Systems give control back to operational managersby
FT.com site; May 04, 2005
By Richard Waters
Systems that automatically consult and update each other offer the chance to put more control back in the hands of operational managers, writes Richard Waters.
Is the corporate world ready for another big vision from the information technology industry? The afterburn from the last one has certainly taken a long time to clear. The 1990s surge in tech spending, fuelled by automation and the internet boom, left most big companies nursing a collection of largely incompatible IT systems.
The promise of seamless electronic commerce was undermined by balkanised departmental-level applications, warring proprietary technologies and the sheer cost of trying to make all this electronic wizardry work together.
"The cost to integrate the applications, to move all the data around, is too high," says Ed Tobin, chief information officer at Colgate-Palmolive, the US household products group. The past half-decade, in short, has been about consolidating and patching together systems that were never really designed to exploit the full potential of the networked age.
That has not stopped the big software companies from conjuring up their next notion of digital nirvana - which, like all new things from the tech world, comes complete with a new set of buzzwords. "Service-oriented architecture", or SOA, is the ungainly banner under which the movement is being advanced, though it also depends on a set of other technologies that together are meant to tame corporate computing, making it more flexible, efficient and subservient to the interests of business managers.
The companies that master this technology will operate more efficiently than their competitors and be quicker to adapt to changing business conditions in their industries, promises Shai Agassi, head of the product and technology group at SAP, the German producer of applications used to automate corporate functions from human resources to customer relationship management. Think Wal-Mart and Dell, he adds: companies that have turned their mastery of process into a powerful competitive advantage. "This will be the age of process innovation."
After the disappointments of earlier generations of technology, it is not surprising that even some of the software companies that broadly agree with this view are cautious about how it will be received. The possible reaction from many customers is that this is just "another pie-in-the-sky idea from a software company", concedes James Utzschneider, head of strategy for the Microsoft division that produces business applications software.
There are also doubts within the tech industry about whether the promises for this new generation of software will be fulfilled. "I've seen this movie before," says Chuck Phillips, co-president of Oracle, the US maker of database software. "I've seen it in several generations (of software). I'm not in the camp that believes in magic."
Yet the ground is being laid for a fundamental shift in the software architecture that underpins corporate IT systems, with potentially wide- ranging effects both on how companies operate and on the working lives of many office staff.
One sign of this is the corporate manoeuvring going on as the biggest software producers position themselves for the changes. Hostilities have broken out most directly between Oracle and SAP, which this year fought to buy a small maker of software for the retail industry called Retek. The bidding war over Retek, won by Oracle, is a side-effect of what have become known as the "stack wars" - the rivalry among a handful of big software companies that each wants to provide the complete package, or "stack", of software on which corporate systems run.
This has taken the software makers away from their previously discrete specialisations. Oracle, with its roots in databases, has been reaching out into applications, most forcefully with its purchase last year of PeopleSoft. SAP, the biggest applications maker, is reaching down into "middleware" - a layer of software plumbing on which applications that run over the internet depend, and which has until now been dominated by IBM, Microsoft and BEA Systems, a specialist US company.
Through grand-sounding initiatives ("Project Fusion" at Oracle, "Project Green" at Microsoft), the companies are also trying to rewrite much of their software, to consolidate on single code-bases after recent acquisitions and to prepare for the intended architectural shift. SAP has declared 2005 a "year of investment" as it sets its sights on a big pay-off starting in 2007. Behind these moves lies a belief that a new wave of spending on corporate software is about to begin, which could shape the software business for the next decade.
The problem that bedevils most corporate IT systems stems from the way companies have automated their activities in the past. This has typically involved "a bunch of separate departments and lines of business trying to optimise their own functions", says Danny Sabbah, head of strategy in IBM's software division.
While boosting the efficiency of individual departments, this approach has fallen short in a number of ways. With a focus on individual functions, there is less attention to broader business objectives. The same "silos" of information that divide departments within a company also prevent companies from co-operating more closely. Many programmers are employed to write fixes that overcome shortcomings.
The solution to this mess could come, broadly speaking, in three parts.
One involves standardising the information held within a company's IT systems so that the same information is made available to all its software applications. "Consolidating the data and having one version of the truth" lies at the heart of this approach, says Mr Phillips at Oracle.
The standardisation of how data are presented also has to stretch beyond individual companies. Work has been under way since the late 1990s on technical standards to allow the automatic interchange of information between different companies' IT systems. This so-called "web services" initiative, which has been driven largely by an alliance between Microsoft and IBM, usually arch-rivals, would introduce a common language for computing.
The second part of the solution involves dismantling the applications themselves to break down their heavy departmental focus. Microsoft, for instance, says it has identified 10 generic processes that take place in any company and span departmental boundaries (the sequence of steps that leads from accepting an order to receiving cash from a customer, for instance, or the process of generating a new customer). These main functions can then be broken down into 1,500 standardised tasks for which software can be written, says Mr Utzschneider.
