Three weeks after moving from .ai to .com. The numbers.

In post 1 I said I'd post the numbers whichever direction they went. Three weeks in, here they are. Two of them surprised me.

The question nobody had an answer to

Post 1 ended on this. Copilot had cited BrandScreen 487 times, every citation pointed at a brandscreen.ai URL, and I couldn't find anyone who had written down what AI engines do with a redirected domain.

Here's what happened.

AI citations follow a 301. It took twelve days.

The .com started getting cited on August 10, three days after the redirects went live. It crossed above the .ai on August 19. Over the last seven days: 90 citations to the .com, 71 to the .ai.

The old domain hasn't gone to zero and probably won't for a while. But the handoff happened, on roughly the same timescale as Google's, and I didn't do anything except leave the redirects up.

So if you're moving a domain and worried about AI visibility, it transfers. Budget about two weeks.

The number that looks like a win and isn't

My Google impressions went from about 87 a day to about 870 a day.

That's a 10x and it means almost nothing.

One page, my domain checker tool, accounts for 14,339 of 14,798 impressions. That's 97%. It sits at position 90 for queries like "domain checker" and "check domain name quality," and it has produced zero clicks.

Take that page out and the new domain is doing 27 impressions a day. The old one was doing 87.

So my traffic went down, and a single page ranking at the bottom of page nine covered it up with a number that looks like growth.

This is the second time in two posts I've caught this. In post 1 it was Google exploring 27 newly indexed pages and producing an impression spike with no clicks. Same thing here, bigger, lasting longer.

I'm starting to think impressions are close to useless as a metric for a small site. A handful of pages surfacing at position 90 will drown out everything real.

But the impressions are telling me something

I said the 10x is meaningless. As a traffic number it is. The queries behind it aren't.

Here's what my domain checker page is surfacing for:

Query	Impressions	Position
domain checker	518	86.1
domain check	455	91.1
check domain	370	91.6
website domain check	206	89.9
check domain name quality	175	84.9
domain trust checker	175	85.3
how good is my domain name	127	90.5

That's real demand. People are searching for the thing I built, in volume, and Google has decided my page is relevant enough to show. It hasn't decided the page is good enough to rank.

Position 90 is results 891 to 900. Page nine. Nobody goes there, which is why 14,339 impressions produced zero clicks.

So I wouldn't say we're getting traffic. Google now knows this page exists and roughly what it's about, and has put it at the very bottom of a competitive query. That's the starting line.

The useful part is the query list. I didn't know "domain trust checker" and "how good is my domain name" were things people typed until Google told me by showing them my page. I have four other tools, trademark, entity search, social handles, brand score, and no idea what people search to find those, because those pages aren't surfacing for anything yet.

My instinct is to build more tool pages and get them surfacing too. I'm not sure that's right. The best page I've ever had isn't a tool page. It's a blog post about how to check if a name is trademarked, 2,090 impressions at position 35 on the old domain. Fifty-five positions better than the tool page, from writing instead of building.

So what I'm stuck on is whether the tool pages are thin. If a page is mostly a search box, position 90 is Google saying so, and shipping three more like it gets me three more pages at position 90.

I'd rather find out why one page is stuck at 90 than build four more that are.

The prediction I made, and where it stands

Post 1 said this:

My domain checker tool page was at position 39 on the .ai. On the .com it came back at position 90. I expect it to climb back. If it is still at 90 in mid September, something is wrong and I will say so here.

It's still at 90. Average position 90.09 over the last two weeks, no upward trend.

Mid-September isn't here yet so I'm not calling it. But it isn't climbing, and the volume of impressions at that position looks like Google has settled on where it thinks the page belongs rather than still working it out.

What actually moved

Clicks barely changed, and clicks are the only number here that matters. Before the switch: 43 clicks over 56 days, about 0.77 a day. After: 20 clicks over 17 days across both domains, about 1.18 a day.

Up slightly. On numbers this small that's noise.

The brand query re-pointed immediately and stayed there. "Brandscreen" is position 1 on the .com with a 36% click-through rate. That was the one thing I was actually worried about and it took less than a week.

Indexing is where the real damage is

The old domain had 335 pages indexed after two months. The new domain has 115 after three weeks.

Google knows about more URLs on the new domain than it ever did on the old one, 430 against 353, and has indexed a third as many.

Here's why the other 315 aren't in:

Reason	Pages
Discovered — currently not indexed	296
Duplicate without user-selected canonical	7
Excluded by 'noindex' tag	6
Not found (404)	4
Page with redirect	2

296 of the 315 are one category and it isn't an error. "Discovered — currently not indexed" means Google found the URL, queued it, and hasn't been.

I pulled the drilldown to see which pages. 290 of the 296 are my programmatic name pages. The last-crawled date on every one is 1969-12-31, the Unix epoch, which is Google's way of saying never.

Google didn't crawl those pages and reject them. It has never visited one of them.

Nothing on my site links into that cluster except the pages themselves linking to each other. The only way Google knows they exist is my sitemap, and a sitemap-only URL on a three-week-old domain with 141 referring domains is about the lowest priority thing you can hand a crawler.

The queue went from 324 to 296 in eleven days. About three pages a day. At that rate it clears around Christmas.

Here's the part that told me it isn't just the name pages. Six ordinary marketing pages are stuck in the same queue. /about, /pricing, /contact, /privacy, /blog and a blog category. Those are linked from my own navigation, on every page of the site, and Google still hasn't visited them.

So this is a new domain getting very little crawl budget, and me handing it 417 URLs at once and expecting it to work through them.

It's a bit discouraging that I'm still not indexing, but am trusting the process and giving it at least 60 days for the full indexing.

The thing I'd do differently: I moved the whole sitemap on day one. If I'd submitted the pages that matter first and let the programmatic cluster follow once the domain had some standing, I think the six pages I actually care about would be indexed by now.

What I'd tell someone doing this

The migration was a non-event. Redirects went in clean, brand search re-pointed in days, AI citations followed within two weeks, and nothing broke that a redirect caused.

I think that we made the right choice and we have a plan, but the jury is still out if it was worth the $10k that I spent on the domain.

Next

Post 3 is what broke. A React provider that had been dead for three weeks. A middleware file that had never executed in production. A contact form that accepted submissions and delivered nothing for 81 days. Share cards that said .ai for a fortnight because the wordmark is split across two spans and my grep couldn't see it.

None of those were caused by the migration. All of them were found because of it.

If you're picking a name right now, BrandScreen checks the trademark, the domain and the handles before you commit. It's free to run one.