Is the Real Secret Sauce of SaaS Discipline not Multi-tenant Architecture?

There is a prevailing view that I see in the press and blogging world that business application software is moving inexorably to the SaaS model.  It is true that you would be hard pressed to find a vendor who started in the HCM software market in the last five years that did not utilize the SaaS model.  However, does that mean that vendors that started with a single tenant, perpetual license, on-premise installed solution are tomorrow's legacy vendors?  No, not necessarily.  The implication has been that if those vendors do not rearchitect their solutions (or create next generation, multi-tenant solutions) they are going to be displaced.  I am not going to address this perception directly.  Instead, I want to try to make the case that a vendor can offer a hosted solution that delivers many of the same benefits if the vendor has the right discipline.

Let's take some of the key perceived benefits from SaaS and see how a hosted solution might meet those needs:

• Lower up front costs - There is nothing that would prevent a vendor from offering an on-premise or hosted single tenant solution with a subscription license (some vendors already do it)
• Easier upgrades - There are things that vendors with single tenant solutions can do to make upgrades easier for customers.  First, they can offer remote application management services to customers (some are doing this already) where they can apply patches, updates, and new releases for the customer at their site.  They can also host the solution for the customer and do those activities as well.  This does not deal with the difficulty of the upgrade (if there are many customizations or changes in the data model with the new release), but it can help.  There is no reason that vendors of single tenant solutions cannot work with customers the same way (at least moving forward) as SaaS vendors do - just do not allow customizations that impact the ability to upgrade.  If they host the solution, they can enforce that discipline
• More frequent new functionality delivered - There is no reason that traditional vendors cannot do this.  SAP has demonstrated that it can do it with its switch framework and enhancement packs (they started with enhancement packs every 3-4 months, but have scaled back to every ~9 months because SAP felt customers were having trouble adopting the new capabilities that quickly).  It becomes more of a question of resetting customer expectations and patterns.  SaaS sets this expectation and enforces the discipline, but that does not mean the discipline cannot be applied with a hosted solution.

I think you can see where I am going with this argument.  The effectiveness of SaaS is as much about the discipline it enforces in implementation and upgrade practices as it is about technical architecture. 

Now, I also know I am going to hear about the ability for vendors to cost effectively deliver this - SaaS has inherent advantages right?  There is truth there.  However, there are technologies that can help lower hosting costs.  Virtualization technology can make it easier to manage multiple instances (and keep them in sync in terms of release levels).  Grids and Blades can make the hardware environment less expensive and more scalable.  Cloud computing also can help make the compute services used more elastic (think development and test environments that do not use production data).

Does that mean a hosted, single tenant solution can match SaaS on economics?  Maybe not, but I am not sure it has to, it just needs to be in the neighborhood.  The real challenge is the discipline.  Can the vendor resist the temptation to allow customization and increase the tempo of new functionality delivered?

So, what am I missing?  I did not try to detail every benefit of SaaS, but is there a flaw in the logic?  Let me also mention that I am a big fan of SaaS as a delivery model (before the flames start to be thrown).  I just do not think it has to be the only game in town.  As far as I know, no single tenant solution vendor has ever tried to do hosted offering where they enforced these disciplines.  If you know one that has, please comment.

Comments

[working girl]

Great reminder that SaaS just means Software as a Service, with no mention of any particular architecture. I think marketing might be a challenge, though - for example, if your big claim to fame is your easy customizability, 'No more customization but we'll be happy to host you for just a bit more money than the multitenant SaaS guys,' sounds a bit weak. But I agree that discipline is the secret in most successful sauces.

[Jim]

Good points. There is certainly a marketing challenge. Perhaps, the vendor would set up a separate business unit (even brand). It would not necessarily be for existing customers, just new customers.

[Peter Pada]

Hi Jim,

interesting argument. I am always a fan of supporting an underdog if the fame of the dominant model seems to me more due to fashion trends than to real delivered value.

Furthering on your arguments, the following remarks come to my mind.

1. The virtualization of hosting hardware and software allows to recreate a hosting environment basically anywhere. So theoretically, if SaaS production environments are standard enough, and simple enough, the possibility exist to host the application anywhere, including in a hosting cloud belonging to the client. Multi-hosting or flex-hosting of a unique code is a trend that I would not be surprised to see soon in the SaaS industry.

In the end, the only thing that SaaS vendors are really selling remains their software. The brain juice of their developers and product managers, for the achievement of client benefits. So yes the wrapping is attractive, but of course, nobody will buy a bad software, just because it's a SaaS.

There already exist SaaS offering that are based on a single source code and hosted where the client sees fit

2. An added value of the SaaS model is that by construction the company cannot survive by enforcing the discipline you are referring to. This very strong incentive to do what seems at first glance to be contrary to client satisfaction is driving a number of company processes. A single tenant software editor doesn't have this incentive, and furthermore, has the contrary incentive to deliver client satisfaction by delivering customized code.

So I would go a bit further, and say that as far as SaaS is concerned, the discipline you are talking about is so difficult to implement, and comes with so many commitments in return, that companies can be constructed to deliver this discipline/commitment or not at all.

3. If a single tenant software is hosted by the vendor or by the client, if code unicity is protected by company discipline and if the pricing model is a subscription, then actually it is a SaaS. It is software delivered as a service.

What do you think?

[Jim]

Peter,

Thanks for the comments (and sorry for the long approval time - I just missed it). I think that most people think of SaaS as multi-tenant software. The point of the post was to show that multi-tenant software was not the only way to achieve many of the same benefits. What multi-tenant software has done is enforce a discipline around customization and upgrades that was not there previously and a lot of the benefits (though not all) can be attributed to this. However, it is not the only way to enforce that discipline.

[John Martin]

Jim: you're correct that upgrade discipline, along with building for/operating a single platform, are the key benefit drivers of SaaS, rather than multi-tenant per se (see http://buildingsaas.typepad.com/blog/2006/08/salesforcecom_i.html for more details). The real key is upgrade timing - requiring clients to upgrade allows supporting only one version and having only one upgrade path.

However, it's not as if on-premises providers simply lack the "discipline" to become SaaS -- their development progress is held back by the albatross of supporting years of old releases and porting/testing/supporting across a myriad of platforms. Beyond development, other significant hurdles have to be addressed such as sales compensation, revenue recognition, and acquiring the skills to successfully host sofwtare to world-class SLAs.

If it were as simple as becoming more disciplined, then we'd see lots of on-premise software vendors successfully moving to SaaS; instead, successful transitions are a rarity and every SaaS leader is a new player.