I think one of the most important questions companies should ask before starting any project is: "Will doing X provide value to our customers or differentiate us in the marketplace?"
If the answer is "no" then X shouldn't be done. Simple enough, right? There are several companies that really get this. However, this is just the tip of the iceberg.
A more difficult issue arises when doing X is a prerequisite for a project that actually will differentiate you in the marketplace. Most of the time this involves procuring a server, installing it in your data-center, and hiring people to manage it. All of this work requires vast amounts of time, energy, and resources from the smart people in your company without providing any real business value.
The solution to this problem is simple: let someone else handle the details for you. Why deal with all of the issues that come with owning your own data center when there are several cloud providers out there who can do it better, faster, cheaper, secure(r), and more reliably than you?
Most businesses must create justification and obtain approval before moving a system to the cloud. I think that's backward. You should have to create a business case for keeping something on-premises. Cloud-first should be standard.
Free yourself up to focus on the things that truly matter.