Rivers, Not Trees: The Challenge to Organizational IA

I feel like the emphasis of organizational IA (intranets especially) has been misplaced for a long time. It’s stuck back managing trees, when it should be managing rivers.

Let’s back up a second –

Information has a logical shape. We discussed this in a post about the concept of content geographies last year. Individual units of information – “pages,” lets say – connect to each other in such a way that they form a structure.

For most public websites, this structure is a tree. You have a home page, and a series of top-level pages, then secondary pages, and so on that spread out below. This is such a common metaphor that most web content management systems have baked in support for it. They either have a content tree, or they have a menuing system in which you arbitrarily organize pages into such a tree. (Well, an upside-down tree, in most representations, assuming the visitor moves “down,” or “deeper” into the tree.)

The tree metaphor works for organizing a repository of static information. People can relate to it. It’s a natural way for visitors to narrow down information, moving from broader concept to narrower concept until they find what they’re explicitly looking for.

However, getting back to organization communications (anything that serves The Indoctrinated Audience), this metaphor only fits part of your content. And it’s probably a smaller part than you think.

To explain, let’s talk about Donald Rumsfeld for a minute.

Mr. Rumsfeld was the U.S. Secretary of Defense during the opening years of the Iraq War. Here’s a famous quote he made at one point during the early stages of the war:

“[…] there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns – the ones we don’t know we don’t know.”

The press jumped on this as incomprehensible double-talk, but Rumsfeld was unrepentant (in fact, he wrote his memoirs last year, and titled the book “Known and Unknown”). I’m on Rumsfeld’s side here – his point was clear.

In essence, there are three types of information:

• Known Knowns: This is information that you know exists, and that you believe know the value of. I know that there is a periodic table of elements, and that the abbreviation for gold is “Au.” (Quick mnemonic: a famous James Bond villain was AUric GOLDfinger.)

• Known Unknowns: This is information that you know exists, but that you do not know the value of. I know that there is an abbreviation for the element Boron. This information exists somewhere, but I do not know the value of it. I would have to go looking for this, but the important fact is that I know it’s out there somewhere.

• Unknown Unknowns: This is information that you don’t know exists, and (obviously) don’t know the value of. My eight-year-old daughter doesn’t know the abbreviation for Boron either. The difference between her and I is that she doesn’t even know that this piece of information exists. It’s a crucial difference.

It’s this last one that’s really dangerous. You don’t even know what you don’t know.

How does this relate to intranets? My experience tells me that Unknown Unknowns constitute the majority of the information an intranet will convey.

Let’s tie this back to user behavior for a minute. In the Polar Bear Book, Louis Rosenfeld elucidated several types of user-searching behaviors. One of these was “Known Item Searching.” This is information that a user knows exists, they just have to find it.

Does this sound familiar? It should – it’s a Known Unknown.

• I know the chemical symbol for Boron exists, I just need to find it.

• I know that my company has a dress code. I’ve even read it at one point. But I can’t remember if I can wear a pink necktie, so I need to go reference this piece of information.

• I know that Client XYZ has a profile on the intranet. I haven’t seen it, but Bob told me it’s out there.

These are valid information seeking tasks, but they’re all predicated on me knowing that the information exists. Knowing this, I’ll go search some tree of data to find it, digging down through layers of pages in my intranet to locate it.

Now, consider some Unknown Unknowns:

• I don’t know that the north parking lot is going to be resurfaced tomorrow and that I shouldn’t park there.

• I don’t know that an analyst just mentioned us negatively in an industry report.

• I don’t know that Mary over in Accounting just gave birth to a bouncing baby boy named Raymond and won’t be in the office for 10 weeks.

• I don’t know that Congress is discussing a legislative change that affects our industry in a very, very big way.

• I don’t know that the CEO wants to tell me all about “vision” and “synergies” and that I should read this thing he wrote because he’s going to ask me about it at the Christmas Party next week and I don’t want to look like a stooge.

