Personal service builders

"From web services and service oriented architecture (SOA) will emerge a new kind of application development using tools called 'personal service builders'," according to Jonathan Sapir, quoted in an ADTmag.com news story this week.

Sapir is CEO of InfoPower Systems, a software development company which has just such a tool in alpha testing. "He sees personal services assembled into larger business applications as the next trend in corporate IT, but with most of the actual development being done outside IT by individual business users."

Sapir is not alone in making this claim. Proponents of next-generation business process management (BPM) talk about allowing business users to directly manipulate process automation within the enterprise. Howard Smith and Peter Fingar, co-authors of Business Process Management: The Third Wave explicitly liken this desktop manipulation of process to the breakthrough numerical manipulation enabled by spreadsheets on the PC:

"To the business, the PC loaded with a spreadsheet meant a radical simplification of routine calculations, transferring to the everyday businessperson a function that had once required special programming skills. Today, a similar symbiotic relationship is emerging between web services and BPM ... What the spreadsheet did for numerical computation BPM will do for process work."

(The article from which this quote is taken, by the way, is a riposte to Nigel Carr's controversial HBR article, IT Doesn't Matter, and answers it with an eloquent exposition of how we are currently in the middle of a shift from "Information Technology" to "Process Technology".)

Microsoft is investing in the same trend. The next generation of Windows/Office/Outlook/InfoPath is being designed with the intention of becoming the world's favorite PSB (to borrow Jonathan Sapir's acronym). Bill Gates gave a flavor of what's coming in a briefing to financial analysts this week: "There will be major advances in the user interface … Web services will be built in as part of the interface ... Longhorn will be built around scenarios, and making these as easy as possible. Longhorn is not just a release of the Windows client, but it will also involve Office and our server products."

So InfoPower has tapped into a trend that is already quite powerful, but in doing so has given us a new acronym to describe it that I rather like. I see Personal Service Builders (PSBs) as the desktop element of the emerging trend in on-demand assembly of process automation. By happy coincidence, it partners well with the up-and-coming acronym on which the underlying service infrastructure is going to be built — ESB. To emphasize the synergy between the two (and in deference to Jeff Schneider's sensitivities), perhaps the latter acronym should be tweaked so that the 'B' no longer stands for Bus, but for Builder. Either way, the two concepts of ESB and PSB will go hand-in-hand as we make our way forward into the service-oriented future of business automation.
posted by Phil Wainewright 2:07 AM (GMT) | comments | link

Early adopters put management first

BAT, BT, AgentWare and Providence Health System all advise putting in web services management right from the start in our latest feature article, Getting to grips with web services. It's exactly what the vendors have been saying all along, but then they would say that, wouldn't they? It makes a difference when it's their customers who say it, especially when the conviction with which they say it fairly jumps off the page at you.

The need for management that these early adopters describe in such illuminating detail is exactly the kind of SOA Reality that CBDI Forum's analysts conjure up with their latest commentary: "[SOA] is not the silver bullet that many are suggesting; it's plain hard work." As our article shows, managing SOA is a real headache that requires an investment of forethought and resources.

The CBDI analysis continues with a useful breakdown of the definition of SOA, and then goes on to discuss what's involved in enabling true loose coupling: "It is in the component architecture and design that many of the issues that impact loose coupling will be resolved."

There's an interesting discussion brewing below the analysis, with a reader asking, "Would document style web services not be looser coupled than RPC style?" CBDI's David Sprott replies, "It would be a mistake to think that document style services are inherently loose coupled — there may well be tight business rule dependencies," while Lawrence Wilkes adds: "There are often hidden 'back door' dependencies in existing systems that prevent the switching of services no matter how loosely coupled the messaging appears to be." All important considerations.

Another warning voice about rushing into web services is John Hagel's: "Opportunistic, one-off deployments of the technology may solve near-term business problems but, unless they are designed to be consistent with a broader architectural vision, they will contribute nothing towards longer-term value creation. In some cases, they may actually make it more difficult to implement a consistent architecture later."

