Do you still control the old domain?

Don't rely on the "site move" tool to talk to the crawler.

You need to 301 redirect your pages from the old domain to the new domain.

I see this all the time with clients who skip some basics when changing domains.

I'm happy to talk you through this via HN comments

The old site was hosted on Github Pages, so there isn't a way to do a server-side redirect AFAIK, but I replaced each page on the old site with a client-side redirect in the following format:

    <link rel="canonical" href="new page URL" />
    <meta http-equiv="refresh" content="0;URL=new page URL">

This might be helpful:

https://dev.to/steveblue/setup-a-redirect-on-github-pages-1o...

Also, these redirects won't help with your Google results until next time they crawl the site, so you may have to wait a few days.

Meta refresh redirects just don't work I am sorry to say.

Hypothesis: CoA (change of address) request expired, redirects dont get recognized or are not here any longer.

Did you use a custom domain for old version?

When you say it was on github pages, do you mean it was on a github subdomain like example.github.io ?

Yes.

Were you (1) using example.github.io as a CNAME with another domain, or (2) were people going to example.github.io directly?

I'm still not 100% clear on how the previous domain was configured with github pages

if it's 1) then you should be able to set up redirects for the domain using something like cloudflare to serve the 301s to the old domain.

if it's 2) then there really aren't any more "simple" solutions that you can implement purely technically. You've already added the canonical reference and the meta refresh. You could also add an A tag link to "Download the software on our new website" in addition to the other measures. Users are already sent over via the meta refresh, but the link will encourage crawlers who skip that tag to find you as well.

If your goal is to show up when people search the name of the software (example: "iterm2"), then the solution is to register on every possible social media website you can find and link them to your new domain. In addition to that it's worth using an online PR service to distribute a press release with a link to your new site announcing the new domain.

If there's a more generic term (example: "mac terminal software") then you've got a longer road ahead of you with a regular SEO campaign.

It's the second one.

Most of my previous traffic was from direct searches for the name of the software, so I think I'll try the social media approach - thanks!

Sounds good, good luck. Assuming your name is relatively unique then this should be something that resolves within 4-6 weeks. If it takes longer you might need to do some high profile PR to get the name out there and linked to the right place.

I think it sees your new domain as just a blatant copy of the old one. Quickly set up a server on the old domain and redirect to the new domain, so you tell Google that the new one is the authoritative site.

Sorry for not including this in the original post, but I did replace the old site with a set of redirects: https://news.ycombinator.com/item?id=23884449

I think the person above is telling you to set up a webserver to 301 redirect pages from your old domain to your new one. I think that will tell Google that you have chosen to move the site to a new domain, not just copied the site off someone else

The old domain was a *.github.io domain (which I migrated to a custom domain), so I don't think it's possible to use my own server for it.

Google did pick up the relationship correctly, at least in the beginning - the new domain got the same ranking as the old one, and the "backlinks" section of the search console shows links to the old domain under the property information for the new domain. So it seems like this type of redirect does have an effect, unless it somehow expires after a few months and a 301 redirect doesn't?

Yes, this really seems worth a shot, if possible.

If the old site was on a domain that you own, remove the CNAME to github pages, and point it instead to a server you control, that returns 301 redirects to the new site.

The 301's are a standard way to do this, while the meta refresh tags are dubious at best.

> There seems to be no way to contact an actual human, and since nothing on the site has changed, I can't think of anything to revert.

There are some Google people on Twitter responding to concerns such as yours. @JohnMu is one of them. You can send a private message to him about it.

We've experienced similar issues, but possibly with a different cause, see https://seotool.ee/indexed-not-submitted-in-sitemap

Once Google doesn't like your site, it's not easy to get it back. It involves being patient and trying to understand what's wrong. Sometimes though, there is nothing to be done and your site just comes back up. Google is a big black box in this regard, and doesn't give any message in search console about what may be wrong.

Why should Twitter be the space for Google's customer support service? This is absurd

The actual answer to your question is probably that it's better for Google if it's difficult to reach a human without sufficient effort, since it lowers their support costs. Providing a escape hatch in their support process might mean that too many people use it.

Also, the person in this case is neither a Google user nor a customer, so there's a particularly small incentive to provide easy support.

Maybe a lot of it comes down to Google's tendency to rely on AI, but the fact that so many companies have crappy support to me indicates that it's a harder problem than one might think.

An indexed site is not Google's customer. From Google's perspective, the Web that they're indexing is just a resource that they can choose (more or less) to do what they please with.

Advertisers, on the other hand, are Google's customers.

I think I actually read your post earlier when I was trying to figure this out, so thanks for that! The symptoms do seem pretty similar, although the fact that the timeframe is different makes me think it's a different issue.

> There are some Google people on Twitter responding to concerns such as yours. @JohnMu is one of them. You can send a private message to him about it.

Thanks! I'll try that if I can't find a solution.

I’ve done this a few times. Even if you do 301 redirects properly, it still can take a month or two for google traffic to day get back to previous level. You might just have to be patient.