How could I go searching for any of this information? I can’t, because I don’t know the information exists. In every case, I would eventually find out, but that would be sub-optimal. I would eventually find out about the bus bearing down on me at 50 m.p.h. too, but the goal would to find out early enough to step out of the way.

And what is Rosenfeld’s label for this type information seeking behaviour? He doesn’t have one, because you can’t look for what you don’t know is there.

How do I get this information? News.

You know, news – the ubiquitous “announcements” or “latest news” widgets. This is the information that starts at the top of the home page and scrolls down, latest information on the top, older information dropping off the bottom. News is usually time-sensitive – it’s important that I know the parking lot is being re-surfaced before it happens, but after it’s done, it’s not relevant.

News is timely. It is temporal. It’s not rooted in one place (you know, like a tree). Rather, it rushes past you (you know, like a river).

Employees search for Known Unknowns. Employees need to be force-fed Unknown Unknowns. While employees will gladly go climb a tree to find something they know is there, they need to be dropped into a river of news in order to find out the things they don’t know.

I maintain that the majority of information imparted by your intranet will be river information, not tree information. In every intranet I’ve built or used, the most prolific source of information was the stream of news, usually on the home page. Sure, there was reference information deeper down, but how often did I need to go find that? (And when I did need to look, did I browse though the navigation, or did I just search? Did I climb the tree, or just chop it down to my level? But, that’s an entirely different conversation…)

I checked in on the analytics for one of my clients’ intranets. Sure enough, content down the “news” URL branch was accessed 150% more often than all the content down the “departments” branch (even more significant when you realize that 90% of the “news” accesses were items from just the last 30 days). Additionally, there had been vastly more content added to “news” than had been added to – or edited in – the “departments” content.

(Postscript: I emailed this article to that client. Their take: “The news is where the info is at. All the other pages are there just in case.”)

Sadly, intranet IA is often fixated on tree content to the exclusion of river content. We spend so much time organizing it and obsessing over it. Nielsen Normal Group will sell you a $400 report all about it. Then you will no-doubt draw a big architecture diagram showing how all the branches fit together. Then you can do reverse-tree testing to make sure your tree is climbable. And then you can implement a great search system so people can ignore all of that.

But I promise you that your users are going to get more value and depend more fully on news. Not just for information, but for a sense of organizational identity.

Last year, I read Mitchell Stevens “A History of News.” In it, he discusses the use of news as a force for societal unity, and why news simply must exist for a group of people to identify or bond together.

Societies depend for their unity and coherence on a sense of group identity. […] a society too depends on the flow of a stream pf perceptions and sentiments from a shared perspective – in this case a societal perspective – to provide its members with day to day, minute to minute, reminders of the existence and the significance of the group. To think a society’s thoughts is to belong to that society. News provides the requisite set of shared thoughts.

News is shared experience. It informs, sure, yet it also binds. The news that a group shares is a continuing current or channel of unified concern. It’s the lifeblood of the day-to-day existence of the collective.

I’m not saying tree content isn’t important, but can your corporate dress code do that?

The content portion of intranets should fundamentally be about rivers and enabling employees to manage the information running past them. Rather than organize static information into trees, let’s determine what rivers exists and give employees the tools to find them, filter them, combine them, share them, get notified about them, and push information into them.

Too many times, I’ve seen a skyscraper’s worth of architecture and analysis put into a vast site map and information architecture, and then heard something like, “and we’ll have some announcements on the home page too.” Often, they don’t even verbalize this – it exists only as some box on the home page wireframe because someone saw it on someone else’s intranet and figured they should have it too.

Just like trees can’t survive without water, it’s a continual influx of fresh information that’s the backbone of your organization and that which most intranet projects spend the least amount of time analyzing and planning.

Just once, I’d like to see an intranet project start with the news feed, and have everything else line up behind that.