This is the story of how I decided to migrate from a .net to a .com domain, and despite doing everything right losing 90% of my Google search traffic, then eventually — after months — recovering.
This story is a log of progress. I'm over a year into a painful recovery. I'll keep updating it as it progresses. Last update is September 2023. This is just over a year post-migration, and recovery is still in progress
I welcome discussing this offline, even to just commiserate. If you want to vent about your own situation, please email me!
As background, I started a website with a .net domain back in early 2020, just as the world was going into lockdown. (I'd rather not disclose it because SEOs can be competitive.)
Surprisingly, I managed to grow it into a website with 3000 page views per day, making about $100-150 of revenue daily (from ads and affiliate links) after two years.
I thought: Man, this could really go somewhere. I had my eye set on $1,000 a day. All I needed to do was increase the number of pages.
But to get there, I'd need to get serious. I'd need more content. I'd need an app. And — what seemed easiest — I'd need that coveted .com domain suffix.
So I decided to buy the .com domain. It was listed for $3,000 and I bought it for that — the investment company that owned it (Huge Domains) never negotiates. Whatever, I could pay it back in a month or two... or could I?
Context — About Me and SEO
Before the story, for the sake of context, by the time I wrote this, I considered myself to be a fairly competent technical and functional SEO. (Hah!)
There are a few kinds of people who "know" SEO. I'm a self-taught white-hat. In other words, I make good content, and then I make websites that are technically correct. A technically correct website doesn't necessarily help with search presence — it just doesn't get in the way.
I'm self-taught by reading a lot of websites, some "how-to" books I paid for, and figuring stuff out along the way. Eventually we built our websites to over 10K daily sessions, which isn't the moon, but we're at least not failing.
Technical SEO is just the foundation. As an analogy, if you're writing a great book, then having the equivalent of "technical SEO" for your book is having a catchy title, an attractive cover, easy to navigate structure, a pleasant layout, and appropriate typeface. It just helps the reader get to the information.
This was my fourth successful website — the fourth I had built with organic traffic from the ground up to thousands of pageviews a day.
I don't do any link-building campaigns. I just make good content, and people come and find it.
I started this website in March 2020, when in lockdown. My inspiration for this project (as with many other websites I've started) was a page on another website that was ranking very well and bringing in traffic. I realised "Hey, I can make a whole website out of this", and voilà, it worked.
By the end of 2020, the first year, my new website was doing a measly 500 pageviews a day. This slow growth wasn't because it was a bad idea. It was actually because I thought it had failed mid-year and I had abandoned it. It turns out that I just had mis-configured Google Analytics (d'oh).
When I realised it was doing well, I doubled down in 2021, and quickly got it to qualify for MediaVine and to start bringing in good Amazon affiliate money. On its best days, it topped US$200/day, with 2-3000 daily sessions.
Then in mid 2022, I decided to migrate it to a .com domain. I bought the .com, and prevaricated for a while, not wanting to screw things up.
I had migrated websites before, but I had never migrated an entire domain. Previous migrations had gone well — traffic initially dipped, then recovered, then increased as I added new content. Basically, the promise that 301 redirects don't penalise "link juice" seemed to be true.
So what I expected was a roughly 25% drop in traffic, then a recovery over a period of a month or so. But that's not what I got.
Why migrate from .net to .com?
Many experienced website builders, SEOs, and entrepreneurs asked me this question after the fact. Basically, they said "You shouldn't have done it." The value of a .com suffix just isn't there next to a .net, compared with the risks.
Well, I agree now. I just didn't get that strong opinion not to before did the migration. It's easy to have 20-20 hindsight, and I didn't have expert opinion backed by science.
I'm just a small business guy, not someone running a multi-million dollar revenue media company. I couldn't find an SEO consultant who could provide that kind of opinion from a position of experience.
So here's why I did the migration from a .net to a .com suffix.
Reason 1: A .com domain seems more authoritative. If I (as a user) am going to click on three sites, and one is ".info", one is ".net", and one is ".com", I'll always think the ".com" is the most authoritative. Much more than .info, and slightly more than .net.
The reason for this is just a prejudice. As a web guy, I know it is hard to get a .com domain so I'll assume whoever's building that site has put more effort into it.
But I do always wonder why a .net or a .org website has that suffix.
Then, if a suffix is something else like .info or .plus — I might just think it's spam. It used to be generally true.
