Summary: An analysis of intranet portals found slimmer information architectures and a renewed emphasis on fresh content and useful applications. Past findings, including those on role-based personalization, were confirmed.
It's been almost three years since our last project assessing the usability of intranet portals . This is a good time to take a fresh look at portal design, since three years is also the average time between intranet redesigns .
To find out what's working now, we analyzed a new series of portal projects and compared them with the conclusions from the previous round.
In the larger scheme of things, three years is not much time, so it's not surprising that all of the findings from our initial portal project still hold. In fact, none of the forty-five best practices documented in the report's first edition have changed . Yes, we've gained many new insights, but what was good three years ago continues to be good today.
In particular, the new project confirmed the following findings from the previous project:
- Portal solutions still don't offer satisfactory usability out of the box. This is more of a disgrace now than it was in the past, because we now know so much more about intranet usability. Vendors need to integrate this knowledge into their software.
- Single sign-on is still more a dream than a reality. It's one of the most desired portal features and creates huge savings in help-desk calls, but most companies are not yet there. Users still must log in again and again. Multiple sign-on does offer one usability benefit, however: it can help employees feel more comfortable about information privacy when accessing highly sensitive data.
- Personalization for individual users is still rare. Organizations continue to find role-based personalization more useful and to use it more frequently. For example, some companies present certain information or portlets only to people with a particular job title or people who work in a particular location.
- Governance has always been more important to portal success than technical issues, and this finding was even stronger in the new study. One popular approach is to create a steering group representing various business areas. Projects also need to establish firm rules for enforcing design consistency and migrating content and applications into the portal.
- ROI is woefully under-documented . Too few portal projects collect good productivity metrics, though some companies are now beginning to measure themselves against our intranet testing report's time-on-task benchmarks and using this data to compute their savings relative to average intranets. More typically, portal projects measure user satisfaction and usage. For example, Fujitsu Siemens Computers in Germany found that its intranet use tripled after the portal went online. (Doubled use is more common across the projects we've studied.) Since intranet use is completely voluntary, increased use is a strong indication that a portal helps employees do their work, though it's still an indirect metric.
First, portals are becoming too popular for their own good. Some companies have multiple "portals" because many software vendors include portals with their products to make them seem more important and to get a shot at being on top. This obviously doesn't work, because an enterprise portal by definition must be the one entry point to the intranet and all of its content and applications. So, you can only have one portal. Everything else must be subservient to it for the concept to work.
Second, as time passes, the amount of stale portal content grows. Portal managers have therefore started explicit efforts to combat outdated information. The Defense Finance and Accounting Service's Project CleanSweep audits pages across its portal and tells owners of outdated content that it will be removed by a certain date if it's not refreshed. Other portals use their content management system (CMS) to track expected revision or expiration dates for all pages, and authors must estimate these dates whenever they create or update content.
Third, because much portal software continues to be overly complex, portal managers have recognized a need for training in the correct use of portal features. For example, one big company found that 26% of pages contained miscategorized metadata. You can't just assume that people will understand a complex concept like metadata without instruction.
Fourth, user acceptance requires explicit attention. A big company (that prefers to remain anonymous) had its portal project fail completely. Performance was miserably slow, usability was poor, and people simply refused to use the portal--to the extent that they used the old intranet pages more than eight times as much as the new portal. The portal project had been in vain; to save the intranet, the portal was removed after less than a year.
Portal projects are more likely to have a happier outcome if they emphasize user needs from the beginning, and also employ various tricks to enhance user acceptance. For example, we've begun to see good use of multimedia and video clips on portal homepages; even something seemingly frivolous like a user poll can help make the portal more fun and appealing.
Fifth, information architecture (IA) is becoming streamlined. For example, Sprint has reduced the number of tabs on its portal from eight to five due to a corresponding reduction in the number of major categories. As we discover more about how to build a good intranet user experience, portal teams can take advantage of accumulated usability findings across companies and emphasize the most useful areas.
Portlets on the portal homepage follow a similar trend: fewer, and more useful. In the past, portal projects typically offered users portlets simply because they could -- even though organizations generally didn't provide enough content or features to justify the portlets' prominent placement. In another design simplification, offering two layers of tabs seems to be a thing of the past. Good riddance.
Finally, we're seeing a shift in how organizations use portals. Where portals were once pure information aggregators, organizations are increasingly using them as application organizers . Many companies now offer a variety of killer apps and are moving more of them into the ports. This also helps with user acceptance, because a tools-oriented approach to the portal makes it more obviously useful.
All in all, intranet portals are definitely maturing. The user interfaces are simpler, and teams are getting better at managing both the portal itself and its relationship with other departments and content providers. We definitely see a continuing need for better ROI measures . Nonetheless, many portal projects achieve significant improvements — which are almost as easy to gauge as the spectacular failures.
The full research report with case studies of intranet portal design is available for download.