Friday Rant: Don't be a SaaS Bigot
In software circles these days, it's almost held to be a pagan ritual to believe in and deploy traditional behind-the-firewall enterprise software. It seems everyone has not only drunk the SaaS koolaid -- they've bathed in hosted holy water as well. Even SAP, the king of highly configured uber business workflows that cost millions of dollars each once you're done paying off the consultants to make it work, recently announced it's getting SaaS religion. Many will say it's about time. But is it?
I've gotten a surprising number of questions from users and consultants since the New Year about companies considering both SaaS and non-SaaS approaches to tackling Spend Management problems. Many are concerned, for example, about the limits of multi-ERP integration in the P2P space. Others voice similar security concerns -- many of which are based on incorrect assumptions, yet they continue to linger -- to those that date back to the origins of on-demand in the early years. But more are interested in exploring the installed options because they realize that their environments are so different than the next guy that they worry SaaS approaches will fall short of what they need.
Now, this is not to say that installed applications will thrive across the Spend Management spectrum. For example, I highly doubt that we'll ever see more than 10% of e-sourcing deployments go anything but SaaS again. But I do reckon that in larger companies, there will continue to be many aspects of P2P that reside within the firewall -- even new deployments. I also believe that contract management -- especially contract management deployments tied around spend and process compliance that must connect with dozens of internal and external systems and data sources -- are not necessarily going to go the way of SaaS. Spend visibility is another question entirely, although I suspect that we'll see more and more hybrid approaches where companies outsource data enrichment, cleansing and rationalization on a weekly, monthly or quarterly basis but keep the analytics component within their four walls -- most likely built on-top of a BI or custom reporting platform.
So where is this leading us? I believe that while a SaaS backlash may be overstated, I do suspect that we'll see a number of installed fans quietly getting together in the shadows. Like smokers who feel shamed congregating outside office buildings and restaurants when they need to light up, these installed purists will face the smirks and grimaces of their colleagues. But maybe -- at least in some areas -- they'll be the ones getting the last laugh (or puff). At least within the confines of some of the more complicated and fortified Global 2000 walls.
- Jason Busch
Tweet
TweetBacks
So, let's stop reckoning and start thinking. First, let's accurately define "SaaS" as "browser-only client," and throw the rest of the marketing nonsense out the window.
Next, any business software that works well in the old CICS model (fill out a form -- send the form -- get another form back -- fill it out -- send it -- get another form back -- etc.) ought to be deployed with SaaS. The advantages of doing so are obvious.
Next, just because other apps -- analytics, office suite, photo editing, a long list -- can be force-fit into a browser-client model, doesn't mean that it's a good idea to do so. The result is a square peg in a round hole, with significant compromises. If all I'm interested in is red-eye elimination from the latest photo of my gap-toothed niece, I'd be happy with some B.S. web-based photo editor. But not if I want to do serious work.
It will be very interesting to see what happens in this economic downturn. Many of these solutions are now proven, and serious buyers may be taking a longer-view perspective. I expect that many of them might opt to buy down their long-term costs (e.g. buy hosted software upfront rather than lease for 7 years), or they may attempt to lock-in extremely favorable SaaS rates with 3 to 5 year agreements.
It is interesting to note that Salesforce.com is already seeing price pressures across the board. Many of their customers are simply saying "we pay how much per year?" and looking to slash recurring operating costs...
In the enterprise space there will be a handful of mega-deals that will come up in the next 18 months. Flexibility in pricing models and even software delivery modes could advantage vendors in the short term.
I agree with Jason's observations that eSourcing and Spend Analysis Data services will continue to be provisioned on a SaaS basis. It makes little to no sense to bring these offerings in-house. Some corner cases exist, but these are very rare.
eProcurement deployments have mostly been behind the firewall throughout the past decade. I believe that if one were to count the total deployments within companies, having $2B+ in revenues, the results will be very much tilted towards installed software. Remember, you've got a very large Ariba Buyer legacy base, as well as SAP EBP, Oracle iProcure, PeopleSoft, etc. out there. I don't see much movement towards eProcurement as SaaS within large companies. The current economic climate is not conducive for revisiting and replacing this particular module. It will be interesting to see if adoption of eProcurement in the mid-market, where SaaS delivery makes lots of sense, accelerates or decelerates within the next two years.
Contract Management will likely continue its 50/50 SaaS vs. Installed software trend. In addition to integration concerns companies have always been reluctant to "let go" of their contracts.
If you take a 10 year view, it is not unlikely that many of today's solutions will split into two primary components. One component that lives within an enterprise, and another which lives on the net and marshals the connectivity to suppliers, mashes external data sources and links to various web-services. Each vendor will need to determine how they want to evolve to capitalize on this opportunity. For some like SAP and Rearden, the decision should be relatively easy, the company's offerings and DNA is focused on just one offering type. Others, like Ariba and Emptoris play on both sides of the fence (in different ways, which today work effectively for both). It will be interesting to see if they choose to continue to do so. Only time will tell what is the "right" choice for each vendor.
As for the next areas of high-value solutions? There is little doubt that they will emerge in the cloud - as SaaS offerings, as new types of web-services, and as mash-up applications. There is lots of room here for both point solutions and for elegant SaaS platforms that smartly pull it all together.
A behind the firewall software platform that intelligently leverages these new web-services could be a huge winner. But for that to happen a visionary behind-the-firewall vendor would need to emerge. We've not seen one of those in over 16 years, and there is little to no movement in this direction from the current players...
- Ammiel Kamon
The money saved is in support and personnel. If a company is big enough to have their own support already, or too small to care much about real support, then owning your own equipment makes a lot of sense. How about those companies started by us technical people who just enjoy playing around with new servers?
Makes you think, huh?
Not that simple. It's how much it costs over the lifetime of the product. You buy a big in-house system, you're going to have it 5, 7, maybe even 10 years. To properly cost a SaaS offering, you have to look at it over the same time window.
On-Premise comes with a big up-front license fee, annual maintenance fee of probably 22% (or more), hardware (and upgrade) fees every 3 years, internal support costs every year (which go up on an annual basis), database/web-server/middleware license costs every year, initial implementation & integration costs, upgrade costs for major releases every few years (not covered in the maintenance agreement), and training (and retraining after every major upgrade) costs. Over five years, that's a hefty price tag.
SaaS is just a monthly subscription fee, up-front training fees, and, depending on the type of solution, possilby an integration fee to link it in to your in-house systems.
If the monthly fee is low, even multiplied by 60, 90, or 120, it will still be much less than the TCO of many traditional on-premise solution. However, if the monthly fee is high, and traditional vendors, desparate to make quarterly numbers, are willing to slash list prices, then we have a different scenario.
Coupa has a simple calculator on their site at http://www.coupa.com/tco/procurement-software/ which gives you an overview of the costs. I have a more in-depth and editable spreadsheet that I'll be including in an upcoming post.
Even more compelling though than the customer-side benefits and trade-offs, for software providers (especially ones without an on-premise installed base) it's a no-brainer to go SaaS - no provider going forward will willingly sign up to support a bazillion platform combinations across a half decade of old versions. Providing a complete end-user-focused solution and getting much faster feedback are more fulfilling and allow fast continuous evolution, even over the long-term.
Finally, I think that some of the "religion" is simply designed to try to convince some of the late adopters to let go of their "gotta control everything" on-premise preconceptions and be open to evaluating a newer but (almost always) better way to buy and deliver software, even for large enterprises.
Best regards - John Martin, IQNavigator CTO, www.BuildingSaaS.com