More trust means more clicks, which sends positive signals to Google, which means better ranking. More trust also means more shares and backlinks.
I've seen other websites with non-traditional domains — but which have great content — lose out next to newer sites with better domain suffixes.
Reason 2: I like .com. It feels nicer to say.
This is just dumb. But all my other sites are dot coms. I didn't want to say I have this dot com, this dot com, and this dot net.
This was especially true when for a brief period it was our largest revenue earner.
It's vanity, but I admit it was there.
Reason 3: Someone told me the resale value of a website with the .com domain would be higher.
I'm not sure if this is true. But I know that as a website buyer I've definitely prevaricated before buying websites with .net domains. I wouldn't even consider buying one that has anything else.
In hindsight, I could have built the site on .net, and just sold it with a .com, and that would have been fine.
Reason 4: I thought it wouldn't hurt my traffic to do a migration. I was wrong here.
The reason I didn't think it would was a lot of research, like assurances from Google that 301s pass on "link juice".
A lot of people reported an initial ~25% dip in traffic, which recovered after a couple of months. This seemed worth it to me.
Now, I admit that you can counter all these arguments. I know, I know, there are many successful websites/businesses with other domains. I'm just letting you know what was in my head.
How I did the migration (or how my host did it)
I did my migration on 1 July 2022. I chose this date because ad spend usually drops on the first day of a quarter, so I wouldn't be losing as much. I could have waited three or six months, but that's also another three or six months of potentially lost revenue or lost ground to a competitor.
My host is BigScoots. They're a great full-service website host. I have a VPS with them that hosts three of my most important websites.
I chose BigScoots because they have such great service quality. Case in point: they did this entire domain migration for me. They do it all the time! It's just included in their service fee (I pay them $35 a month).
The way they did it was they made an entire copy of my website, changed every internal reference from .net to .com, let me verify it, and then published the website.
They did a redirect at a server level on Cloudflare directly using a rule. It was a 301 redirect.
I confirmed the new .com website worked after they deleted the old one.
Post-migration, everything seemed to work for a couple of days. Whenever I did a Google search for one of my top keywords I'd get the .net domain in search results as per usual. Clicking it I'd be immediately redirected to the .com domain as expected.
Things I checked/did:
- I confirmed the results were all 301 redirects.
- I created a new Google Search Console property and used the Google Change of Address Tool to inform Google to migrate search traffic.
- I updated the domain names in Google Analytics, but didn't change the tracking code.
- I did an Ahrefs scan of my site and made sure there were no redirect loops or other major errors.
That's supposed to be the core of it. So what went wrong with my domain migration?
What went wrong with my domain migration
Then after a day or two, I fell off the SERP.
Here's an image of how traffic looked. I left it a few days (weeks actually), thinking it might recover.
Disaster! Google Search traffic fell to a trickle. It dropped from 1000+ a day to around 200 a day, then 100, and then 50. Bing traffic outranked Google traffic. Most of my traffic started to be non-Google traffic.
What happened? I had some theories, like
- I did the redirects incorrectly
- Migrations just take a lot longer than I thought — "Google lied to me"
- My .com domain had some bad history I couldn't find
- Something about my original domain never really deserving the traffic it got, so Google took it away
Checks I did myself
The first thing I did was make sure the redirects all worked. I did google searches with
site:thenewwebsite.com in the search bar and made sure I could still pull up the new website.
I also went through old pages that ranked well and made sure they redirected. I used the Terminal/Unix command
traceroute and also various 301 checking web services to make sure the 301s were correct — they were.
In search console, I checked the new .com domain for a bad history. There were no penalties on it.
I went through Internet Archive and looked for old versions of the .com website. There were some versions dating back ten years, but they were just landing pages selling the site, nothing nefarious. The old .net domain had the same.
I was running out of ideas, and losing hundreds of dollars a day, so I thought I may as well pay a consultant to check it out.
Clean up — The SEO Consultant
It's kind of hard to find a good SEO consultant.
The reasons for this is probably because I can't afford a good one. And SEO is mostly theory, anyway — most SEOs don't have the resources to do lots of laboratory tests to see what happens to sites as we make changes.
I'm sure a good SEO, like any good consultant (like me), costs thousands of dollars a day!
So I looked on Upwork. After filtering through the usual dross, I found a pitch by one smart-sounding dude from Serbia who didn't make massive promises but said he'd do a deep analysis and see what I thought. He charged US$500 and, looking at his billing history, usually charged a lot more.
