The claims made for portal technology by some of its most high-profile proponents are compelling: computing giant Hewlett-Packard claims that its portal implementation generated $50 million in savings in just six months; car manufacturer Ford says that, over the course of a year, its corporate portal saves each of its 170,000 employees worldwide over 30 hours of wasted time spent searching for information and switching between disparate applications; and global public relations company Ketchum believes its portal has generated around $12 million in additional sales through improved employee productivity.
When the first portal products emerged more than four years ago, critics dismissed them as little more than a simple web interface branded with a software company's logo. Increasingly, however, portal software vendors such as Epicentric and Plumtree, alongside enterprise application vendors such as SAP and PeopleSoft, argue that their products are now mature enough to allow users to not only view data from underlying applications, but to interact with those applications in a fully integrated manner.
Coupled with a number of convincing blue-chip return on investment (ROI) studies, such claims mean portal technology is back on the list of project priorities within many IT departments.
"Interest in portals has certainly been revived by the slowdown in the economy," explains Laura Ramos, an analyst at Giga Information Group. "A couple of years ago, any department with an IT budget invested in a portal because having a web presence was seen to reflect their standing in the company. Now the investment decision is more considered – a case of 'How can we use this to make more efficient use of resources across departments and functions?'."
The benefits of deploying a corporate portal are well-documented: access to applications and content is centralised, so users do not spend hours searching for information; the management of disparate applications can be centralised; and by allowing users to access 'self-service' applications in areas such as human resources and procurement, organisations spend less time on manual processing tasks.
But how far can prospective customers rely on the ROI figures presented to them by portal technology vendors? In particular, how reliable are the 'ROI calculators' these vendors supply on their web sites?
Iype Abraham, IT director of UK hotel chain Radisson Edwardian, says he paid "no attention" to the ROI calculator offered by his portal supplier, French software company Mediapps. Instead, Abraham chose to focus on tangible benefits that he could measure himself.
"I didn't pay any attention to the ROI calculator because it's quite difficult to prove the actual returns. A lot of it is subjective. What I can measure is the fact that we haven't replaced our desktop hardware for the last three years after deploying a thin-client infrastructure."
The nature of investing in portal technology also makes it hard to assess where the likely returns will come from. Unlike most software investments, the calculation is more complex than a comparison between the cost of the software and the incremental revenue or savings it creates.
One reason for that is that organisations can deploy portals in very different ways and at very different cost levels. They might build a simple web front-end onto existing applications, or integrate and present applications and content through an application server product. Alternatively, they might use a thin-client interface from connectivity specialists such as Citrix Systems. Finally, they might invest in a full-scale portal product from a specialist vendor, where the cost of an enterprise portal software licence can be anywhere between $250,000 and $1 million, depending on the number of users.
Furthermore, the cost of the actual portal technology is minimal compared to the subsequent investments in integration and deployment that IT departments will need to make. "The entry cost of the technology, depending on the choices you make, can be in the hundreds of thousands of pounds," says Abraham of Radisson. "But you don't have to spend that amount of money. The substantial effort is in the integration and component development."
Integration is, in many cases, the most significant cost burden. If users are to transact and interact efficiently with corporate applications via a portal, there needs to be some level of application integration between those applications and the portal interface itself. UK news organisation Reuters, for example, found it had more than 600 individual web sites scattered around the globe. It has since aggregated all this content into one secure portal, supported by process management tools and workflow from ebusiness integration software supplier Tibco.
Charlie Abrahams, European managing director of Plumtree, an enterprise portal software vendor, admits that the services required to install and configure the software for a portal, build the user interface, identify application integration requirements, develop user directories and build secure access, could run into hundreds of thousands of dollars on top of the price of the software licence itself. But, says Abrahams, these costs are multiplied if organisations attempt to build the portal framework themselves.
Plumtree customer Boeing, for example, tried to build its own portal framework several times before it invested in a packaged enterprise portal product. "The first phase of building a portal in-house seems perfectly reasonable, but as soon as you want to add new audiences and functions to that portal, it becomes unmanageable in terms of systems management and maintenance," says Abrahams.
Joachim Zetterlund, European vice president of portal software vendor Corechange agrees: "If you just invest in the bare portal framework it's not costly at all. If you're looking to add applications and new content, analyse it and display that analysis in a graphic format – that involves a lot of additional development work," he says.
There are also a raft of supplementary costs that might not be immediately obvious. For example, the fact that most portal software is hosted centrally and accessed over the corporate network will create new burdens for the staff that manage that network, and also means organisations have to factor in costs such as load balancing and increased server capacity to cope with peaks in usage.
Most organisations will also have to invest in further database capacity to support the rising levels of content going through the portal. And although many portal vendors supply so-called 'portlets' – pieces of middleware that connect directly into most enterprise applications – every time that an application is customised or an organisation upgrades that application, the portlet becomes useless because it is 'hardwired' into one particular configuration.
Given the potential for costs to spiral out of control, it is not surprising that some portal projects never get off the ground. Hewlett-Packard's @HP portal project, for example, almost got culled because some HP executives believed it was "something untested and unknown".
Now the portal covers almost every function within HP, including those resulting from its merger with Compaq, and has created $50 million in savings in the human resources department alone. According to Barry O'Connell, HP's general manager of business-to-employee (B2E) solutions, this is because every phase of the project was focused on one overall goal, and was tied to demonstrable benefits at each stage. Just distributing payslips electronically through the portal can save up to $9 a month per employee, says O'Connell.
This focus on tangible benefits tied to specific business processes, concludes Ramos at Giga, is what organisations should be aiming for. "Ultimately, the business model for a portal should focus on a specific audience, and how using that portal will make a business process run better," she says. "If companies judge the success of their portal project on woolly metrics such as improved productivity, they're going down the wrong path."