This bit in the tweet chain was more interesting to me:
Google Domains *was* profitable, I confirmed internally. Growing extremely fast (~$120M ARR now, ~growing $40M/year)
Google Domains *was* profitable, I confirmed internally. Growing extremely fast (~$120M ARR now, ~growing $40M/year)
We don't know what the expenses were. It's still possible it wasn't profitable.
The only real expenses would be dealing with customer service and card fraud. While it’s possible that those costs could be substantial, Google isn’t exactly known for throwing a lot of money at customer service.
So it doesnt matter, i speculate, if it was earning $1M in profit a year for 10M domains if they need to employ 50 call center staff - and for the next 10M domains they need to hire another 50 call center staff— I think they would rather pull the plug.
Which I personally would never do, but, it is in line with other decisions they’ve made and how I have heard key people talk about googles leadership externally.
This is 1000% true. The culture is that diminishing returns means linear scale will eventually turn into a cost center. Execs only care about 10x initiatives. My understanding is that Larry personally canceled gFiber b/c the margins, while profitable, were too thin.
It's a huge pain in the ass to deploy fiber. The cities that got Google Fiber mostly had to bend over backwards in order to do so. There's a lot of logistical (and legal) factors that just don't exist for pretty much any other Google product.
Like what approvals does Google need to release Bard? But you could make pages of approvals needed to deploy fiber to a city.
Well, for one they still need Ireland's DPA approval before they release it to the EU.
The point is that fiber would require a lot more approvals than many of the pure software projects. Also afaik, fiber never launched in EU and exists in many less locations than Bard.
That makes sense given "nothing linear", but I can't understand why they would have thought it could have been better. Same for domains. In what scenario would they grow exponentially?
like someone else in this thread said; in theory it’s a global distribution of 1kb in a database and a user interface.
if you are the best at AI, maybe you can solve the support burden. Or at least thats what they attempted.
Nevermind that non-linearity should have been achievable[0] via lobbying at the state/federal level.
[0]An advantage only really achievable by a company as large as google, with money and weight to throw around.
Surely that's a lot more trivially fixed with a small price bump than throwing the entire revenue stream & user sentiment into the trash?
Google Domains is currently charging $12/year for a .com domain, which we know from cloudflare the "at cost" price of that is $9.15/year. Comparing the $12/year to other providers it's kinda middle-of-the-road, not especially cheap or anything. Shopify is $15/year while porkbun is ~$10/year, so $12/year certainly appears to be a price that has adequate margin to either be profitable or if not only needs a small bump to be profitable?
The ROI on Domains is likely to be no where near that of, I don't know, GKE or BigQuery. You can argue all you want about how those are totally different projects with totally different considerations. But at the end of the day, unless Google wants to have a strategic investment in maintaining an end user facing Domain Registrar service, the ROI is all that matters. They are focusing and consolidating all their "developer facing" services in GCP. What makes sense will eventually move there, what doesn't will get cut.
ARR and YoY growth tell you nothing about profitability or the ROI. Google never shied away from deprecating services that they don't care about to focus on things they care about.