For that, he did a full Screaming Frog crawl and Ahrefs crawl, looked through Search Console for big mistakes, and made me some pretty reports with recommendations.
I didn't expect them to find much. They did find some bugs, though.
- The sitemap was broken, and the old sitemap was returning a 404 (it didn't redirect correctly). In fact, I don't think the sitemap ever worked.
- I didn't correctly redirect non-slash links to slash links. This happened a few months previously, though.
Aside from that, they found a ton of small technical errors, and prioritised them for me. For example, some broken external links, some over-sized images, etc.
In the end, I don't think they found anything critical (I don't believe sitemaps were that critical as most of my pages are just discovered through crawl). But what they did give me was peace of mind that generally the migration looked OK and there was nothing massive getting in the way.
So while I don't mind having spent the $500, I wouldn't do it again.
I did also get quotes for other SEO consultants. While they seemed competent, their fees were from $3K-5K, and I didn't think they'd do anything meaningful.
Case Studies in Traffic Loss & Recovery from Asking Around
My next port of call was googling around to see who else had been through this, and to make a post explaining my own travails on the Google Search community.
My own post didn't lead to any helpful suggestions. Someone suggested that maybe my site was a thin affiliate site, and that by migrating it, I had "alerted" Google about what it was, and they downranked me to zero. I didn't buy that, as by Google's own descriptions my site was definitely not a "thin affiliate" — it has too much unique content. And those sites had lost traffic over a month before mine had. It's too big of a coincidence that my traffic would drop after a migration. And finally, I don't think Google needs to be "alerted".
But again, they didn't find anything massively wrong.
Secondly, I looked around at other migration traffic drop threads. I found three main ones.
In all of those situations, people had posted that they had lost most of their Google Search traffic after what they thought was a successful technical migration and were at a loss as to what to do.
The usual responses weren't that helpful — basically "It has only been a month, that's nothing, it can take 6-12 months" or that "of course a new site has no traffic, Google thinks it's a new site".
None of that jibes with what many SEOs know, which is that 301 redirects should pass all the link authority from an old site, or at least the majority of it.
Again, what I expected was a maybe 25% drop, then a recovery over a period of a month or so.
I decided to contact the three people who had the most interesting posts.
I did my diligence on them first. I used Ahrefs and looked at the old domain traffic then their new domain traffic, and saw that they recovered — but only after a minimum of three months, and a maximum of 12 months!
Two of them responded and here's what they said.
(Privacy notice: Data screenshots below are using publicly available data on Ahrefs, and the details of the domains were posted publicly on internet forums.)
Case study 1 — Ancient.eu to Worldhistory.org
In this case study, the owners of ancient.eu decided to migrate to worldhistory.org. They were doing over 3M monthly sessions before migration (per Ahrefs), and had a long history of stable traffic.
Here's how that went.
Immediately, they lost nearly all search traffic and were in the doldrums for around 6 weeks. Then traffic started bouncing back, and gradually recovered. After a year, they got to and exceeded former traffic levels.
When you're running a business with employees and you're dealing with your major site, losing most of your search traffic can be very stressful. Jan posted on a number of places, including the Google Search Console Community where I found him.
I found Jan's email address and emailed him. He responded super quickly and was very helpful.
Jan said he was shocked at what happened, and was happy that they eventually recovered, but didn't do anything magical to do so. But they kept working on the site: making improvements to the structure, speeding up the site, modifying content to improve it, and doing a concerted link-building campaign.
But in his words: "But I cannot tell you how much each of these helped."
Mostly, Jan said it's important to have cash reserves to keep going as times become lean.
Case study 2 — vihara.nl to meditatieinstituut.nl
In this migration posted on GSC, traffic to a website dropped from about 2.5K daily sessions to almost nothing.
Traffic stayed in the doldrums for half a year when it suddenly began recovering, too.
I found the owner of both sites, Gertjan, and sent him an email. He responded pretty quickly too, especially considering since it was the European August vacation month.
The advice Gertjan gave was that he created four properties in Google Search Console, both for http:// and https://, and for www for both, and he ran the Change of Address tool on all of them. The traffic began recovering shortly after that.
I did the same — to a limited extent. Google has updated the way Google Search Console works since then. So I couldn't do anything for the http:// property as that's considered the same as https://. But I could do it for the www property that I created.
