Task orientated intranet redesign

May 11, 2006 3:09 AM

Looking to re-develop an intranet around task orientation rather than business structure. Have you done this?

I work on a large-ish intranet (~15,000 documents) and we're looking to change the IA from being orientated around business structure to being task orientated. But where to start?

Have you gone through a similar exercise and, if so, how did you do it? What went well and what would you do differently if you had to do it all again?

posted by TheDonF to Computers & Internet (8 answers total)

I've done this (I'm an information architect) for various intranets (sales, telecom, enterprise billing, etc). Where are you on the project? How long do you have? What budget constraints are you working with? Are you inside or outside the company?

Off the top of my head, these questions are the tip of the iceberg:

1. Who are your users? Different user groups will have different tasks, with or without overlap; differently-prioritized tasks; or order those tasks in a different sequence. Often this breaks down by responsibility level, so you may want to start by separating out users by level and then teasing out differences within each level. I like doing this bottom-up, because the lower echelons often have a better working knowledge of the system and how things really get done, warts and all.

2. Don't forget to talk to the people who train others on the system. Take a day to be "trained," look at help documentation/training manuals, talk to support/helpdesk staff.

3. Make sure you look at the complaint logs and "suggestions box". Alternatively, ask users "what do you hate doing most? why? what if you could offload this?"

4. Don't be the user-experience asshole. Respect the culture of the company and how they get things done. Suggest good UX but don't roll your eyes when they balk at it. Try to visualize more than one way of doing something so you can give your clients a choice (they'll feel empowered rather than as if you are forcing The New on them).

There's so much more, but really, #4 is the one that has gotten me the farthest.
posted by mdiskin at 5:51 AM on May 11, 2006 [2 favorites]

Response by poster: Cheers mdiskin. Project-wise, we're here:

Need to start looking at the IA in the next couple of weeks. Have done a thorough audit of our content and trashed a lot of older, out of date stuff. It needs to launch in the next 4 or 5 months, no budget to speak of and I'm inside the company and have been here for about 5 years so I've got (at least I believe I've got) a fairly good understanding of the business.

Do you mind if I contact you away from this page - I'll feel a bit better about everything not being quite so public.
posted by TheDonF at 7:39 AM on May 11, 2006

I don't want to derail the question or anything, but if you do a web search for these solutions, use the word oriented, because orientated isn't a word.
posted by blue_beetle at 7:39 AM on May 11, 2006 [1 favorite]

Sure thing -- see my profile. Content audits are no fun but they bring lots of demons to light. Good luck.
posted by mdiskin at 8:09 AM on May 11, 2006

I was eager to get in to the thread and talk about a subject I'm actually knowledgeable about, but it looks like mdiskin has already covered the major points. Number four is important even if, like me, you are naturally snarky. One small addition: another reason "to visualize more than one way of doing something" is to give users more than one way of accomplishing their tasks. What may seem to be the obvious way to organize something may not be that intuitive for some people, so you may want to find out the non-obvious way and offer them that path as well.
posted by lackutrol at 8:33 AM on May 11, 2006

If anything comes to light in your discussions that can be safely anonymised, would you mind posting it here? I'm very interested in what mdiskin might have to say.
posted by Pericles at 8:33 AM on May 11, 2006

I am bookmarking this page as well. In the meantime, I am assuming you know about Jakob Nielsen's extensive reviews of intranets, and best practices guide
posted by seawallrunner at 9:24 AM on May 11, 2006

Full disclosure: I used to be that UX a-hole, until I realized that NO ONE DIES if my suggestions aren't used -- I'm not designing cockpits or weapons-systems UIs. My suggestions are good, but if they don't get taken, a company may lose money. That's their issue, not mine. Sometimes the internal realization that you are right is just fine -- there's no need to be the punishing nemesis of the marketing department.

I wholeheartedly agree with lackutrol that multiple paths to a task and its performance are good.

Another thing, which usually crops up in user interviews when people are trying to tell you how to fix something: Try to avoid getting into the morass of deciding on controls (radio buttons, etc) too early in the discussion. Likewise -- avoid larger visual metaphors (folders/tabs)until you've sifted the information enough to realize how people move through it. Nothing's worse than deciding to use a tab metaphor and finding that it only works for one user group, or only under certain conditions, etc...

This is all basic IA, but it seems to be more crucial for intranets, perhaps because there's so much input from the in-house users.
posted by mdiskin at 9:34 AM on May 11, 2006