When Debating “The Economics of SaaS,” Lets Not Use Fuzzy Math

I recently read with great interest a story in BusinessWeek titled, “The Economics of SaaS,” penned by a group of writers known as the "Staff of the Corporate Executive Board.” In it they meticulously detail out the cost savings realized from SaaS solutions vs. the equivalent on-premise solutions. Personally, I am shocked at how this type of comparison continues unabated. Is this a last ditch effort from some of the die hard on premise hold-outs trying desperately to prove to the world that it actually makes more sense to install a separate copy of software at each customer site? I think it is important to look at the following considerations, to make sure they are being contemplated in this comparison:

1. Generally speaking, a SaaS solution that has a competitive on-premise offering will be priced around the same as the maintenance fees the on-premise vendor will charge. Yes, if you take Salesforce.com's published list price and do the math times 500 users it will look expensive, but it doesn't work that way. The unit price goes down as the volume of users increases.
   
2. Integration does not cost more when integrating SaaS, it costs less.  If you use legacy integration products, then yes it will cost more as you will have to code adapters to support the SaaS application(s) you are trying to integrate, install software into your DMZ, potentially spin up instances on Amazon, etc. However, some integration providers serving the SaaS market offer a product that is itself SaaS, driving the cost of integration down to a fraction of what it used to be in the on-premise world.
   
3. All bets are off if the approach of the "SaaS provider" is to install their single-tenant software in Amazon and call it "SaaS".  The primary principle of SaaS, multi-tenancy, is eroded.  This provider will have a much higher baseline of costs because they need to provisioning hardware (regardless if physical or virtual) for each customer they sign up, and they will simply pass onto the customer.  Contrast that with Salesforce.com, that claims 500 servers are powering its 60,000 SFA customers. That is the benefit of multi-tenancy realized.  

The favorite point that people like to make when comparing SaaS to on-premise is how in years two and three the cost savings for SaaS become marginalized against the investment made to "own" your on-premise software.  But what if we factored in to this math a comparison of the number of releases made by the vendor and utilized by the customer?  One of the top reasons on-premise customers move to SaaS is the "rev lock" they face, where due to the complexity and cost of the customizations and integrations made to their specific version of some on-premise application, the cost to upgrade to the latest release rivals the initial investment made in product in the first place.  Yet regardless of this, the customer must still pay their maintenance fees. Compare that to SaaS, where generally providers do four significant releases per year, and 100% of their customers get 100% of these releases.  That's because the SaaS vendor must build the product in such a way that they can upgrade it for their customers, and not break them or have any impact on their business operations at all.  This model works and is proven not just by the leaders such as Salesforce.com, but look at most well known SaaS ISV's and they will showcase how well this architecture works.

Comments

[abm]

I wonder if the costs argument has been won, while the cases that highlight indemnification and remediation for business continuity issues start to look more onerous.

Right now, now matter how SAAS (Even more so PAAS) holds up to costs and operational advantages, the SME that operates real time, commercial infrastructure is having a devil of a time getting major professional lines insurance coverage. why is this?

1. Most SAAS and PAAS companies are private, under-capitalized, and closed to outside certifications and ratings.

2. Underwriters can walk onto a SME premises and inspect the server closet -

3. The insurance industry has just not caught up to the SAAS and PAAS vendors, even the most august who have gained SAS70.

[ranc]

I”ve had some clients really explore the discovery phase and approve wireframes and then design and the result was something I actually wanted in my portfolio and it worked great for the client too