But this doesn't mean you should hold back from going ahead. As John explains, it's quite possible to get immediate returns from implementing business-driven projects now, without sacrificing the greater long-term benefits of adopting a strategically planned SOA framework.

The users we've spoken to provide palpable evidence that some organizations are indeed doing both.
posted by Phil Wainewright 8:57 AM (GMT) | comments | link

Reality check

Jeff Schneider poured some much-needed cold water over the hot new ESB acronym in his weblog a few days ago. His assortment of spoof definitions hits home, in particular this one: "It is a means for a message queue company (like Sonic and Spirit) to make it seem like they do web services when they really spent the majority of their time doing JMS based queues."

He's also right to point out how stupid it is to invent a term that describes a service network as a bus, of all things. But sometimes the buzzwords we end up with — just like the technologies they describe — aren't always the most elegant, precisely-engineered phrases we could have chosen. They just happen to be the ones everyone agrees to use. To my mind, ESB will do just fine, so long as everyone remembers: it's not a bus, and it doesn't depend on JMS.

Anyway, I was glad I got pointed to Jeff's blog because it's a while since I've taken a look, and frankly we should add it to the Loosely Coupled news and blogs aggregator page. The main reason we haven't is because the page isn't yet a true aggregator. It's waiting on some development work that will let it show postings according to when they're published rather than who wrote them. And that in turn is waiting on some clarity to emerge from recent developments in the RSS arena.

Jeff was commenting earlier this month about web service frameworks (WSF), providing some useful context to all the current standards debates: "I am of the opinion that the WSF is where MS & IBM wil have their next major battle. So, they HAVE to agree on a certain sub-set of functionality (those concerns which must be remedied at the protcol level), but beyond protocols, the field is open and the battle will begin shortly."

Another entry discussing databases and SOA asks some interesting questions about the web services API for salesforce.com's sforce initiative: "After you look at these questions, you can begin to see some of the problems that will be surfacing in providing external access to internal data. The art of service enabling the db is still early — let's learn from sForce and advance the art."

All in all, Jeff's blog is a goldmine of refreshingly grounded commentary and well worth a regular visit.
posted by Phil Wainewright 7:25 AM (GMT) | comments | link

Missing the (inflection) point

I never cease to be amazed at how consistently and comprehensively Siebel has failed to understand the true nature and threat of hosted services. Future generations of business school students will marvel at the company's unwitting prowess in becoming a perfect case study illustration of Clayton Christensen's The Innovator's Dilemma, achieved while that book remains in the business bestseller lists, and yet without ever once having any inkling of how perfectly it fits the profile: a successful company doomed to failure by its inabilty to embrace an emerging disruptive technology.

The latest episode in this sorry saga comes in a CNET news story, Siebel, IBM preparing online CRM service, which reports sources saying the two vendors are preparing "a service in which Siebel's business applications will be hosted by IBM Global Services." The calamitous extent to which these two companies just do not get it is revealed in the very next paragraph:

"Like similar services from Salesforce.com and Upshot, the partnership between Siebel and IBM will offer a slimmed-down version of Siebel's sprawling CRM applications delivered over the Web for a recurring fee, the sources said."

Frankly, I find myself shocked and appalled that there are still people around who remain convinced that the ASP model is simply a matter of what for several years I've characterized as serving up "cut-down software for cut-down companies." If that's all that Siebel and IBM have in mind (and unfortunately there's every reason to suppose it is), then let me put the record straight here and now. That concept bears no relationship whatsoever to the hosted services offered by Salesforce.com, UpShot and their like.

Siebel's error is to perceive hosting as what Christensen calls a 'sustaining technology' — one that marginally enhances the established client-server paradigm on which Siebel's current products are founded. The company's commitment to that paradigm is so ingrained that it is incapable of recognizing the 'disruptive technology' underlying the hosted services of the emerging companies who now appear destined to usurp its market position.