Like his sites, I never had a
www. property on mine — but I did it anyway. It didn't create any immediate change, but traffic did start creeping up a couple of weeks later.
Other case studies
Along the way, I found other case studies for sites whose traffic never recovered. I wrote to one but he never responded — I can imagine it'd be hard to talk about. Maybe he's just busy now, too.
Anyway, it's worthwhile pointing out that not every story had a happy ending.
Case study conclusions
Based on the above case studies, I generally concluded that I should make sure my technical SEO was actually as good as I thought it was, and then be patient.
Patience is the hard part!
Anyway, I first had to fix other gremlins.
Gremlins 1: Broken Backlinks
I next did some sleuthing of my own. I decided "hey, maybe my old domain had some link juice that never passed on". So I used four tools to find broken backlinks and started to fix them.
These four tools were:
- Rankmath (a plugin in wordpress). You can use many website tools to find 404 errors. I found a few 404 errors in Rankmath and created redirects for them. Some were for images, but some were for pages.
- Google Search Console. I used Search Console for the old domain to find pages that Google had tried to crawl but for which it got a 404 result. Many of these were strange files I had never seen, but they DID have keywords in them relevant to my site. I found those PDFs and images, created a folder called "seofiles" on my site, and created redirects for them.
- Ahrefs. I have an Ahrefs subscription and it pays off constantly. I found pages that are highly valued by backlinks but which were returning 404s, and created redirects for them.
- Moz. I signed up for a month for moz and found it didn't give me any new information on top of Ahrefs. Happy - because Ahrefs costs much less.
I know it's valuable to fix my own 404s, but not sure how much it contributed to my recovery.
Similarly to the above case studies, I have no idea how much what I did contributed to any recovery. But it couldn't hurt (I think). I actually think it was just keeping me busy while I did the more important thing — practise patience.
Gremlins 2: Search Console Symptoms
Here are some other symptoms of the failed migration. They're all from Search Console.
I'm adding this detail nine months after migration, when traffic is about at 50% of original, pre-migration levels, and traffic is at about 1000 sessions a day, with many pages getting 10-20 a day.
1. No Google Images search results.
I know that before migration, when I searched for a motorcycle model, I'd often see Google Images search results for it from my site. This is no longer the case.
2. Only one page (the home page / root page) is in Core Web Vitals results.
This to me is very strange. Despite the fact that I'm getting about 1000 sessions a day 9 months in, I still don't have more than one page in the CWV results.
I know this doesn't matter for search results, but it's just so weird.
Note: This has changed — see the September 2023 update.
The Beginnings of Recovery — Google "Testing" my site
It was very interesting watching what Google has done with my website during the recovery process.
In essence, it seems like Google is testing keywords, seeing if this new website is worthy of ranking.
Over around the first six weeks, Google would sometimes expose various keywords to users. The way this would look to a user, for example, would be one user Googles a term and gets my site in top position, and another user doesn't get my site at all.
I think what Google does is experiments. They're exposing my website to users to check various things — not sure what, but it could include
- Is this website as good as the one referring to it? (From the 301 or search console migration)
- Is this website still one that users would like? (Good content, speed, presentation)
The way this would look is that for a given keywords sometimes I'd be in #1 position, and sometimes not there at all. Very weird behaviour.
Google kept doing this with many of my keywords on and off. I'd always be in the #1 position when they did an experiment.
On August 8 after around five weeks, I noticed a sudden uptick in traffic. I saw it was an uptick in search traffic, and specifically an uptick in Google (and not other search) traffic. Could this be the day?
A day or two later, traffic held steady at the new baseline, and I realised that maybe partial recovery WAS happening. In one sudden day, my Google search traffic had skyrocketed. I went from about 50-100 sessions of Google search traffic a day to around 7-800 sessions per day.
It still had a way to go, however. I needed it to get back to 2K sessions a day. I expected that to take many months more — maximum a year.
After the initial recovery, I noticed that my average position for target keywords in Search Console was very high — around 1-1.5. Even for the whole site, the average position was 2. This means that Google was still just showing my website for the very best, most likely relevant keywords.
For other websites I have, the average position is somewhere between 10-20. This is healthy — there's a mix of positions. My site may show up in position #1 for its target keywords, in positions #3-10 for long-tail keywords, and beyond that for other random keywords that Google thinks maybe the site is relevant for.
I noticed that over time, Google would continue to experiment with more keywords.
