This afternoon at Structure 08 an interesting discussion was had about the birth, growth, trials and tribulations of Salesforce.com. Om Malik from GigaOm was joined by Michael Copeland from Fortune Magazine and Parker Harris from Salesforce.com. While the ‘fireside chat’ was titled The Endgame for Boxed Software?, the focus was on what lessons can be learned from the venerable CRM vendor.
According to Harris, in the consumer world, everyone must think not only on one level, but on several. Salesforce.com thought about software, scale and the Internet; questioning what it would take to build this ‘thing’ they were thinking about.
“When we met Marc [Benioff], he had a vision and had written a two page e-mail detailing that we need to make this as easy as buying a book on Amazon,’ Harris says. At the time, there was Amazon, Google and E-bay.
In early 1999, they built a prototype. Stealing tabs from Amazon and buttons from Yahoo!, they built a site.
Om: What has the transition been from then and now and why?
Harris: As a technology, we wanted to build a platform first. One that is beautiful from a technology perspective. But you risk losing touch with why you’re doing it. Originally it was going to be for sales people and there were various abstract layers; we quickly realized the need to unify them. We continued adding layers in response to our customers; columns, end user interface. At that point we realized it could be a platform; we didn’t consciously do that.
When we started, a lot of people said they didn’t want to trust their information to anyone; data, privacy, these were all big things, and people were concerned with trusting their customer records and leads with another company. We made the huge step, and now people trust the information services.
It’s still very early days in platforms–are you going to use someone else’s platform (app engine from Google, Amazon) or use your own?
It’s all about the database. Databases have evolved. Today you don’t really go into business to write it, you find one. I see platforms heading in the same direction.
We sell to business. We do a lot of work on compliance and security issues to meet the needs of enterprise. For business we are very appropriate; for websites, go to Google. Amazon gives you CPU time or storage. We too have a lot of technology that is pre-built that you can leverage if you want; analytics, search, 24.7 web services, API, it’s all there.
Michael: What about compliance and security and is anything Salesforce.com can do to get apps to go viral.
Harris: We want to have a directory where people can go find pages quickly. Free pricing models are one way to drive a viral model. We are looking at ways how to work with Facebook. We are looking to involve developers more.
Om: Infrastructure; what are the challenges
Harris: About two years ago we did go through some serious issues and part of it was due to eBay. They could give us more power but no space, or space and no power – so we moved everything to new data centers.
We made two big mistakes: we changed everything all at once, which is a big risk for any business, and moved onto large scale systems. Pushing scale in a vertical sense created a complexity that was just too hard. Then we hired the guy from eBay to make sure it never happened again.
There is no blueprint on building a scalable service; you never replace it with V2 so you are forever changing it. There are best practices that have evolved, and people should follow them if they want to do this.
Early on we thought of tech innovation as software and infrastructure, and that innovation can occur on one or the other. To truly innovate you need to think of it as one thing and change all areas.