What Siebel is thinking of is application outsourcing, in which software is hosted in a provider's data center rather than being installed at the customer site. In some circumstances, there can be cost savings, but it's marginal. Coincidentally, I recorded an IT Conversation with Doug Kaye on this very topic last week, in which I describe the specific conditions in which that form of application outsourcing makes sense. Also last week, I cited an analysis by Microsoft Research scientist Jim Gray, which demonstrates why the economics of conventional application outsourcing don't stack up. Unless you offer a truly shared service, you're just adding telecoms costs to an already largely defunct model.

That's why, as the eponymous company's CEO Tom Siebel himself said last year, "it's just not how people want to buy software." The really bad news for Siebel — along with Oracle, PeopleSoft, SAP and every other packaged enterprise applications vendor — is that people don't want to buy software at all. They don't even want to pay for it — and with hosted services, they don't have to.

Hosted services use a shared-service model that, instead of dedicating a bundle of software code to each individual customer, relies on a completely shared, service-oriented application infrastructure, with delegated, self-service configuration at the process and presentation layer. This shares the cost of the software at a much deeper level than conventional application outsourcing. But there's more.

The most successful of these hosted service ASPs cut their costs even further by using free software. Their services are built on open-source platforms such as Linux, Apache and Perl. They've eliminated software as a cost component.

Tim O'Reilly has been giving interviews recently about how Google and Amazon exemplify this same approach to software: "Don't try to be proprietary. Build software and deliver the services over the net, and be profitable with those services," he says in The Independent today. Earlier this month, he expanded on the theme for InfoWorld: "Once you're no longer distributing an application, none of the licenses mean squat ... I think we're going to find more and more places where that happens, where somebody gets a critical mass of customers and data and that becomes their source of value ..."

Amazon last week expanded on that strategy when it said it will start offering its online payment processing as a service to partner websites, including support for controlled access to web content (ie subscription services). This is a hosted service with all the same characteristics as those of web-native ASPs, and Tim is right: this is the future of software (or perhaps, as I called it recently, The end of software).

Of course, there are many obstacles that still have to be overcome. Some of the most challenging ones are outlined in a transcript of the middle section of my IT Conversation with Doug Kaye, which we published here on Loosely Coupled earlier today. But some insiders have already seen the writing on the wall for the established enterprise software vendors — last week salesforce.com announced Peter Gassner, a former VP and general manager at PeopleSoft, has jumped ship to head up the ASP's hosted service integration initiative, sforce.

Will any of this make any impact at Siebel? Unlikely. The company's last online venture, sales.com, which masqueraded as a hosted service play, was in reality an attempt to cash in on the dot-com boom with a quick-fire IPO that, like so many similar initiatives, crashed and burned with the pricking of the bubble. Siebel has never seriously investigated hosted services and seems destined to miss the point yet again — or, to borrow former Intel chief Andy Grove's phrase, miss the inflection point, "when the balance of forces shifts from the old structure, from the old ways of doing business and the old ways of competing, to the new" (quoted from Only the Paranoid Survive).

Siebel executives will doubtless be reassured by research from Forrester, quoted in the CNET story, which finds that, while hosted CRM revenue is growing faster than sales of licensed software, "hosted revenue will remain a fraction of the total CRM market." But their calculations will once again be off the mark. Whereas Siebel collects the whole license fee upfront, a hosted provider collects the equivalent sum spread over 36 monthly payments, so a year's hosted revenue equates to three times as much as Siebel revenue. And by offering shared services based on open-source software, hosted competitors can undercut Siebel's prices by an order of magnitude, which adds another 10x factor into the equation.

Put the two together, and it means every dollar of revenue hosted service providers earn at the expense of Siebel is thirty dollars denied to the incumbent — and thus 3.2% market share already equates to parity. The downside for the newcomers is that they have to acquire thirty times as many customers as Siebel to equal its revenues; but at prices like those, it shouldn't take them long.
posted by Phil Wainewright 5:40 PM (GMT) | comments | link