# Veeva *The rails you don't own* --- In 2007 a former Salesforce engineer named Peter Gassner started a software company for one industry: pharma. His first product, a CRM for drug-company sales reps, he built on Salesforce's own platform. It was the fast way to market, renting Salesforce's infrastructure instead of building his own, and it worked. Veeva became the system nearly every large drugmaker ran its sales force on, a company now worth around thirty billion dollars. But it was renting the ground it stood on. --- So in 2011 Veeva began building its own platform, Vault, from scratch. Not for CRM at first, but for the documents and data that run a drug company's R&D and quality functions: regulatory submissions, clinical trials, manufacturing quality records. Vault was proprietary, owned end to end, and over a decade it became the larger half of the company, the half no landlord could touch. Then, in December 2022, Veeva did the hard thing. It announced it would not renew its agreement with Salesforce, and would rebuild its flagship CRM on Vault, its own platform. The Salesforce deal ends in September 2025. The new Vault CRM went generally available in 2024, and Veeva has told customers the old version has a hard end-of-life in 2030, with no extension. --- The cost of that is hidden inside Veeva's own moat. Its systems are validated, regulated and mission-critical: a drug company cannot casually swap them out, because re-validating a system a regulator has signed off on is slow, expensive and risky. It is the same existential stickiness that keeps [[Jack Henry]]'s banks from ever switching their core. That switching cost is why Veeva has long kept net revenue retention above 120%. And now Veeva has to ask those same customers to switch, to migrate the CRM their sales force lives in onto a new platform, on Veeva's timetable. For once the moat runs the wrong way. It is worse than that: the landlord is now a competitor. Salesforce, no longer a partner, has launched its own life-sciences product and is courting Veeva's customers at exactly the moment they are being asked to re-platform. A locked-in customer is, briefly, in play, and the migration window is precisely when. --- Veeva is doing it anyway, and the logic is one every operator on someone else's platform should sit with. A platform you build on is leverage someone else holds over you: your costs, your roadmap, your margins, your independence, all partly in their hands. Renting the ground is usually the right call early, when speed matters more than control. But once the business is valuable enough, the rent becomes a strategic risk, and the only cure is to own the ground, even when laying your own track means asking your customers to move house. Veeva could afford the cure only because it spent a decade quietly building Vault before it needed it. The escape was possible because the alternative already existed. Most companies that depend on someone else's rails never build that option, and learn the cost of not owning the ground at the worst possible moment, when the landlord changes the terms. ---