Time to be honest about SaaS
I thought that those of you who aren't recipients of our monthly newsletter might be interested in this commentary (penned by the other Neil) dissecting some of the problems with the definition (or lack thereof) of software-as-a-service.
Over the past few days we’ve been having an interesting debate here at MWD, in conjunction with the analysts at our close partner
Freeform Dynamics. The question came from Dale Vile at Freeform: what's a good definition of software-as-a-service (SaaS)? The reason for asking the question was that SaaS is a hot topic, and it's something that's considered as a major growth opportunity for a lot of technology suppliers; but although there are a fair few forecasts of growth in demand, it's difficult to get a clear idea of what's actually included in these forecasts. Of course, if there isn't a consistent view of what does and doesn't actually constitute SaaS then that's not really helping anyone. The approach we took to try and provide that consistent view was to look at a long list of things (ranging from Google Search, Google Maps and hosted wikis to Skype's VOIP and messaging services, hosted voice PBXs, online travel agency services and remote backup services), and say whether we thought they "counted" as SaaS offerings.
What came to the fore very quickly was that there was no crisp set of attributes that we could agree characterised SaaS offerings. SaaS isn't defined (as some would tell you) by a particular type of distribution or access technology, a particular technology architecture, or a particular approach to charging for usage.
Yes, SaaS offerings do commonly exhibit particular choices in these areas (use of the web for distribution and access; a "multi-tenant" architecture to efficiently separate the data and customisations of each customer from those of others; and some kind of subscription license). But crucially, these choices aren't unique to what most people would call SaaS offerings. Google's services, and countless millions of other online dynamic websites, have made those same technology choices for distribution and access – and they're commonly lumped into that whole other can of slippery worms, "Web 2.0". Countless online portals (some hosted within organisations, others available to the public) allow users to personalise their experiences and use a multi-tenant architecture to store personalisation data efficiently and effectively. Lastly, all sorts of information- or IT-based capabilities are delivered on a subscription basis (not least, mainframe capacity, and analyst research ;-).
So what is it that marks something out as SaaS (or not)? The only answer that seems to tick all the boxes is that SaaS offerings are those which deliver online, hosted alternatives to things that we have historically experienced through the in-house purchase (or development) and deployment of software systems.
Let's take Customer Relationship Management (CRM) as an example. Historically, CRM capabilities were provided by software that was installed on premise, was managed on premise, supported one organisation, and was paid for through a perpetual license. When Salesforce.com delivers those CRM capabilities from a remote installation, manages them on behalf of multiple organisations, and is paid to do so on the basis of a monthly subscription, it needs a different name: that's "Software-as-a-Service". Following that example, remote backup/restore services, online word processing applications like Google Docs, the Zoho suite and (now Adobe's) Buzzword, and SAP's BusinessByDesign (formerly A1S) all count as SaaS offerings. Google Search and Facebook don't, because they're not delivering capabilities that you would ever have associated with on-premise, perpetually licensed software.
This helps us clarify SaaS' place in the IT industry, but we think it's a problematic conclusion, for three reasons. Firstly, most people use the label without really understanding how context-dependent it is (what you think of as SaaS is primarily defined by your own experience); secondly, if we continue down this road, there can never really be a consistent definition of SaaS that will work for everyone; and thirdly, this is a very IT industry- and supplier-centric way of looking at the world that is only likely to alienate or confuse a very important community – "users" (the people who pay all our salaries).
Perhaps we need to call time on SaaS, and think of some clearer terms and definitions that can really help IT organisations and IT buyers work out how everything fits together. At the very least, as an industry we need to be honest about SaaS – and explain that it’s an industry-driven marketing and positioning term that's primarily about separating "funky new stuff" from "boring old stuff".
Labels: Adobe, Facebook, google, SaaS, Salesforce, SAP, Skype, Web 2.0, Zoho
