A heated TalkBack discussion developed yesterday in response to my posting on the topic of IT shops 'afraid of SaaS'. In the midst of it, SaaS advocate Matt Robinson complained:
"This forum seems to be a feeding frenzy for SaaS detractors, however, the arguments presented aren't providing much constructive criticism ..."
It had all started with a TalkBack comment from Erik Engbrecht, in which he protested that a SaaS solution wouldn't allow him to make the inevitable changes that were bound to arise with any application:
"I've NEVER seen an off-the-shelf application that performed as sold and didn't end up needing a pile of customization.
"So why would I, given that the probability of a vendor meeting my needs hovers around 0%, rely on that vendor to meet them?"
Inevitably, the question of customization quickly became central to the discussion, and whether configuration of the application — as distinct from custom coding — offered enough flexibility. This was Matt Robinson's take:
"I am curious to know if many of the detractors here have actually evaluated any SaaS offerings ... I think the line between configuration and customization is becoming very blurred due to advances in SaaS platform capabilities. Platform vendors like CogHead, SFDC with their new BP/workflow editor (due out soon from what I hear), and NetSuite's NetFlex and associated scripting engine are allowing deeper and broader control over how applications perform and function in their respective SaaS environments. It would be interesting to learn more about what types of things businesses want to actually customize at the code level that cannot be touched in these existing and upcoming SaaS offerings. The more we learn from these needs the more we, as SaaS vendors, can expose and therefore integrate these concepts further."
Contributor jorwell stepped in to raise integration as another matter of concern:
"I have some serious misgivings about how one would sensibly integrate SaaS with other data held elsewhere, I suspect most SaaS vendors are touting either the web services mantra (which I believe to be too inflexible and code intensive to be seriously considered) or it would involve a lot of data redundancy (which is even worse)."
In response, Matt Robinson enlisted enterprise mashups to bolster his argument:
"SaaS platforms should have the capability to add mashup and integration components on the fly for all customers to use as needed ... They will be available as needed in your SaaS applications and will require less and less custom coding versus point & click, drag & drag 'development'. Business users will be able to peform these tasks through powerful but increasingly easy to use UIs."
This was all too much for Erik Engbrecht, who countered:
"Running a business on a piece of spaghetti code that connects six different services that I do not control and could change at any moment is irresponsible."
It was about this time that Matt Robinson made his comment about a 'feeding frenzy' against SaaS among TalkBack contributors. But he couldn't resist a partisan sideswipe of his own against in-house implementations:
"People also tend to forget or overlook the fact that ADP is mostly SaaS, online banking is SaaS, Yahoo Mail and GMail is SaaS, Yahoo and Amazon and Google and eBay are becoming, to varying extents, SaaS platforms — these are some of the most highly utilized applications in existence. Most enterprise apps installed in-house have significantly higher downtime and larger security holes than any of these systems."
It all seemed to endorse the conclusion reached in the original post, that people who work for IT shops tend to start out biased against SaaS. Another contributor, mmankow, advised SaaS vendors to make sure other departments besides IT are involved in the buying decision:
"When selling SaaS, surely the IT department should be involved, but our research shows that the CEO, CMO, CFO should all have a vested interest in SaaS, considering its value proposition."
Then today, Jim Howard, the SaaS vendor CEO interviewed in the original posting, added his own response:
"Many of CrownPeak’s clients are working with a SaaS for the first time. They had to believe in this model and our company before they would buy from us. Happily, today we have a large and diverse customer base, and the amount of faith it takes to believe in an SaaS versus the amount of faith it takes to believe in a custom development project is starting to even out. Soon, we believe, the question will be, 'how can we justify taking on a complex IT project when we can simply flip a switch with SaaS — at lower cost?' ...
"My core message in this briefing is this: The reasons given to eliminate a SaaS from consideration are not typically valid. Poor security, loss of control, inability to integrate, loss of flexibility. Today, CrownPeak and companies like us sell to a portion of the IT audience. Most of that audience rejects SaaS as an option because they don’t yet understand how powerful and mature the solutions have become. Folks haven’t yet learned that SaaS is a great method for developers to do what they do best — deliver very specific customizations for their business. SaaS actually enables those customizations in a better supported, lower cost, lower risk way."
What do you think? Is IT too hidebound to embrace SaaS, or are there good grounds for resisting? The debate continues. To add your views, whether to promote the merits of SaaS or to defend IT for steering clear of it, simply click here.