Enterprise Switching CostsWide moat

Microsoft's deepest moat is inertia — the good kind, the kind that works in the owner's favor. A large company runs its email, its files, its logins, and a great many of its line-of-business applications on Microsoft's stack, and every year of use ties another knot. The result is a switching cost so high that migrating off is not experienced as a purchasing decision at all. It is a multi-year, high-risk, high-disruption project with no obvious upside for the executive who would have to champion it, and considerable career risk if it goes wrong. So, overwhelmingly, it simply does not happen.

The single most powerful thread in that knot is identity. Active Directory, and its cloud successor Entra, is the system that decides who inside a company may log in to what¹ — and once an organization's entire notion of 'who its people are and what they may touch' lives inside Microsoft's identity system, the company is bound in a way that touches every application it runs. Owning the login is very nearly owning the enterprise, because every other piece of software must ask Microsoft's system who the user is before it does anything at all.

Then there is data gravity, a quieter but relentless force. Years of documents, emails, and records accumulate inside Microsoft's applications and storage, and data, once it settles in a place, exerts a pull on everything around it: the new tool a company buys must work with where the data already lives, which usually means it must work with Microsoft. Each new integration deepens the entanglement, until the stack is woven into a thousand distinct workflows that no one fully maps and no one dares disturb.

The depth of that integration is worth picturing concretely, because it is where the abstraction of a 'switching cost' turns into a wall you can feel. A large enterprise does not merely use Microsoft's programs; it has wired them into custom applications, automated reports, security policies, compliance workflows, and the daily muscle memory of tens of thousands of employees. A spreadsheet feeds a macro that feeds a database that feeds a report an executive reads every Monday, and every link in that chain assumes Microsoft. To pull the foundation out from under all of it is not one decision but ten thousand small ones, each carrying its own risk of something quietly breaking — which is precisely why the migration is forever postponed to a someday that never quite arrives.

The commercial consequence is a customer who renews a dependency rather than reconsiders a choice. When contract time comes, the enterprise is not soberly evaluating alternatives and weighing Microsoft against a rival on the merits. It is renewing something it cannot practically live without, and it knows it, and Microsoft knows it. That is what lets the company raise prices and add products with a steadiness that would be reckless in almost any other business — and it is why, of all the moats a software company can dig, being the thing that is simply too embedded to remove is the most comfortable position of all.