Step 2 of 5 · 1999–2005

A monthly subscription Siebel could not match without cutting into its licenses

The monthly price was set before launch, in 1999, and it was the part of the model Siebel could not copy. Salesforce charged $50 per user per month and claimed 90% of the features of Siebel's flagship product at a fraction of the price SF11. A subscription replaced a large upfront license with a monthly bill, and it fit the buyers Siebel was not built to reach: from 1999 to 2004 Salesforce sold mostly to small companies and divisions of larger ones, averaging about 24 seats per customer SFS-1.

Siebel saw the threat and could not answer it without hurting itself. Its 1999 response, Sales.com, was a free website with some of the same features, but Forbes reported that Siebel had to limit it or cut into more than $600 million a year of license revenue SF11. Bruce Cleveland, of Siebel's founding executive team, later wrote that Siebel was built to sell to large companies through a large, expensive direct sales team SFS-1. A cheap monthly service would have competed with the licenses that paid for that team.

Marc Benioff knew this business from the inside. He had spent 13 years as a marketing executive at Oracle and was an original investor in Siebel Systems SF11. In 2005 he described the license model he was attacking: "When I was at Oracle, we watched Computer Associates buy all those mainframe software companies and milk them for their license revenue" SF12.

By January 2005 Salesforce had about 13,900 customers and 227,000 paying subscriptions SF1. Its revenue was still $176 million to Siebel's roughly $1.4 billion, but Siebel's sales had stalled since 2002, and in September 2005 Oracle agreed to buy Siebel for $5.85 billion SFS-1 SF12.

Rivals Siebel CRM OnDemand cost $70 per user per month, above Salesforce's price, and Benioff said in 2003 that Salesforce met Siebel in deals all the time and beat it SFS-2. A 2001 recession that hit Siebel's telecom customers hardest also held back its license sales SFS-1.

Counter-positioningSubscription PricingPenetration PricingFounder Domain Expertise

Step 5 of 5 · 2011–26

Switching costs, reinforced by a trained Salesforce workforce

The single system of step 1 is where each customer's work piles up, and that makes Salesforce hard to leave. Schneider Electric began with Sales Cloud around 2011; by 2016 all its call centers ran on Service Cloud and its partner program on Community Cloud SF4. Fields, workflows and AppExchange apps from step 3 sit on that system, and leaving means rebuilding them elsewhere. The trained staff of step 4 add weight: a company that switches must retrain or replace the people who run its CRM.

Veeva shows both the cost and its limit. It built its life-sciences CRM on Salesforce, then built Vault CRM on its own platform; in October 2025 it reported more than 25 customer migrations done and about 300 to go, with its older CRM reaching end of life in September 2030 SF6. A specialist partner that owns its customers can fund an exit, and the move still takes years.

Growing inside large accounts cost money. Salesforce warned in 2005 that larger customers brought longer sales cycles SF1. It widened its scope by buying Slack in 2021 SF13, and in January 2023 Benioff said hiring had run ahead of the business and cut about 10% of staff SF8. In fiscal 2026 Salesforce reported $41.5 billion of revenue and a 20.1% GAAP operating margin SF9, built on the monthly service Siebel could not afford to offer in 1999.

Rivals Siebel's own configured customers stayed too: Oracle still ships Siebel CRM releases, version 26.3 in 2026 SFS-9. That base kept Siebel alive inside Oracle while Salesforce led the CRM market SFS-8.

Switching costsExpertise network effectsLand and ExpandMulti-Product