At the high end of the platform-as-a-service spectrum, there's a cluster of vendors that offer fully templated but still highly customizable business applications, targeting small to mid-size businesses and departmental managers in larger organizations. Interestingly, in a poll of ZDNet readers I ran earlier in the month asking where you'd prefer to develop SaaS applications, this class of platform was the most popular choice, beating out cloud hosting alternatives such as Amazon EC2, and leaving hosted development platforms such as Salesforce.com's Force.com and the impressive startup Bungee Labs well behind in fourth place. True, the poll was no more scientific than the infamous show-of-hands I orchestrated at last month's SaaS Summit, but perhaps the results shouldn't be dismissed too lightly. As computing shifts to the cloud, the way in which vendors enable customization may become the key determinant of success or failure. It all revolves around what kind of platform for customization the market really wants.
I had an interesting couple of meetings last week relating to this theme. One was with NetSuite's Zach Nelson, who has reappeared after the company's recent IPO sporting a beard (see picture). More on that below.
The other meeting was a pre-briefing with Pankaj Malviya, CEO and founder of Relationals, the company behind the LongJump PaaS offering, which today takes a major step forward with the addition of a visual workflow designer (partial screenshot follows).
I've spoken to a lot of PaaS vendors in recent months and one of the observations I've made is that customization is a curse as well as a blessing. If all a vendor offers is customization, then it's like offering a blank sheet of paper, which has two adverse results. First of all, people don't know where to start, so most of them don't bother. That's frustrating enough, but it gets worse. Those that do bother always seem to come up with needs that aren't yet built into the platform. So here's the paradoxical curse of customization: it's always too much, and it's never enough.
Smart vendors are the ones that, like LongJump, go the extra mile and actually build applications that their business customers will find useful. This tackles both aspects of the curse of customization: you give people somewhere to start, and you constrain them into choices where you at least have some idea of what their needs are likely to be, so you can make sure you've built those needs into the platform already.
LongJump offers a series of ready-built customizable Web applications for sales, marketing, HR and administration that are designed to offer business utility as well as customization. "We don't believe that giving a blank sheet really works," Malviya told me. "You can build an application from scratch but we believe the best way is to use one of our applications and then build on top of it ... We are providing just 14 applications but each one is useful, has been designed comprehensively, and is making money for us."
LongJump's typical customer is a 50-person company or department in a large organization and the flat-rate cost (irrespective of the number of applications deployed to each user) is $19.95 per user per month. The applications are all based around a database with forms and event processing that acts on it. "The concept is a database on demand with a UI on top that enables you build your business processes around that data and make that data very actionable," Malviya said. Available actions include sending object records to a SOAP API for integration with other data sources or applications.
The visual workflow designer introduced today provides the ability to build a process around any object. The conditional workflow is created in a graphical format and code is generated automatically when the workflow is saved. External objects can be included in the workflow by, eg sending an email to an outside party and then pending for a manual update of the response coming back, or a call to an API.
Some people will tell you that the sector of the market Longjump targets simply isn't interested in the kind of extensive customization the company offers — especially not with a $20-a-month price tag on it. These people — quite often industry analysts or product managers for enterprise software companies — insist that 50-person companies have neither the motivation nor the skill to embark on complex customizations. The conventional wisdom says that such companies are looking for cookie-cutter implementations of best practices that they can just get working with right away.
That viewpoint got short shrift in my second encounter of last week, this time with NetSuite's CEO Zach Nelson. He was in London last Tuesday and speaking at a one-day conference on SaaS organized by the software research team at Société Générale Corporate and Investment Banking. Sporting a beard freshly grown over the Christmas/New Year vacation while relaxing after the company's mid-December Nasdaq debut, Nelson is currently in the midst of a multi-week circumnavigation of the globe, meeting customers and prospects in locations as disparate as Toronto, Dublin and Manila.
NetSuite recently released what it calls its Business Operating System, which is the branding it gives to the set of tools that allow customers and partners to customize the NetSuite business application to their own needs. Nelson was very keen for his audience of investors and analysts to understand how crucial this capability is for NetSuite in its target market. "Every company requires customization," he told them. "Nobody buys 'vanilla' NetSuite." To illustrate the point, he invited a customer on stage — Mike McGinn, project implementation manager at Hugh Symons Telecom, a mobile telecoms solution provider based in Poole, with a 45-strong headcount.
Here is a perfect example of a relatively small company, with just one dedicated IT guy and limited resources, but still keen to use IT to gain competitive advantage and improve operational effectiveness. In a live demonstration of how the company has adapted the NetSuite application, McGinn ran through examples of custom roles, custom workflow and custom reports that had been set up. He showed the ISO compliance processes that the company had custom-written into its NetSuite implementation. All of this had taken a little over three months to implement, replacing a dozen or more applications previously in use at the company. McGinn's own estimate is that the first-year cost comes to around $200k — but that developing the same capability using on-premise alternatives would have cost a million or more, beyond what his company could afford.
Nelson contrasted this example with suggestions from SAP that customers of its Business ByDesign SaaS offering will be able to be up and running in a matter of days. "Our experience tells us that not making it customizable is a non-starter," he said. "I hope that's what SAP has developed." But at the same time Nelson warned against going too far in the other direction, citing Salesforce's Force.com platform as an example of a platform that offers too much customization. "[NetSuite is] designed to run a business. That's the data model we created. Don't change the data model."
So the message from both NetSuite and LongJump is that constrained customization is the best path to success. First of all pick a platform to meet your broad business objectives and avoid the curse of too much choice. Then customize within that platform to your specific process needs. But to reap the full blessings, don't limit the customization options any further than that.
Something else I liked about the LongJump approach was that the company also sells through partners — an approach that NetSuite is keen to develop too. I suspect that this type of platform is going to be an opportunity for partners, although as Nelson said last week it will be a painful adjustment for some: "There's a world of hurt coming to the IT services industry based on the SaaS model." Projects will be smaller, but there will be more scope to build up expertise and develop reusable intellectual property in vertical markets. LongJump works through third-party developers and ISVs, who buy at an OEM price and add their own margin. Unlike many others in the PaaS environment, LongJump hosts the products under the ISV's own branding. "These companies want to own the customer relationship," Malviya explained.