The third step is to reassemble these smaller, reusable components of software, each representing discrete units of work, into combinations that span departments and link the operations of different companies more closely. As business needs change, the software components can be recombined at will.
If the last generation of technology produced savings by cutting the time it took to complete tasks within business departments, this next generation will do it by speeding up communications with other companies, says Mr Tobin at Colgate. "We need more standard data structures so that we can more easily move information around (between companies) and use components of applications," he says.
In effect, this makes the technology a requisite of the outsourcing economy: without these standards-based linkages, companies will not be able to integrate their operations deeply enough. "A lot of this is about getting companies to focus on their core competences," says Mr Sabbah at IBM.
The extra efficiency and flexibility this promises could eventually have a big impact on business practices. For instance, it could let a company measure more closely what is happening in its production process, then make adjustments that feed back instantly through its chain of suppliers, says Mr Utzschneider.
For many managers, the technology could change the information at their disposal and how they do their jobs. At the moment, much of the data about a company's business is locked up inside software applications such as those produced by SAP. Mostly, this software is the preserve of specialists and lies outside the reach of most managers, says Mr Tobin at Colgate. "There are people who are tied at the hip to (the SAP system), who can make it dance," he says. "But when you move to the management, or decision-making level, they don't naturally use SAP."
That could change if the applications are opened up and made to work better with two other classes of software - analytical tools that help managers make better decisions and collaboration tools that let groups of workers share information and work on joint projects more easily. icrosoft, with a background in worker productivity and e-mail software, has set its sights on trying to merge these different sources of information - one reason why it made a takeover approach to SAP last year. "It will become easier for people to work between e-mail and other unstructured parts of their jobs and structured, transactional parts," promises Mr Utzschneider at Microsoft.
If this is the tantalising promise of the next wave of corporate software, however, there are also still many unresolved questions. When it will come, how much it will cost - and whether it will change the tech landscape in ways that bring significant benefits to both the shareholders and customers of the software companies - are all issues that remain unclear.
Change does not come quickly to corporate IT systems, most of which have been built up over years by layering new generations of technology on top of old ones. "There will be slow progress, it will take years," says Mr Phillips at Oracle. Also, many aspects of the SOA approach, such as how to make it secure once sensitive data flow outside a company's firewall, and how to make much of the information-sharing automatic, have yet to be agreed, he says.
Yet some companies are already starting to adapt parts of their IT systems, using the new standards-based approach to rewrite elements of their software.
IBM is working with 70 customers on projects such as this, which together will bring in more than Dollars 400m in revenues for the company this year, says Mr Sabbah. While these are custom-designed projects by big companies that see an advantage to being on the cutting edge of technology, Big Blue hopes to use the experience it gains to create more standardised approaches that it can start selling to other companies within 12-18 months, he adds.
Others are also trying to turn the experiments into mass-market products. "It doesn't shift into the mainstream until you can package it and buy it off the shelf," says Mr Agassi. SAP says it will have rewritten the main parts of its software by 2007. For most corporate customers, SOA "could be a 2010 phenomenon", says Bruce Richardson, a consultant at AMR, a technology research firm.
In terms of cost, the next generation of software should have one big advantage over the last. The basic information "engines" on which modern multinational companies run - the transaction processing systems and process automation software - are largely in place. The next stage involves opening up these systems through an incremental series of upgrades - something that the software companies claim can be done without the big spending that characterised the mega-projects of the 1990s.
The longer-term costs for customers - and profits for shareholders of software companies - depend on how the new technologies evolve and what they do to the balance of power between buyers and sellers of technology.
Being standards-based, which makes it easier to connect products from different companies, they should leave customers in a stronger position to play off rival suppliers. But the broader software "platforms" being created by groups such as SAP and Oracle may force users to make strategic choices about which company's technology foundations they want to build their business on as a long-term ally. The trade-off for companies that become closely tied to these broader software platforms, says Mr Agassi, will be an explosion of creativity, as independent developers flock to write applications that run on these more robust and open software foundations.
If he is right, that could echo the PC boom of the 1980s and 1990s, which was fed by the ubiquity of Microsoft's operating systems, or the dotcom boom, which drew developers to the platform created by the internet.
A rival view, though, holds that this is an overly idealistic vision of how the technology world will evolve. "Whatever integration you do with SOA will never be as effective as more tightly integrating parts of the stack," says Mr Phillips at Oracle. The benefits of openness, in other words, go only so far.
For companies that depend increasingly on their mastery of the technology underpinning their operations, the outcome will be of vital concern.
- - - - - - - - - - - - - - - - - - - - - - - - -