[davnicwil]

I heard some advice recently, I think in an article here on HN, to just use a big company registry where it's not their profit center, indeed may even be netural to loss making, to drive other business.

This makes total sense to me. I'm not saying it solves all problems but it eliminates so many of them, including a meta problem: the risk of new classes of problems being unexpectedly introduced (by say a private equity acquisition or similar).

If domains themselves are the profit center, you are likely in trouble if there's really any incentive for them to make incremental revenue in such a competitive market. Doing 'the right thing' just of course will not factor in if there's really no reputation at stake.

[jasongill]

The only problem with this is that having such a complex function as a non-core business unit makes it ripe to get rid of, or try to find a way to make it a profit center.

Cloudflare, for example, removed the ability to change the nameservers for all domains registered with them.

Google got tired of being in the business and sold it to Squarespace.

Being a domain registrar is a total PITA and is not for the faint of heart, so it's the sort of thing that any business that takes it on as a non-core function will eventually tire of.

[duskwuff]

> Cloudflare, for example, removed the ability to change the nameservers for all domains registered with them.

Cloudflare never offered that ability. From the time they started offering domain registrations, it was always with the caveat that the nameservers would be fixed to Cloudflare's.

[jasongill]

Correct, I guess I should have clarified that I mean they removed the ability that most registrars have, in an attempt to make domains into a revenue driver for their primary product.

[jurgenburgen]

Remove implies they previously offered it. I would say they just offer a more opinionated product.

[ccamrobertson]

As someone still smarting from the loss of Google Domains, I would consider this advice carefully.

When domains are not the profit center the company might just arbitrarily turn them off as a feature.

[crossroadsguy]

I think a middle ground could be to use a great "domain registrar" and be prepared to move to another one when the last one stopped being great or sold to someone not so very great. And so on. Luckily this doesn't happen every month, or every year.

[danaris]

The problem with that is, there's not necessarily a signal for "stopped being great or sold to someone not so very great" besides "they gave my domain away to a scammer because they asked."

[crossroadsguy]

There ares some. Their reaction to events, their ownership structure change, support culture change etc etc. There isn't a perfect way to find that out, but there are indications.

[apothegm]

Most of us don’t interact with our registrars often enough to notice changes in support culture, and who has time to be constantly checking for changes in their ownership, if that information is even available? If you’re spending that much energy on your registrar and your work isn’t as a domain trader, something’s wrong.