Economics of Cloud Services

Vinnie Mirchandani at Deal Architect blog has a great post TCO - Total Cloud Ownership which addresses how architectural and business model choices of cloud vendors influence pricing. This topic is near and dear to us at Zoho, and I wanted to provide our perspective. I also want to contrast our approach with that of Salesforce. Before I begin, let me make it clear that I really respect what Salesforce has achieved in making software as a serivce a reality. They are pioneers and paved the way for services like Zoho. My post is not about moral good vs evil, but about different ways of approaching this business. 

To quote Deal Architect:

 While most SaaS/cloud vendors look ridiculously cheap in most TCO elements compared to their on-premise competitors, in the next wave (sooner than you think), they will be compared against each other.

So, last week in our conversation with Parker Harris of salesforce, the question came up – why would customers pay its storage costs compared to rates amazon and now Microsoft with Azure are bringing to market? Will it need to rethink its EMC infrastructure? From there the conversation moved to whether it needed the Oracle DBMS – whether it could afford to keep passing along those costs when vendors like Zoho and RightNow aggressively use open source components.

Cloud vendors have to take control of every aspect of their TCO. Like Google with its commoditized server chips. Like NetSuite imploring its reseller channel to deliver more services as software. Like Microsoft rethinking data center design and economics. Like salesforce going to market with smaller SIs like appirio, Model Metrics and others – not just the traditional SIs. Network costs. Systems management costs. Every aspect of the TCO.

That has exactly been our point of view at Zoho from the beginning. Our goal has always been to bring the efficiencies achieved by Yahoo & Google in architecting their infrastructure,  to business & productivity applications - as an aside, let us not forget that it was Yahoo which began using commodity infrastructure components and open source software well before there even was a Google. Have  you wondered why is it that Yahoo and Google (and now Facebook and many others) are able to serve hundreds of millions of users, with customized applications, with a lot, lot lower infrastructure costs than is typical in enterprise IT? When was the last time Yahoo went down?

At Zoho, we look at cloud economics in 3 dimensions: (a) The cost of developing applications  (b) The cost of insfrastructure to deliver the applications (c) The cost of marketing and sales, of which (b) and (c) are particularly important. Our goal is to deliver efficiencies in each of these dimensions, so that we can offer real value to customers. In terms of development, we reap systematic efficiencies by using common components and frameworks across Zoho services. In terms of infrastructure, we use commodity servers with substantial redundancy, and by using open source components wherever possible. In terms of markeing and sales costs, our business model is to spend less on these and pass on the savings to the customer.

These three factors, particularly (b) and (c) explain why we are able to price our services so affordably. Zoho CRM is the classic example: we charge $15 per user per month for the CRM ($12) + Mail ($3) package, while Salesforce prices their comparable edition at $65 per user per month. We believe they are fundamentally inefficient in all 3 dimensions, and that inefficiency becomes a tax on the customer.

I want to emphasize that we are profitable at our price point, there are no gimmicks here. Contrast that to what Salesforce chooses to do (screen-capture from their pricing page at http://www.salesforce.com/crm/editions-pricing.jsp).

That gap, or may be I should call it the Grand Canyon, between $9 per user per month and $65 per user per month is exactly what you have to resort to when you have a fundamentally inefficient business model that precludes you from dropping your price the honest way. They must really believe this can keep customers from defecting; I can tell them it is not working