The recent history of Software-as-a-Service (SaaS) serves in some capacity as a cautionary tale of the burden of hype. You may remember the halcyon days from the height of the dot com boom where companies large and small – with even Microsoft getting in on the act – spoke in not so hushed tones of the imminent death of shrinkwrapped software, with the alternative of “renting” applications delivered in a network context at a (theoretically) lower cost.
There was just one catch: companies balked at the idea – usually pitched by so-called Application Service Providers (ASPs) – big time. Further, many found the very notion itself abhorrent. Most of the major technology players more or less shelved the idea – at least for the enterprise – and many of the smaller firms simply disappeared (I still remember doing an evaluation of Iconomy).
We could dissect the reasons for this vendor/customer disconnect for days, but from where I sit a few of the major showstoppers were the relative paucity of broadband connections, immature (or worse, inelegantly ported) application sets that led to poor performance, and a enterprise inability to accept the idea conceptually. Even for more open minded enterprises, however, that could accept those limitations or weren’t bound by them, there were two major concerns:
1. Allowing mission critical, highly sensitive data to be stored and accessed on systems outside the firewall
2. The inability to customize the application
Fast forward a few years to the present day, and the climate is very different. Night and day different. As is often the case, it now seems that SaaS may have been overhyped, but undervalued. Salesforce.com is growing apace, and we’re seeing the expansion of the model into new traditional application categories like ERP with Intacct (whose application is PHP based, BTW).
Many of the concerns, be they broadband access or immature application sets have been satisfactorily addressed, and the number of businesses that trust their customer data (and what data is more precious?) to Salesforce would indicate that the comfort level with storing information externally is at least less of a concern than it once was. But nonetheless, one major objection remained: the ability to customize the application.
As any systems integrator can tell you, enterprises love to customize their applications. Live for it, in fact. Despite dire warnings of the difficulty in maintaining or upgraded heavily tweaked applications down the line, the committee/consensus based approach used by just about every management or systems consultant virtually guarantees that an enterprise will bend the software to its unique needs rather than adhere as closely as possible to standard workflows and processes. There are exceptions to this of course – SAP implementations being one partial example – but for those used to molding packaged applications into their desired form, SaaS was a tough sell.
The difficulty lies in the approach: the SaaS approach is predicated on economies of scale, in that you design an application and deliver it via the network to many customers. Every customer, in theory, works off the same version, just with a different dataset. It has inherent advantages in feature rollout and ongoing maintenance, but typically offers little in the way of customization. For enterprise buyers long accustomed to having their packaged applications behave the way they want them to, rather than one size fits all, this is a negative.
That, however, may be changing as Jon Udell beat me to pointing out (love the title, BTW, Jon). With a little Firefox extension called Greasemonkey (mentioned previously here), we’re beginning to see the possibility for a slight reversal of this trend. Jon talks about a
extension script here that allows for the introduction of persistent searches in a Virtual Folder like box into Google’s Gmail application. Via CotÃ©, I stumbled across another Gmail Greasemonkey script that allows for the introduction of a long wished for “Delete” button (pictured).
Why are these minor UI alterations to a consumer based (and still invitation only) webmail system important? Because they indicate that web applications – and thus SaaS – may in fact be more malleable than we all believed. Whether that’s a good or bad thing is a debate for another time (although from the look of the abstract here, Forrester’s missing the potential benefits and focusing on the drawbacks), but either way I think we may be getting closer to chewing up SaaS’s last mile. And who’s showing us the way? Google? IBM? Microsoft? Sun? Nope. A 16kb extension called Greasemonkey.
Update: Simon Willison takes Forrester to task for the report here.
Update 2: Jeremy pointed out that in one instance, I was using extension when I should have been using script. Thanks, Jeremy.