plain-English cause] and lasted about [duration]. If you tried to place an order between [start] and [end] and didn't get a confirmation email, reply to this message and we'll check it for you. Sorry for the disruption.
7. E-commerce apology and credit note
Our store was down for about [duration] on [date], and we're sorry. We know some of you were in the middle of an order. As a thank-you for your patience, here's [X% / $X] off your next order with code [CODE], valid until [date]. If you were charged but didn't receive a confirmation, reply to this email and we'll sort it out.
A credit is optional: send it only if your margins allow, and only to affected customers. Our e-commerce downtime guide explains why stores feel outages hardest.
Where should you post each message?
| Channel | Templates to use | Watch out for |
|---|---|---|
| Site banner or maintenance page | 4, 5 | A fully down site can't show anything, so you need another channel too |
| Status page | All, as the running record | Host it off your main server so it stays up when your site doesn't |
| 3, 6, 7 | Skip the running updates; send the notice, the all-clear and any credit | |
| Social media | 1, 2, 5, 6 | Where people check whether the problem is "just them" |
| Support auto-reply | 1, 2, 5 | Acknowledges the issue before a human has to |
Make one place the source of truth, usually the status page, and point every other channel at it.
How often should you update customers?
There's no official standard, so treat this as a sensible default:
| Stage | When to post |
|---|---|
| Planned maintenance notice | About a week before, plus a reminder the day before |
| First notice (unplanned) | As soon as the outage is confirmed |
| Ongoing updates | Every 30 minutes for the first couple of hours, then every 1 to 2 hours |
| All-clear | Once you've confirmed the site works, including checkout |
| Apology or credit | Within a day of the fix |
Keeping to the cadence matters more than the exact interval.
What should a "down for maintenance" page return to search engines?
Your maintenance page should return an HTTP 503 (Service Unavailable) status with a Retry-After header, not a normal 200 page or a 404. Google Search Central's guidance on planned site downtime explains that a 503 tells crawlers the downtime is temporary, and that Googlebot may use the Retry-After value to decide when to recrawl the URL.
A "back soon" page that returns 200 looks fine to visitors, but crawlers see a normal page that has replaced your real content. The same post calls out sites returning 200 for downtime and placeholder pages as a common problem.
Retry-After accepts either a number of seconds or a date. MDN's Retry-After reference shows both forms. For a two-hour window, the response headers would look like this:
HTTP/1.1 503 Service Unavailable
Retry-After: 7200
Google's current guidance on temporarily disabling a site adds a few rules worth following:
- Use a 503 for short outages of 1 to 2 days. Disabling a site should last "a few days at most."
- Don't return a 503 for robots.txt. Google says this blocks all crawling.
- Keep the page light (static HTML, inline CSS) and say when you expect to be back and how to reach support.
- Confirm the status with
curl -Irather than trusting your plugin or host.
Don't leave the 503 up indefinitely. Google's documentation on HTTP and network errors says server errors slow down crawling, and that URLs which persistently return a server error are removed from the index.
On WordPress, core's update mode handles this: the wp_maintenance() function sends a 503 with Retry-After: 600 and the "Briefly unavailable for scheduled maintenance" message. If you add a custom maintenance.php drop-in, WordPress loads your file instead and skips that header, so your file has to send the 503 itself. For moves, see WordPress migration downtime.
What does a two-hour maintenance window cost?
Know what a window costs before you schedule it. Here's the math the downtime cost calculator uses.
For this example, assume:
| Assumption | Value |
|---|---|
| Business type | Online store |
| Monthly revenue | $60,000 (example figure) |
| Monthly visitors | 40,000 (example figure) |
| Month length | 30 days, or 43,200 minutes, as the calculator assumes |
| Indirect-cost multiplier | 1.5×, the calculator's built-in e-commerce setting |
| Conversion rate | 2%, the calculator's built-in e-commerce assumption |
| Window | 2 hours |
Revenue per minute is $60,000 ÷ 43,200 = $1.39. With the 1.5× multiplier for indirect costs like cart abandonment, that's $2.08 a minute, or $125 an hour. A two-hour window comes to about $250. On traffic, 40,000 visitors ÷ 720 hours is about 56 visitors an hour, and at 2% that's about 1.1 lost orders an hour.
Those figures spread revenue evenly across the day, so a 3 a.m. window usually costs less and a busy-afternoon outage costs more. Schedule maintenance in your quietest hours. The real cost of website downtime explains the per-minute figure in more depth.
How does communication change the cost of downtime?
You can't talk your way out of the sales lost during an outage. The indirect cost is different: customers who don't come back, a pile of support tickets, damaged trust. That's the part clear communication can shrink.
It helps to hear about an outage before your customers do. An uptime monitoring tool can alert you within minutes, so your first notice goes out early. For benchmarks, read the real cost of website downtime.
