Intranet portals (also called "enterprise portals") are hot buzzwords in an industry renowned for gorging on buzzwords. So, are these portals a desperate attempt by software vendors to profit from the chaos and poor usability that define most big intranets, or are they a way to help employees do their jobs better? The answer to both questions seems to be a resounding "maybe."

Portals can help employees find information and perform their jobs, though few portal designs are optimal out-of-the-box. Also, designers can achieve some of the portal usability benefits through simpler means, especially in smaller companies. Most importantly, at all the companies we studied, the key issues in building a good intranet portal were political and organizational -- not technical. Basically, buying software won't get you a good portal unless you also manage internal company politics. Technology accounts for roughly one-third of the work in launching a good portal; internal processes account for the rest.

The real challenge is to get contributors from individual departments to comply with the portal rules, enter decent meta-data, and refrain from fielding maverick intranet servers outside the portal's scope. Many of the portal teams we have talked to had successfully used some of the standard tricks for generating buy-in, such as establishing an advisory group with representatives from important departments and getting this group involved early in the project.

Other standard tricks: ensure top management backing, and make it easier for departments to join the portal than to do their own thing. Interestingly, the current bad economy can help on the latter point, as departments will relish the opportunity to save money that they might otherwise spend running their own intranet sites. As one portal manager said, "They don't have money to spend on intranet development, and we've already done it for them, so why would they sit there and gnash their teeth about how to do it for themselves?"

Problems Solved by Portals

Most intranets have become completely unwieldy and present a highly fragmented and confusing user experience, with no consistency and little navigation support. Portals aim to correct this problem by presenting a single gateway to all corporate information and services.

A consistent look and feel reduces the learning burden and makes it easier for users to recognize where they are and where they can go when navigating a large information space. Also, by integrating services and presenting personalized snippets on the initial screen, intranet portals reduce the need to move around and provide a more efficient environment for job performance.

Ideally, intranet portals would also present a unified security environment and support single sign-on. Unfortunately, almost none of the portals we studied had achieved this goal, even though most help desk calls are about lost passwords, and multiple login barriers are disruptive to users.

Internet vs. Intranet Portals

The entire concept of portals started during the Internet's bubble era, when a range of websites emerged with the goal of being users' one-stop entry to the Web. Names such as AltaVista, Excite, Disney's Go, and Lycos come to mind. Yahoo is still going strong, but most other portals are now either dead or have scaled back their ambitions and focused once again on search.

During 2000 and 2001, several services -- such as Octopus and OnePage -- let users create customized portlet views that were screen-scraped from individual websites. The goal was to spare users from having to visit multiple sites to satiate their information needs: viewing the one portal page would suffice. These services are now dead as well.

Given the miserable track record of Internet portals, why would intranet portals offer any hope? The answer lies in the fundamental difference between the public Web and an internal company intranet: the weaknesses of Internet portals are the strengths of intranet portals.

Business Model

Internet portals relied on advertising, which doesn't work well for informational websites, where users focus on content, not ads. Advertising works very well on search engines because they are the only type of website that people visit in order to find someplace else to go. Portals that originated as search engines have mostly reverted back to them, because that's where the advertising dollars are.

Intranet portals are funded through productivity increases. If they can make employees more efficient at their jobs, the company pockets the gains and can use them to pay for the portal project. (In contrast, an Internet website gains nothing if it helps users complete their tasks faster.) In our interviews with portal teams, we found that many companies did not see productivity as the main argument for portal ROI. Their assumption was that just because employees saved a minute here and a few minutes there didn't mean that the business would make more money. However, I believe that productivity improvements are a true economic benefit. Companies ought to prefer that employees spend time working on business issues, talking to customers, or performing other productive tasks, rather than navigating a confusing intranet or struggling to make a computer do as it's told.

Many teams found the most compelling ROI argument in the elimination of duplicated efforts, both in the technology infrastructure and the work required to place content and services on the intranet. Sprint estimated savings of $15 million per year purely from the hardware, software, and maintenance contracts previously required to run its five biggest intranet sites. Verizon saved 15 million emails per year by making reports available through its portal instead of mailing them to users three times a day. The infrastructure savings are tangible, but I can't help but also mention the productivity gains from reducing users' inbox clutter.

Telephone companies are big and thus report the most impressive numbers, but even smaller companies benefit significantly from eliminating duplicated efforts and streamlining information and services delivery.

User Profile

Internet portals never delivered very targeted information, even on their "my" pages. This is predictable for two reasons. First, a website doesn't know much about its users, and people are reluctant to spend much time customizing their views for reasons of privacy and laziness. Second, a general website is restricted to offering generic services that will appeal to many users. Truly specialized services are rarely available on an Internet portal, despite the fact that people benefit more from services that meet their exact interests than from those that are the same for everybody. World news is less valuable than news about hog farming in Ohio -- if you're a hog farmer in Ohio, that is.

Intranet portals know a lot about their users without having to ask them to perform any additional customization. Role-based personalization uses existing databases to give users information relevant to their job, department, and specific location. For multinational corporations, role-based personalization can also be used to provide country-specific information and to set the preferred language for most users.

It's an open question whether additional personalization is sufficiently useful to be worth the effort. Most of the portal teams we interviewed had decided against it; those portals that did offer individual personalization (beyond role-based personalization) found that most users didn't bother to use it.

Intranet portal applications are much more valuable to users than applications found on Internet portals. For example, mission-critical applications and personal employment benefits are targeted at specific users, and often include elements that they need, rather than those that are just entertaining or nice to have.

Search

Even Yahoo is not entirely self-sufficient, though it does have more internally developed services than any other Internet portal, which may be one of the reasons for its survival. Internet portals depend on the Web for much of their value, especially in regards to their all-important search capability. This is a weakness, because it's hard to effectively integrate external services into a single user interface. Search engines are thus restricted to the impoverished understanding of external sites that can be gleaned from spidering stand-alone pages.

When the same system operates both the search engine and the content, you can substantially improve search performance by using a more structured understanding of the content, proper meta-data, and insights into user behavior and needs based on an analysis of which pages users visit and how they traverse the information space. Also, keyword indexing can be more powerful in domain-specific spaces, such as a bank's portal where the users "speak banking slang," as a portal manager at Credit Suisse put it.

Search has the potential to work substantially better on intranet portals, even though the portal teams we interviewed reported few success stories so far.

Design Process

Internet portals do have a few advantages over intranet portals in terms of their actual development process. Particularly during the boom years, Internet portals had big budgets and big, professional design teams that included dedicated usability professionals to perform frequent user testing. Intranets have always been starved for resources, and portal teams are no exception. They rarely include full-time usability professionals, and the teams we talked to performed much less user testing than they wanted to. Several teams did employ "discount usability methods" and collected valuable user feedback through simplified means. But considering the millions of dollars at stake in terms of employee productivity, much more should be done for portal usability.

Wild Web vs. Single Tool

During the first ten years of their development, intranets were modeled after the Web, with no central control and a profusion of incompatible services. This model has advantages during an experimental phase, but ultimately leads to wasted and duplicated efforts, and results in a substandard user experience.

Intranet portals aim to replace the wild Web model with a tool metaphor, where a company's content and services work together instead of undermining each other. Having a single starting point, a single overview of each user's most important services, a single search, a single navigation scheme and information architecture, and a single set of consistent page design templates all combine to make the intranet portal a more promising corporate information infrastructure.

Full Report

The full research report with case studies of intranet portal design is available for download. (Note: link leads to a newer edition of the report than the one discussed in this article.)