Most people submit their site to Google Search Console and stop there. That was us too, until today, when I found that the thing we registered back in July had quietly gone stale.
Why bother with Bing at all
Bing’s own traffic is the boring answer. The interesting one is that other products read from Bing’s index, and ChatGPT Search is the one that matters right now. If you are only registered with Google, you are absent from a lookup path that a lot of people now use instead of a search box.
It is free, there is no paid tier, and the setup is shorter than Google’s because you can skip most of it.
Setting it up
Go to Bing Webmaster Tools and sign in. If your site is already verified in Search Console, use the import option. It pulls the site across and carries the verification with it, so there is no file to upload and no DNS record to edit. That takes about two minutes.
Add it manually only if you are not on Search Console. Then you get the usual choice: drop an HTML file at your web root, add a meta tag, or add a DNS record.
One thing worth knowing about those verification files. They have random names, so months later they look like junk sitting in your web root, and people delete them while tidying up. Do not. Search engines re-check them, and deleting one un-verifies the site.
IndexNow, if you publish often
IndexNow is an open protocol that lets your site tell Bing the moment a URL is published, updated, or deleted, instead of waiting to be crawled. You generate a key, put the key file at your root, and ping an endpoint on change. Most SEO plugins do the whole thing for you once you paste the key in.
It is worth setting up if you post more than occasionally. It does nothing for Google, which does not participate.
The part I got wrong
Every one of these tools asks for your sitemap. You give it once and never look again. Here is what ours was pointing at.
Sitemap: https://curiogan.com/Sitemap.xml
The address loads, which is exactly why nobody questioned it. Open it and it is not the sitemap WordPress generates. It is a separate static file somebody hand-wrote, 555 bytes, with six addresses in it: the homepage and the first five posts.
The real sitemap had thirty.
| What robots.txt advertised | The actual sitemap | |
|---|---|---|
| File | Sitemap.xml (hand-written) | sitemap_index.xml |
| URLs inside | 6 | 30 |
| Last updated | frozen on 22 July | every time I publish |
So for over two weeks we told crawlers this was a six-page site. A sister blog on the same server was fine, so this was our configuration, not a server fault.
Did the other posts stay out of search?
No, and I want to be straight about that rather than make the mistake sound worse than it was. I went through the server access logs for the same period: Googlebot fetched 27 of our 29 article URLs, all returning 200. It walked in through the archive pages and the links inside posts.
A sitemap is a map, not the only door. A stale map does not make the house disappear.
Fix it anyway. There is no reason to leave a reason for new posts to be found late, and anything sizing up your site quickly sees six pages instead of thirty. The fix was one line: point the Sitemap directive at the address your CMS actually generates. With Yoast that is /sitemap_index.xml. Then open it and count.
What to check after you register
Submitting takes five minutes. Checking takes five more, and skipping the second half is how you end up where we did.
- Open
yourdomain.com/robots.txtand read it. Confirm nothing added a blocking rule. - Copy the address on the Sitemap line and open that.
- Count the URLs in it. Compare against how many posts you actually have.
- Confirm the verification file is still sitting where you left it.
Step three is the one that matters. That is precisely where I failed: I checked that the address responded and never looked at what came back. A 200 means a file exists. It does not mean it is the right file.
This connects to URL structure too. Change your permalink scheme later and your sitemap and your live addresses drift apart, which is the argument in the post on picking a permalink structure once. And if you are wondering whether optimising for AI answers is a separate discipline from this, I went through Google’s own documentation on that in the piece on SEO, AEO and GEO.
Registering looks like a one-time job. What you registered is what goes stale. Re-reading those four lines once a quarter covers it.