CompanyGraph
Salesforce Inc
CRM · NYSE Arca · United States
Runs a shared cloud platform where thousands of companies manage their sales, service, and marketing data.
Salesforce runs a shared CRM platform where every change a customer makes — new data fields, workflow rules, screen layouts — is saved as a configuration record rather than written as standalone code, and the platform reconstructs each customer's environment from those records every time a request comes in. Because thousands of customers share the same underlying computers and databases, adding a new customer costs almost nothing extra, but over time each customer's configuration grows more intricate, and the platform has to do more work on every request to parse those layers — so the same architecture that makes growth cheap also makes the heaviest customers expensive to serve. A customer who wants to leave faces the problem that their Apex code, custom objects, and automated workflows are expressed in force.com's proprietary language, so moving to a competitor means rewriting years of work from scratch rather than copying files across. The structure that makes switching so hard also creates the central tension: if regulators force Salesforce to split its shared infrastructure into separate regional systems — one for Europe to satisfy GDPR, another for U.S. government customers to meet FedRAMP rules — each isolated instance has to parse tenant configurations independently, multiplying the cost of running the platform without adding a single new customer.
How does this company make money?
Most revenue comes from subscriptions — companies pay a recurring fee based on how many employees use the platform and which features those employees can access. On top of that, customers who send very high volumes of API requests pay additional per-transaction fees. Companies that use Tableau, MuleSoft, or Slack pay separate licensing fees for those products, which Salesforce acquired and now sells alongside its core platform.
What makes this company hard to replace?
A customer's Apex code is typically wired into their ERP systems through MuleSoft connectors — replacing those connections means rebuilding integrations that took years to configure. Executives rely on Tableau dashboards built on top of Salesforce data, and those dashboards would stop working on a different platform. Slack workspaces at many companies are set up to automatically receive CRM alerts and notifications, and those automated workflows are tied specifically to Salesforce's system.
What limits this company?
The shared parser that reads each customer's configuration at the moment of every request is the ceiling. Large enterprise customers build up years of interlinked Apex code, custom objects, and workflow rules, and each request forces the parser to trace through all of those connections. When several large customers are active at the same time, every customer on the shared system can slow down — not just the ones causing the load.
What does this company depend on?
Salesforce cannot operate without Amazon Web Services, which hosts the primary infrastructure. It relies on Oracle database licenses for the underlying data layer that stores all customer metadata. Akamai CDN delivers content to users around the world. The Tableau analytics engine, which Salesforce now owns, is embedded across the platform. And reaching customers on phones requires Apple App Store and Google Play Store to distribute the mobile CRM app.
Who depends on this company?
Enterprise sales teams rely on Sales Cloud to see their pipeline and forecast revenue — when it goes down, that visibility disappears. Contact centers use Service Cloud to route customer service cases, and an outage causes those cases to stop reaching the right agents. Marketing teams run automated campaigns through Marketing Cloud, and a connectivity failure stops those campaigns mid-execution.
How does this company scale?
Adding a new customer is cheap — they share the same computers, databases, and software that existing customers already use, so no new deployment is needed. What does not scale easily is the engineering team needed to maintain the metadata-driven multi-tenant architecture itself. Understanding the force.com platform at the level required to keep it running takes years of specialized experience, and that kind of expertise cannot be hired or trained quickly.
What external forces can significantly affect this company?
GDPR and U.S. state privacy laws require data to stay within specific borders, which conflicts directly with the efficiency of running one global shared system. FedRAMP compliance rules impose separate architectural requirements for government customers. U.S.-China technology export restrictions limit how data can move across borders for customers operating in both countries. Each of these forces pushes toward fragmentation, which is exactly what the multi-tenant model is designed to avoid.
Where is this company structurally vulnerable?
If regulators — through a GDPR enforcement ruling or a FedRAMP architectural requirement — forced Salesforce to split its shared infrastructure into separate systems for each jurisdiction, the economics would fall apart. Right now, one shared system serves everyone efficiently. Broken into isolated per-jurisdiction copies, each copy would have to independently parse every tenant's metadata graph, multiplying the cost of running the platform without bringing in more revenue.