Coda: How To Build A Cloud Company

Software vendors are coming under increasing pressure to offer customers software hosted in the cloud as an alternative to the software license agreements that have characterized the software industry for more than twenty years.

Once confined to the ghetto of small-bore applications that help address a simple problem, mostly for small and mid-sized companies that can't afford high-end products sold by the likes of Oracle, SAP and Microsoft, software-as-a-service (SaaS) offerings have become ubiquitous across industries and among customers of all sizes; meanwhile, Salesforce.com, the most successful SaaS company by far, has surpassed $1 billion in revenues and has created a platform on which other companies build complementary applications, thus making SaaS even more attractive to customers while creating a new revenue stream from the royalties it earns from partners.

But you can look no further than the troubles afflicting SAP to see that switching from one business model to another is easier said than done. When they weren't busy denigrating SaaS as unreliable and unsophisticated, software vendors have been trying to introduce SaaS versions of their own products for several years, with varying degrees of success -- or failure. Of the big three, only SAP has gone back to the drawing board and re-engineered its legacy application suite to make it fully compatible with the Web. But despite having officially launched Business ByDesign over a year ago to approximately one hundred pilot customers, SAP still can't produce a reference customer and has not released the product for general availability.

Coda, which sells financial management software, has been in business since 1979, and with some three thousand customers for its on-premise application, seems like an unlikely convert to the SaaS delivery model. It has an experienced sales staff bringing in a steady stream of business, but has nowhere like the resources of an SAP, which can afford to create a parallel technology. But Jeremy Roche, the company's CEO, explained over breakfast that creating a SaaS offering for small and medium-sized businesses "can open a whole new market for us."

Roche admitted that the business will take several years to become profitable; customers will pay $125 per user per month for the new service, Coda To Go. Most Coda customers only need five or six users for this type of application, so Coda will generate about $8,000 per year for a typical customer, in contrast to the traditional software model which generates upwards of $100,000 per year per customer. Roche also recognized that some customers who might have paid for the legacy product are likely to defect to the cheaper service model. "There will be an element of cannibalization, but I'd rather cannibalize my own customer base than let someone else do it," he told me. Moreover, "multiply that [$8,000] by several thousand customers and you start to build a reasonable business."

Another challenge with keeping both on-premise and SaaS versions of the same product is that a salesperson used to earning commissions on deals worth several hundred thousand dollars can't be expected to sell deals worth tens of thousands. That's one reason Coda To Go is managed separately from the traditional software business, and has a separate sales staff. Roche said sales people for the new service have different kinds of incentives, such as getting customers to commit for 12 months or more.

But the biggest challenge for Coda was creating a new technology foundation for the service; vendors that have tried to reuse the architecture underlying their software products for SaaS have universally failed to deliver satisfactory user experiences.

That's where Force.com, the platform created by Salesforce.com, came in. Although it meant using a proprietary language and becoming a sort of vassal to Salesforce.com, Roche came to see building the application on Force.com as an opportunity that couldn't be missed. "Hitching yourself to the most successful business in that space didn't seem like a bad strategy to me," he said. Moreover, said Roche, the platform has proven its reliability in the years that Salesforce has been in business. "What makes us different is we've inherited ten years of Force.com platform, and lets us ramp well ahead of what would have been if we'd built our own platform," he said.