Ten years ago, Salesforce.com Inc. was founded by former Oracle executive Marc Benioff as a pioneer in delivering enterprise Software-as-a-Service. Their Customer Relationship Management (CRM) product was a huge success; after several years of development, its radically extended and enhanced foundation was exposed to developers as the Force.com Platform, allowing external developers to create CRM add-ons and also to build completely new applications that remain hosted on salesforce.com's multi-tenant infrastructure.

Launched in 2005, AppExchange is a directory of applications offered to the salesforce.com marketplace by third-party developers: users can subscribe to these offerings and integrate them into their individual organizations’ salesforce.com environments. There are over 800 applications available from more than 450 Independent Software Vendors (ISVs). 

We caught up with salesforce.com’s Director of Platform Research, Peter Coffee, in a conversation with Detecon’s Thorsten Claus to learn more about the experiences driving these decisions and where the company plans to go next in its relentless efforts to continue growing through a down economy.

DMR: In 2000 we had a lot Application Service Providers (ASPs). Then everyone moved to “On Demand”, with a lot of people having a lot of different opinions what that actually means. Now everyone is talking about “Cloud”. What is different from 2000?

Peter: The fundamental difference is that application service provisioning took the load of configuring and administering a copy of an application for a customer out of the customer’s hands and allowed the customer merely to use the function, but that workload didn’t go away – it was just assumed by the service provider. What that meant is that the complexity remained, the difficulties of taxing and updating software remained, the difficulties of coordinating a complex and brittle stack of software products remained. 

And that is the core distinction between the old ASP model that simply relocated complexity out of the customer’s basement and into the service provider’s basement: In the old model, because it was so disruptive to perform a major software upgrade, customers would tend to postpone those upgrades and sometimes even skip an entire release of a product as has happened in the desktop environment over the last 5 or 6 years.  This is bad for two reasons: One, it means customers aren’t getting the benefit of new technology because they postpone the acceptance of an upgrade; two, the vendor now has to support and provide security patches and other issues have to be addressed for many versions of their technologies simultaneously. The vendor’s efforts are diluted while the customers’ benefits are delayed.