A few months in, mid October (3.5 months after migration), traffic is still down. My average position is still around 2. There's also some seasonal drop as my traffic is quite summer-heavy. But I'd still have expected a natural uptick by now — traffic is far down over last year, for example.
More recovery in January 2023
Just a quick note to say I got another ~50% bump in traffic in Jan 2023. It's still far off original numbers.
I've also been periodically doing a search on Google for assets on my old domain vs my new one, like this: "site:olddomain.net".
I've noticed a strange thing.
- About 2 months ago (four months post migration), there were a few pages of results.
- About a month ago, there was one page of results.
Over the last two weeks, that has been dwindling down. Right now there's just TWO results.
At the same time, I do the same search for my new site ("site:olddomain.com") and that is gradually ticking up. It's not a scientific test, but I do like seeing the old one go down and the new one go up.
However, even now, just over six months post-migration, my "position" in Google Search Console is still around 2.5. I'm only ranking for top keywords, not long tail keywords. I'd like to see that average position go down to 20-30.
More Recovery in September 2023
Some big changes have happened in September 2023.
Firstly, I noticed traffic was going up over the last few days. I started getting around 1400-1500 sessions a day, which isn't awful, especially since I haven't touched this website in months now.
Curious, I wondered what was going on in Search Console. And wow! Some HUGE changes there.
Average position, for the first time, has dropped below 2. I've started getting a "noisy" ~10 average position. It's not a start, but it's a sign I'm getting a lot more keywords.
And yes, I have a lot more keywords to my name, now. Whereas Search Console previously tracked under 1000 keywords, it now tracks over 1,000.
Other things are tracking well too. Doing site:newsite.com give me many more Google Images results, and many more results on the main page.
And one of the earlier weird "gremlins", the fact that I had no pages passing CWV, changed suddenly.
It seems like Google is finally acknowledging my website exists and isn't a bad one, and is now crawling it.
It also seems Google still has to sort things out, as those numbers are bouncing around, but it seems like the migration is finally underway, a year after starting it.
For this site, this represents around $50K in lost revenue, even in a bear case that assumes it was hit by algo updates and so on. But, live and learn, that's the price of an education.
Side note — Why can Bing get it right?
One of the things that bamboozles me is that Google didn't understand the migration at all — but Bing did right away, and continues to, with featured rich snippets for tons of keywords.
Another couple of technical things
I read the article "Move a site with URL changes" again, wondering if I had missed any details.
I realised I hadn't done something from it:
"Be sure to verify all variants of both the old and new sites. For example, verify
example.com, and include both the HTTPS and HTTP site variants if you use HTTPS URLs. Do this for both old and new sites."
I didn't think that it was terribly important for the new site, because I did redirects from http to https, and redirects from www to without www. So I had never done it. But following that advice, I authenticated the new site for those protocols. I did this on October 17, 2022, a few months after migration.
Nothing significant happened after this point.
A dark thought... Was this traffic drop inevitable?
Looking over my 2022 search console data, combining the old site's data with the new one, shows an interesting picture.
Yes, there's a huge dip right when migration happened. But I can't help but wonder if this downward trend was always inevitable.
I drew that arrow on myself, of course. Curve fitting is dangerous, so don't be fooled into thinking it's the only truth. But it's a possibility.
While my traffic might have dropped anyway, there's too much other weird stuff going on (e.g. zero images results, very high search position, no long-tail keywords) for me to think that this is the only thing that happened.
During that period (July / August 2022), Google said they released the product reviews update. This has nothing to do with my site as it's not a product reviews site.
I also examined the trend in search position for the whole of 2022 (I migrated mid-2022).
As you can see, the average position was dropping in the months before migration. There was also a commensurate drop in traffic as you can see above — traffic was becoming much more focused on the core content (I suppose this was probably related to the "Helpful Content" updates).
But the average position dropped a lot since then. Now, the site only ranks for the titles of posts, and rarely for any of the content. I mean, why bother having content, then?
What I should have done
In retrospect it's very hard to say what I "should" have done. It's easy to have 20-20 hindsight.
Probably the best option: Keep the .net, but own the .com as an asset to sell with the website.
The main problem is that I still don't know why my traffic plummeted. I'm in the minority of cases. Most people get a small drop, then full recovery.
Unless I know the actual reason, it's impossible to know what caused the traffic drop. So if I did this again, I'd be tempted to do exactly as I did this time!
Anyway, for now I'll keep experimenting, building the site, and hoping traffic will recover.