Moving Accounting into the Clouds

My personal cloud computing "buzz-o-meter" hit an all time high this week.

The Wall Street Journal had a front page story on cloud computing - which started with Larry Ellison calling it "Gibberish" last year and ended with him announcing his new "cloud-computing ready" products on Oracle's earnings call last week.

Then I picked up Business Week which ran not one but two separate cloud computing stories - one encouraging IBM to acquire Sun to get more into cloud computing, and the other trying to educate consumers about being smart in using the cloud. NPR got into the act too, with another consumer focused story.

I've been having lots of meetings with thought leaders from the accounting profession lately - in one last week I heard that Intacct should start talking about "cloud computing" instead of SaaS or on-demand or Internet and in the other that the best way to describe Intacct was to call it "Cloud Accounting."

Then yesterday I heard through my network that our new president wants to learn a lot more more about cloud computing too.

So what does it all mean?

Just about everyone in the technology profession (and apparently now even Larry Ellison) agrees that Cloud Computing is a disruptive new shift - even a bigger deal than the last major architectural transition from mainframe to client server computing. But our industry is full of buzzwords that come and go, so it's sometimes hard to sort out real vs hype.

My week above says that cloud computing has moved from a Silicon Valley buzzword to becoming something that has capturing the attention of mainstream consumers, government and business. I think the major factor driving this wide level of interest is that Cloud Computing is pervasive across our experience as consumers as well as in the business world.

And when multiple thought leaders from the CPA world, one of the most conservative professions of all, start talking about moving CPA firms to cloud computing and Intacct as "Cloud Accounting" - that says to me that we are really on to something.

New Microsoft Study - SaaS Adoption by SMB Grows to 86%

A new Microsoft Study, released today, called the 2009 SMB Insight Report, says, among other things, that SaaS adoption by small and mid-sized businesses will grow to 86% in 2009.

The key trends that Microsoft identifies for Small and Midsized Businesses for 2009 all bode well for SaaS:

1. Lower Budgets Mean Less Has to Be More.
2. Strategic IT Investments Bolster the Bottom Line
3. Service and Responsiveness Are Key to Customer Retention.
4. Smaller Companies Will Increasingly Adopt Cloud Computing.
5. Better Together: SMBs Rely on Trusted Partners.

Somewhat strangely, the highest interest reported by the study for a Microsoft SaaS offering was for Microsoft Silverlight - their competitor to Adobe flash - which I have a hard time thinking of as a SaaS or Cloud Computing kind of thing at all.

So while there is nothing earthshattering here, just more good validation that SaaS is become more and more widely adopted and why, one interesting thing that may not be obvious is that Microsoft is now starting to use the terms "cloud computing" and SaaS - for those of you that don't follow this closely, Microsoft have been pushing the term "Software + Services" as their alternative to SaaS, the idea being you combine traditional desktop software with Internet services. Perhaps now that they are seeing 86% of businesses wanting to adopt SaaS, they've decided to get a little bit SaaSy too.

Cloud computing is the next generation of SaaS ?

Earlier this week I found an interest article on a blog written by Bruce Richardson of AMR research. Bruce is musing about whether the combination of Compiere, an open source ERP application running on Amazon EC2 could be the next disruptive thing in ERP. The catalyst was a discussion with and an e-mail exchange he had with the Don Klaiss, ex SVP of Oracle Applications and now the CEO of Compiere, who told Bruce:

Cloud computing is the next-generation of software as a service, in which a complete software environment is licensed as a subscription from a software vendor and low-cost, secure, and dependable IT hardware infrastructure is ‘rented’ from a utility-computing provider on demand.

This new definition of cloud computing made me raise at least one eyebrow. Running your applications on the Amazon elastic compute cloud is a bit like virtualization - and I agree with Don that renting hardware in the cloud can be cheaper than installing hardware on premises. I also think it's particularly attractive for prototyping and for bursty applications, but that's for another post. But running applications on virtual hardware in the cloud is just a fraction of what SaaS is all about - and to claim it's the next generation of SaaS just doesn't pass the sniff test.

With SaaS, the vendor takes on the costs and risks of purchasing, maintaining and operating all of the infrastructure required to run the applications. This includes things like Oracle licenses, SOA platforms, security software, and a lot more. The SaaS vendor also operates the system, handling security, tuning, backup, upgrades, maintenance, disaster recovery, etc. SaaS vendors also spend lots of money and resources on operating procedures, SLAs and SAS 70 audits to guarantee things like like uptime, performance, security and quality of service.

There is also a lot of magic in multi-tenancy at the application level - with major associated benefits and economies in scaling and operations. Old single-tenant applications running in a virtual machine in EC2 by definition don't get this - above the hardware level it's still the old single instance, single tenancy, single maintenance model.

If you are running your applications on Amazon EC2, by definition you are doing all of this stuff yourself. Sure you get virtualized hardware and that's good, but buying and maintaining infrastructure is up to you, as is installing, maintaining and operating your applications. There are no SLAs, that's also up to you, and you're also on the hook for backup, disaster recovery, performance tuning, etc.

The way the SaaS business model works is that the vendors are taking all the money they save on sharing multi-tenant infrastructure and operations across thousands of clients and operations and returning it to their customers. SaaS vendors dis-intermediate both hardware and software infrastructure vendors, and are far more efficient at operations than nearly any individual customer could be. All of this savings goes right back into the pockets of SaaS customers. In the model that Don describes, you only achieve these efficiencies at the virtualized hardware level - which is only around 5% of the overall TCO for running a business application.

Even in the old ASP model, the attraction was that your vendor remotely operated your single instance, single tenant applications for you so you could get both hardware and operational scalability.

A TCO analysis would show that for business applications, all of the multi-tenancy and sharing you get with SaaS ends up being far less expensive than the costs of individual companies each running their own applications themselves on EC2. SaaS vendors get scale across everything - sharing infrastucture and operations costs, not just hardware costs, across large numbers of customers. In other words, you've got to consider the other 95% of costs that are not just related to running virtualized hardware.

So don't get me wrong - I think Amazon EC2 is great stuff - But to call putting a single tenant application on EC2 the next generation of SaaS - well I don't think so.