How to choose a SaaS provider: understanding the selection box

Mars mini chocolates in a bowl

The choice of business software used to be comparatively straightforward. A fairly basic algorithmic calculation combining usefulness (these days we say ‘functionality’) and price would typically find the winner. But things have changed.

These days, companies need to look at delivery options: they need to weigh up how their software is delivered, the costs involved (and that’s not always a straightforward calculation) and then there is integration with legacy software to consider.

Has buying business software become easier, or in fact more complex? Has the advent of cloud opened up the applications market so much that we now have almost too much choice? Do we need some kind of barometer or gauge to assess how much “functionality” we are getting in comparison to how much “ease of integration” (for example) or is there no trade off as such to be made?

If cloud means that we now have a broad SaaS “selection box” to choose from, how do we avoid getting the sickly chocolate orange cream that nobody wants?

CA Technologies' UK CEO Mike Gregoire provided a brief history of enterprise software time at his company’s recent ‘expo ’13’ event in London this winter. Think back perhaps three paradigm shift cycles (they come in five year blocks right?) urged Gregoire; something like around 15 years ago we were very focused on custom/bespoke application development at the enterprise level and this was the accepted norm.

Then jump forward to about a decade ago. Firms like SAP and IBM had made sure we were focused on a higher-level route to applications through chunky ERP suites that would act as platforms to so-called ‘systems of record’ and, potentially, all the enterprise functionality we needed eventually.

Cloud shook things up a bit as did Bring Your Own Device mobile application usage and ubiquitous broadband web connectivity. This meant that the last five years has given rise to what we call the ‘application centric’ world. So as we turn to cloud computing to get these applications, how do we rate one SaaS offering off against another?

Out there in the real world, is it really mostly like for like? ie is an HR app from one SaaS provider relatively “vanilla-flavoured software” in terms of core functional performance and, if so, is it therefore down to value add layers like services and maintenance for any individual SaaS provider to distinguish their market proposition?

You only have to look at Rackspace’s enthusiastically named Fanatical Support proposition to see where that particular business model rests its foundations. For some customers, having a real person on the end of the phone 24/7 for support is what cuts the mustard. Amazon is largely perceived to do things a little differently and is known for its vast libraries of technical documentation which, in effect, form a substantial element of its support proposition. Microsoft’s Azure is of course fairly meaty, but not quite differentiated in the same way as the two other firms’ businesses.

SaaS size matters

So should smaller SME/SMB customers opt for ‘boutique’ hosting providers who tailor services to the smaller end of the market, or doesn’t it make any difference?

Steve Kaplan is VP of channel and strategy sales at SAN-less virtualised datacentre platform company Nutanix. Kaplan says that the size of the SaaS provider is probably “not very relevant” other than in the context of effectively providing support and other essential operational attributes.

“Smaller SME/SMB customers typically lack the luxury of maintaining developers on staff and of course the applications they do select, whether on-premise or hosted, are key to their business success. These firms’ SaaS selection focus should be entirely based upon selecting the right applications for their business. Although that might sounds obvious at first, the need to perform the same levels of due diligence in terms of reputation for support, performance, security, reliability and disaster recovery remain constant at whatever size SaaS implementation,” said Kaplan.

The real secret to SaaS selection lies in perhaps in realising that the same enterprise application functionality can be delivered both on-premise or via SaaS. So we should remember that functionality is “separate from and not dependent upon” the delivery model. This is the view of Martin Moran, managing director for international at service, subscriptions and support focused SaaS company ServiceSource.

“Think of the number of current enterprise application providers offering both [cloud and on-premise],” said Moran. “Whilst SaaS enables consumption more immediately and to a wider range of customers, it also enables those customers to switch allegiance to other providers. This means that the ‘trauma cost’ of moving a legacy premise-based enterprise application suite is significantly reduced if it can simply be ‘switched on and off’ like a tap. So customer loyalty is much more the active choice of the customer rather than being affected by legacy choices made in the past.”

What Moran hints at here is that selecting a SaaS choice is all about remembering those words ‘choice’ and ‘selection’, ie, the ability to switch providers and switch applications openly should be an engrained element of our IT architecture as we move to cloud.

This will mean that we will need to manage and store our data more intelligently if it is to be transferrable across application streams, but this is not an impossibility if you look at the way open source big data is evolving in areas such as Apache Hadoop and the inroads that firms like Hortonworks (for want of just one example) are making.

Developer evangelist at SAP Ian Thain wrote on his mobile-focused blog recently that we should not be afraid to fail cheaply and repeatedly. Although Thain is talking specifically about mobile application design thinking, mobile is cloud for much of its enterprise existence these days anyway.

Choice of SaaS is important, but if we continue down the current road with data analytics and build our intelligence in managing what will be an increasingly unstructured world of data with the Internet of Things, could SaaS options become more commoditised? Not just yet perhaps, in the meantime, pass the green wrapper hazelnut triangle please.