By Rachel Chalmers

Blame Merrill Lynch. It seems to be the one that coined the phrase Enterprise Information Portals (EIP). As the market heats up and vendors start to stake out their claims, it remains far from clear what an EIP actually is – and more to the point, exactly whose product qualifies. The story began in November 1998, when ML’s Christopher Shilakes and Julie Tylman released an in-depth report on the emerging space. Their radical conclusion was that EIPs will eventually reach or exceed the investment opportunities provided by the Enterprise Resource Planning (ERP) market. ML’s hard numbers are a $4.4bn market last year, reaching $14.8bn by 2002. A bigger opportunity than SAP, Peoplesoft or Baan? No wonder people are getting hot under the collar.

Tylman and Shilakes identified three market forces which they say are driving the emergence of EIP vendors. First, corporations have started to appreciate that their stored data is a goldmine. Second, off-the-shelf packages encourage IT shops to buy, rather than to build, new access systems. The third – and the one that distinguishes EIP from mere data mining or decision support – is the availability of affordable, ubiquitous distribution channels. That’s the internet and intranet to you and me. We believe Enterprise Information Portals will emerge from a consolidation within and between the business intelligence, content management, datawarehouse, data mart and data management markets, the researchers said.

That breadth of functionality yields a motley crew of players from a variety of backgrounds. Cognos, Hyperion, SAS and SQRIBE muscle in under the business intelligence banner. Documentum and Open Text qualify as content management vendors. Datawarehousing opens the door to IBM, Microsoft, Oracle, Informatica and Sagent, while Prism, Constellar and once-proud Platinum Technology sneak in as data managers. Naturally, every company wants to exaggerate the strengths of its approach and downplay the weaknesses. As a result, it seems no two competitors can agree on exactly what an EIP is.

Consider Brio Technology Inc, whose products ML notes were originally architected for a client/server environment. In 1997, the company introduced its Brio Enterprise web server. Its Enterprise Suite now includes both server products and web, email and C/S-based software. Brio’s offering is said to be particularly strong in the area of analytical calculation, supporting as it does if then else logic, standard deviation, data manipulation functions and ranking. Other strengths include offline analysis, scheduling and printing of analysis and reports.

Turn then to Viador Inc, formerly Infospace, a company which latched onto the November report from Merrill Lynch with considerable glee (CI No 3,612). In stark contrast to Brio’s C/S heritage, Viador’s products are built on an enterprise Java server. Asked whether Viador’s software supports analytical calculation and in particular, the four functions Brio boasts of, VP of US marketing Steven Dille says: Yes, we support all of these capabilities as part of the analyst version of Viador Sage. Fine, but what about offline analysis, scheduling and printing?

We support off line analysis with export capabilities to Excel, says Dille. But he goes on to argue that Brio is pushing the definition of an EIP into the territory in which it feels most comfortable. One notable difference here between us and a client/server tool (like Brio for example since you mention them) is that Viador is more suitable for an enterprise portal, Dille maintains, because we are 100% thin client (only web browser required for any function – no exception). This is the only true scalable portal model. Client/Server tool vendors talk about offline processing because they have client software installed on the desktop…

A portal is different, Dille contends. Portals are browser knowledge systems designed to support all types of users in a mass deployment not a client/server tool just designed to work for analysts doing SQL queries… Analysts are a subset of the requirements and actually 80% of the users never do that type of computing, they simply locate and view information. As he points out, parts of a system like Brio’s are client/server and parts have been made to work on the web. If you want to build an e- business application and connect your partner companies or multiple offices to a portal information server, are you going to require 100 companies to accept a plug-in in their standard desktop configuration? he demands. True thin client is the only real way to scale a portal to 1000s of users.

Pick any two of the EIP contenders listed in the Merrill Lynch report and you’ll likely face a similar argument – different details, but the same main point. Based on the complex interaction of their history and technology, tool vendors have very different approaches to a single problem, and they’re all prepared to argue that theirs is the best and only approach. Ask a third – say, SageMaker Inc, which didn’t even make it into the Merrill Lynch report, but which has been providing information services to the energy industry since 1993 – and you’ll get another answer again. SageMaker integrates publications and publication data, and provides tools for cross publication queries and collation of data, explains the company’s Steve Schiff. Specifically, the product does not support these functions, but rather takes advantage of these functions as they exist in the underlying databases and languages which SageMaker utilizes to create its products.

SageMaker’s Mike Spinney is something of an EIP fundamentalist. He calls on Viador and other players to adhere to the definition set out in the Merrill Lynch report – a definition he summarizes in three points. First, an EIP should be customizable by the user. Second, it should come pre-packaged with external content. Third, it should allow the user to be immediately productive. By his reckoning, Viador misses most of the above, while SageMaker canters home.

Everyone wants to be a market leader. Define any market in terms as complex as enterprise information portal and you can expect hordes of vendors to fight to annex that space. The only way for CIOs to fight through the morass is to have an extremely clear view of what it is they want, defined not in terms of what it is, since category names change as quickly as fashions, but in terms of what it actually does. If deep analysis tools are important to a company, they may be prepared to put up with Brio’s client/server ancestry. If it’s imperative to perform all interaction through a browser interface, Viador might be the way to go. SageMaker’s background in the energy industry should bring another set of skills to the game. As always, the marketing buzzword of the day is sometimes little more than a smokescreen for a complicated minefield of engineering trade-offs.