> look much more competent and friendlier to customers

How exactly are you measuring "more competent"? Once you pay for support levels you also don't have the problem that Google is a black hole where you don't get any answers like if you are just a free user on Gmail, YouTube,...

It also depends on which services you use, if you are just running everything in Kubernetes and are not locked into proprietary services of any cloud provider it's not that hard to move between clouds. I find it a bit odd to make blanket statements about a company based on some consumer products or Google Reader being killed off years ago.

Azure is not competent or customer friendly at all. I've been using Azure on and off since its very first days and despite me telling every client that they're better off by using AWS or GCP because it will cause their dev teams less headache, and cost them overall a lot less in engineering cost (time spent by their engineers fighting with Azure instead of building a product) there's still every so often a non technical idiot who got sucked off by some Microsoft Partner manager and trapped in this shitshow of inconsistent, slow, buggy and constantly broken cloud services which Microsoft offers.

Not discounting your experience, but we have clients in all three, and our own experience is that Azure support and service is _significantly_ better than AWS and GCP, to the degree that we primarily recommend Azure for that reason.

My team has also chosen Google Cloud.

Here's two reasons: 1. They have I think the most mature manged kubernetes service. 2. They have a managed Airflow offering called Cloud Composer.

If you are just using the basics (VMs, managed SQL, storage), then it probably makes sense to pick primarily based on factors such as price and customer service.

Once you get into more complicated use cases like say big data, distributed systems, or ML, there are important differences between what the cloud providers are offering.