First things first. Although all responsibility for this post is mine, I want to offer my sincere appreciation to Michael Krigsman (@mkrigsman on Twitter) for his suggestions on how I could write a better, more useful and balanced post and to Dennis Moore (@dbmoore on Twitter) for his insights (with which I agree entirely) about those situations for which true SaaS is not the right application architecture/deployment method/business model. I also want to note that this post applies strictly to HRM software being offered now (so with all due respect for legacy applications which may have been the bees knees in their time), which is my major focus, and may not be applicable universally to all other enterprise business domains.
Larry Ellison and I disagree, and not just about the America’s Cup being run where he’ll get great views from his San Francisco house. It seems that Mr. Ellison is not impressed with multi-tenancy as a core design principle of true SaaS. Nor is he impressed with his major competitors.
In my view, this is classic Ellison selling what he’s got no matter what’s best for his customers. Virtualization is quite useful if you’re hosting single tenant software as well as for data center resource management, but it doesn’t give you the most important benefits of true SaaS, of SaaS InFullBloom. What it does do is to impact TCO in ways that will help Oracle’s profit margins. Anyone want more hardware (sorry, now they’re called appliances)? How about some more RDBMS licenses?
But how easy is it to handle inheritance properly across tenants, e.g. for the type of embedded intelligence that makes data entry-style self service into self sufficiency, without true SaaS? How about aggregating data across tenants (obviously with all the relevant permissions) to create benchmarks or shared pools of candidates or job descriptions? What about the real cost (to include managing the risk of errors and rework) of applying upgrades across a virtualized versus a true SaaS architecture? And can you really keep everyone on the same release when some of them are implemented on-premise, some are being hosted by the vendor single tenant, and some are being hosted by the vendor multi-tenant (assuming that Oracle can ever get there with Fusion?)?
It’s entirely possible that Salesforce.com’s architecture needs refreshing (and is no doubt getting it) because of when they started building. It’s also possible to build badly designed but multi-tenant software, and there are certainly situations (e.g. you’re responsible for HRM of our covert operations teams) when true SaaS just isn’t the right answer, even for new applications. Realistically, we’re going to have a mix of true SaaS and everything but true SaaS for many, many years to come — just as we’ve still got flat files and COBOL at the heart of many applications (we’ll mention no names of where that old COBOL pops up in now Oracle products). But it’s Mr. Ellison’s opinion of true SaaS that’s the crock, not multi-tenancy. True SaaS is the best model for contemporary HRM software in all but a few extreme instances, and I believe that Mr. Ellison knows that even as he’s saying otherwise. Did I mention that SAP, Workday, most of the TM vendors, NetSuite, Salesforce, most of the consumer Web sites and many more agree with me?
Mr. Ellison is a brilliant businessman when it comes to making money for his investors and for himself. Furthermore, I think that Oracle has a large part of the IT community by their collective tender parts. And with his investment in NetSuite, he’s playing both sides against the middle, because NetSuite is true SaaS. But Mr. Ellison and his firm are never going to win the contest for most customer-focused, at least not among HR leaders. And just wait until you’ve digested their Fusion HCM price list!
What’s really sad is that the Fusion HCM team has done a very good job within the limits imposed upon them by Mr. Ellison’s business strategy. I believe they’ve pushed their fresh HRM thinking and product design efforts as far as they possibly could within the allowed technology stack and while preserving sufficient backward compatibility with their installed base to retain those maintenance streams well into the future. But without true SaaS, without the ability to keep everyone on the exact same code base, without the many other benefits that only accrue to their customers when a vendor delivers well-done true SaaS, I think Mr. Ellison’s America’s Cup entry has a better chance of beating its competitors than does Fusion HCM when it’s compared, all in on a “rip and replace” basis, to the emerging “fleet” of true SaaS HCM products.
And yes, in the interests of full disclosure, Oracle is not a client (although I did do work for them during Joel Summer’s respected tenure as head of the EBS HCM product line), I’m not on Larry’s most favored analyst/influencer list, and I don’t expect to be invited to watch the America’s Cup as his guest. I would also note that this post was not written as a result of any Oracle briefing and, therefore, was not subject to Oracle’s review.