I have tried this 'perfect switch over' for a major domain after many months of planning. Everything worked fine for a month and then the traffic dropped to 30%. The traffic went back to normal level after few months.

As usual there was no response from Google/ Forums etc. All I could speculate is the loss of backlinks. It takes months to reach the original level.

Keep improving the site and you should be back and kicking in 6 months.

This doesn't help you now, but why didn't you keep the old site and just 301'd to the new page for a year or so?

There is a really good spec by google on how to move sites: https://support.google.com/webmasters/answer/6033049?hl=en Googles best spec in my eyes. 301 (what, how, and what not) change of address request, sitemap.xml,

Always... Always setup your site with Google search console and submit your sitemap to it.

You don't need to setup the old site on a server to 301 everything. You can just CNAME the old domain to the new one. And yes, you can do this with a GitHub.io domain, just have a .CNAME file in the repo. Then setup 301s on the new domain for the old paths to point to the new ones. The Redirections plugins for WordPress make this easy as hell if you are running WordPress.

hopefully this will help you out and get your new site index in no time

I've been through this multiple times in the last couple of years, it's worked well for me to move domain authority around in this manner. Maintaining a 301 redirect map on the original subdomain should work fine (probably better than meta headers on reinstated pages, I haven't heard of Google paying much attention to that) and you should be able to claw back your rankings. Running something like jekyll-redirect-from within your old github pages project looks like the ticket. Good luck!

I have had something similar happen to me. It was resolved by getting more sites to link into mine. Getting new incoming linkes just immediately set things right.

I think for your software product, you may have underestimated the value of GitHub.io domain. Your post makde it sound like it went from one custom self owned domain to another, that's not the case at all. I would move it back to github.io, and build your other domain over time until they are 1&2 in results (or there abouts)

Had exactly this experience. Had some very popular content up on a platform site, ranked very well. Wanted to move to my own domain. Found it nearly impossible to show up in search results at all. Even searching for the exact title wouldn't bring it up despite seemingly being "indexed".

> I feel it's very risky to depend on Google for your traffic at this point.

Yes, but what choice do you have?

This might help if you can edit your html's meta tags on your old domain: https://moz.com/learn/seo/canonicalization

Put up your old domain with content and then do a rel canonical from each page on the old domain to the corresponding page on new domain.

I faced same issue couple years back. Hope it helps.

Sorry if the same answer has been given before mine.

So the old domain should have the same content as the new one? Right now, I have blank pages on the old domain with a canonical link and a redirect. I could try duplicating the content, but it seems like having duplicate content might cause Google to penalize me even further?

Sounds like you need an SEO audit - contact someone who understand technical SEO that can provide you an audit. It may be a small/simple issue causing your website to be de-indexed.

How did you drive traffic to your site in the first place? I'm starting an online SaaS but having trouble getting my first users.

It helps to write more. A lot more. Getting traffic from Google only works if you have a ton of content.

Your video is great, but doesn't tell me why I would use this over Trello or something similar.

- "Pricing is free" feels ominous (to me), say it's a free beta or something instead.

- It's not entirely clear what "roadmap" means on your landing page. Is it your product roadmap?

- Your colour contrast hurts my eyes (maybe I'm weird), particularly black + purple.

Thanks for the feedback! Perhaps I need an animation showing how when you click on a card it takes you into that cards Kanban board. (The feature differentiator from Trello)

As for the color contrast I plan to add more options in the future such as light mode.

What are you doing to speak to your customers?

I've tried places like HN and a little bit of Google Ads (too expensive). I don't know how to reach my customers...

Depends on your niche. I'd recommend directly engaging on Twitter, and launching a podcast and/or YouTube channel. Look at Reddit and see if there are any subreddits you can start contributing to. Reddit advertising can be somewhat effective, but some subreddits' ad inventory are bought out often.

Give us more info: current domain, old domain

Could be lots of self made issues.

How do the 4 SEO Tests look like? Outlined here in the "is it me or is it google" section https://medium.com/@franz.enzenhofer/what-to-do-when-a-googl...

Thanks for the suggestions. I think those 4 all look OK. The site content is purely static HTML (and some images) that hasn't changed significantly in months, so I'm guessing that's not the issue.

I'd rather not post the site URL here; is it OK if I email you?

Honestly, as with most other tools that you have to care about but don't want to: just follow the spec https://support.google.com/webmasters/answer/6033049?hl=en

DuckDuckGo doesn't really have their own index. They mostly just fall through to Google and Bing for their actual results. It's cool they supposedly don't track you, but they are really just a proxy rather than a full fledged index themselves.

Bing/DDG has the site indexed correctly, and I get a decent amount of traffic from them, but Google's market share is large enough that it's not really feasible to just ignore them.

(And I'm not buying ads on either one, this is about organic ranking)

If OP's potential users are Google users, optimising for DDG (bing) won't do anything. The fact is most people are using Google and if you only care about DDG you are throwing away most of your potential visitors.