Hacker News Reader

Comment by paulgb | Hacker News Reader

Until Domains, I mostly trusted google to keep running a paid b2b product with lots of users. Most of the famous products they've killed have been ones that either didn't bring in revenue, or never became popular.

After the Domains announcement, I can't in good conscience bring myself to spin up any new GCP services, and have started leaning more heavily on AWS. Part of me thinks that Domains is a signal that Google is planning a larger exit from the cloud business, given the GCP / Domains integration, and the absolutely atrocious comms from Google around Domains has not reassured me at all.

View on HN

> that either didn't bring in revenue

Domain registration is famously difficult to profit from unless the registrar is able to upsell

They used it to upsell Google Workspace, as well as their own TLDs (.dev, .app). And although they didn't upsell GCP through it, the integration between the two was a nice (small) incentive to use GCP.

What makes it so difficult to profit from? A yearly subscription to maintain some text entries doesn't seem that crazy to me.

The costs are known, the competition is high, the margins are low. You still have to deal with support/abuse. There is no real “high volume customer”.

There’s a lot of contractual obligations and edge cases in domain names, and a lot of potential for fraud and disputes. It’s a business with high legal fees.

Also how domains should be managed evolves with time. For example GDPR has a huge impact on WHOIS, especially since the ICANN took so much time to acknowledge that in their specs, so there was a time where it was impossible for a registrar to be compliant with GDPR and ICANN rules. It’s still not 100% clear how it’ll be handled by RDAP (future successor of WHOIS)

And last but not least, some registries for ccTLDs are real pains in the ass to work with, and won’t hesitate to slap you with big fines if you make any mistake.

They killed GCP IOT core as well. Forced us to find another vendor.

We've had multiple multi-cloud customers signal a desire to move away from GCP for years now because savvy CXOs are just fed up with the unreliability. When you're looking to resolve a ton of tech debt that you have because of the ground constantly shifting beneath your feet for 1-2 decades... well, you just get fed up and want the reliability of serious enterprise vendors.

It makes innovation much harder due to the downstream effects, but at least I'm not carrying more and more risk forward thanks to Google/Alphabet's ADHD approach to product, brand, and support.

What exactly do you mean by unreliability?

Not having to ask "will the product I rely on today exist tomorrow?"

Probably the constant deprecation of API / SDK from Google, you have to keep up with their crazy deprecation policy. On the other side I still have some APPS calling AWS API since 10y without any issues.

"When will Alphabet opaquely and unilaterally depreciate this product upon which we've come to rely, as they do with many other products all the time?"

Alphabet is a hopelessly disorganized bureaucracy whose leadership appears, at best indifferent to customer needs, at worst hostile to customers that don't toe their line.

It's just a risk calculation you have to make when you're spending millions to build systems of record upon which your enterprise will depend for years to come.

I don't know if its fair or even accurate – but it's the feedback I get from client CTOs and CIOs at sizable companies.

Reply on news.ycombinator.com