CODA2Go and the economics of PaaS
I had the opportunity to talk money last week with Jeremy Roche, CEO of CODA, the venerable UK-based business software vendor that has become the poster child for Salesforce.com's platform ambitions after the release at DreamForce Europe of its new on-demand financials application, built and delivered entirely on Force.com. Why does a company with a 30-year history of writing finance applications and 2,400 customers — including well-known brands such as Ikea, Avis, Unilever and HSBC — entrust its on-demand future to a new, untested platform? [Disclosure: Salesforce.com is a recent client].
That time saved translates into millions saved on development costs. "We've saved at least two years in elapsed time and at least 25 man-years of development time," he said. "That figure could be as high as 50," he added, explaining that it's difficult to quantify the exact figure with no direct experience of what it takes to develop an on-demand infrastructure. Either way, the savings are huge. At an average UK developer salary of £50k ($100k) per year, Coda has saved between $2.5 million and $5 million on the development cost.
I'd assumed that the quid pro quo for those savings would be a higher monthly subscription cost — after all, Salesforce.com presumably can charge a hefty premium for all that added value. But Roche surprised me. "It is certainly cost-effective," he said, calling it "comparable" with the cost of servers and bandwidth if Coda were doing it in-house. Later that day, I had a chance to verify what Roche told me with Salesforce.com's CEO Marc Benioff. CODA2Go is priced at $125 per user per month. Clearly Coda isn't paying Salesforce's list price for Force.com of $50 per user per month, but to be competitive with the cost of goods Salesforce achieves for its own applications, Coda would have to be paying around $15-$18, I suggested. Benioff's answer didn't confirm or deny, but I was left with the distinct impression that PaaS partners get the steep discount that Roche had implied:
"We will have general pricing that we quote in a press release but every customer needs customized pricing to deliver their product," Benioff told me. "We've put together a price that we believe is cost-effective for the vendor and [at which] we can be profitable."
This of course is one of the most crucial calculations for any platform vendor, and even more so with PaaS. Pitch the price too high, and any applications built on the platform won't be cost-effective. Pitch it too low, and the platform loses money. PaaS vendors have the extra complication of having to predict variable costs that could rise as easily as fall, for example if unforeseen scalability, security or reliability issues come up that need to be resolved.
There's also the cost of training up partners. Coda has been getting a brain-dump of Salesforce.com's best practices in marketing, sales and support, including visits to the vendor's Dublin, Ireland support center to see how it executes on customer service, as well as implementing the 30-day trial sales methodology. Coda's developers have adopted Salesforce.com's scrum agile development model, aligning with the vendor's own internal two-weekly scrum cycles so that they can synchronize their own development schedules with the release of new functionality from Salesforce.com. Coda can also piggyback on separate developments such as the recent release of Salesforce for Google Apps, which Coda has already taken advantage of to integrate CODA2Go with Google spreadsheets. "We're benefitting from these other investments Salesforce is making in the infrastructure. It just becomes available to us," said Roche.
The on-demand sales and support model is not so different from how Coda already works with its conventional on-premise products, Roche told me, except for the greater emphasis on Web-based selling at the low end. The primary difference, he said, comes in the consulting model, where Coda will be using Salesforce.com partners such as Appirio and Model Metrics rather than its own in-house consultants. One of the advantages of this approach is that partners will typically implement CODA as an add-on to existing Salesforce.com implementations. "We'll appear to partners as an extension of Salesforce.com," said Roche.
Where the gamble will really pay off for Coda is if its application takes off with the Salesforce customer base. "One of the key benefits to us is Salesforce has 40,000 customers," said Roche, "some of which want accounting that's seamlessly integrated with Salesforce.com — on the same account and the same server." That's a unique for Coda until other vendors follow in its pioneering footsteps. Although other vendors such as Intacct have longer track records as Salesforce.com partners, Roche believes the tightness of the integration that comes from running on Force.com will be a killer feature. "I defy anyone else to be able to do that."
If Coda does get growth — especially in North America, which currently accounts for just 10 percent of its customer base — then it will consider the experiment a success. But it will still have to watch the economics carefully, weighing up the costs of sales and marketing, support and its Force.com licenses against its revenues. Equally, Salesforce.com will be keeping a close eye on whether it too can profit from the arrangement after taking into account all its infrastructure costs. The economics of PaaS are not yet proven, but one thing I feel sure of is that Coda wouldn't be making this move into multitenant SaaS if the Force.com option were not there. "Coda's an application company," Roche told me. "I don't want to do things like set up data centers." Nor is that expected to change as Coda gets more comfortable with the on-demand model. "It is not part of the strategy for this product to come off the Salesforce platform," he said.