Saturday, September 01, 2007

SaaS lovers want it both ways - a short rant

A lot of SaaS proponents take great pride in distinguishing the new Software as a Service era from the hosted applications or ASP model - and insist that there is only one right way of doing SaaS. This is in direct contrast to what I am seeing a lot of ISVs do in real life - which is, adopt a range of delivery model options to fit the customers need and economics of their particular business. But, every time an established software company talks about its success in SaaS by pointing to the wide range of options in the SaaS business, the SaaS purists go up in arms. This would be perfectly fine if their definition of SaaS was consistent when it came to only including shared everything multi-tenant services with its much hyped poster child Salesforce.

The problem is that when they tout the growth and size of SaaS and its wide adoption, they refer to billions of dollars in revenue which includes all variations of SaaS.

So here is my point - either go with a "purist" SaaS definition and accept SaaS as a relatively small niche market today with limited adoption or expand your definition to include different SaaS models. Don't mix and match.

Related posts:


Mukund makes an excellent point in the comments that multi-tenant model of SaaS delivery is a preferred architecture - and I agree that it is, for many applications. I don't argue in my rant that multi-tenancy is unimportant or irrelevant- in fact, quite the contrary, I am of the opinion that it is the most suitable model for many ISVs and has several architectural and business beneefits. My argument is with the attempt to restrict the definition of SaaS to only one model and yet continue to conveniently include other models when it helps SaaS puritanism proponents make their point.

Update 2:

The post has elicited a series of responses. Phil Wainewright asks "So who exactly is trying to have it both ways?" in his post on ZDNet and argues against my viewpoint - even though I think that he and I agree on a lot of things as I mention in my comment on his blog including the importance of multi-tenancy. Bob Warfield doesn't take sides but points to what he considers the more important issue of economics (than architecture) . Sinclair Schuller follows up with his post asking Are there REALLY multiple strategies for SaaS ISVs?

This is turning out to be pretty interesting conversation. What do you think about the so called purist vs. realist SaaS debate? Do you think that there is only one true blue SaaS architecture that qualifes as SaaS?


Mukund Mohan said...

Good point, but there's no easy way to move from ASP to SaaS (architecturally). A lot easier if you have architected a solution to be SaaS to move to an ASP if needed.

Anshu Sharma said...

I think multi-tenant SaaS model of delivery has several advantages that make it the preferred model for several applications.


vinnie said...

I am not a SaaS purtist by any means...what I like about SaaS it is pricing at $ 75 a user a month versus the 20 to 30X traditional on-premise ERP license, maitnenance, impl and support have cost thousands of companies (when you amortize TCO over a 5-8 year basis). Don't say you are cheaper than SAP and leave it at that. You may be marginally better in some areas, higher in others but all on-premise ERP, CRM etc is way too pricey. Nothing to do with elegance of architecture, just good solid economics...

Anshu Sharma said...

Agree with you that SaaS model has economic advantages in many situations. My only argument is with identifying SaaS with one particular architecture and then using numbers that include all architectures to show growth/adoption.

Steve Bobrowski said...

I posted this at, but got no response. I think it us relevant to this discussion.

"After much reading and thinking, virtualization technology seems to be an elegant solution to most of the problems presented by SaaS development ***AND*** management. During development, there’s no recoding necessary to bring existing single-tenant apps to market or train developers on multi-tenancy issues. In production, use the VMM to quickly provision instances and manage resources shared among VMs, configure high availability options for automatic VM failover, and the list goes on. I’d be very interested to hear about honest experiences/problems people are having when they deploy an on-demand service using the methods proposed by SWSoft ( and WrappedApps ( Please, no marketing hype from vendors!"

Anonymous said...

The fundamental truth in making any architecture successful is

1. Cash
As long as a company generates and has cash in its coffers it can make and take the right decisions to lead the industry.

2. Execution
There may be myriad of architectures as discussed in this post and others connecting to this post. However, a company that has a execution strategy to keep its customers happy and has the focus to penetrate international markets, large enterprises will win.

In the end, people will talk about the architecture that has won over large long lasting customers.

History has shown us this.

Anonymous said...

I have gone through your fine article on future of software(SaaS), ideal for non IT people to subscribe the service to meet their day to day needs as well as due to it's nature of service it has proved cost effective & higher adoption rate.

SaaS business model will be effective and functional in indian market especially for the large scale of existing
beneficiaries,i.e.SMB segment of indian industry due to it's hassle free nature of service.No additional expenses to purchase licensed software copy, IT maintenance, upgade the application, host at data centre. Just pay nominal monthly subscription and use the service. I use to write web site content as free lancer and luckily at present I am associated with a start up company as a content writer.They are also going to launch SaaS product within a month targetting the unorganized sections of indian industry. Having seen your deepen interest with SaaS delivery model I am eager to be in touch with you to share knowledge regarding SaaS.


e-mail me: