Tear Down These Walls!?!?

Euan Semple responded to my last post with his trademark blend of acuity and courtesy, and I’m glad to see that he thinks we’re violently agreeing; I do, too.  His former colleague at the BBC John Howard also responded via a comment to my post, which is worth reproducing:

Andrew I think your worries over walled gardens are over cautious. These sorts of tools (if you choose the right ones!) all produce RSS and are driven through a browser. Data no longer sits in a database hidden behind an opaque data access layer, it’s available as RSS and URLs, and can be linked too. To really make this stuff fly in an organisation you need an aggregation tool to close the loop.

Howard’s comment highlights an excellent question:  what’s the real problem if some E2.0 environments are mutually inaccessible walled gardens?  If, for example, I’m a member of three distinct corporate wikis, each of which is accessible only to its members?  If I work in sales, which has set up an internal ‘blogosphere’ open only to the sales staff, and also for the North American division, which has done the same thing?  After all, as Howard points out, my RSS reader will let me know when anything of interest has changed in any of these environments, and my browser will let me skip among them with no effort at all.  So how big a deal is it that these environments are walled gardens?

The only honest answer is that we don’t really know yet, and maybe this kind of technology Balkanization will turn out to be no big deal within enterprises.  But I can think of two reasons why it might be a problem, or at least sub-optimal.  First, if employees can’t search, link, and tag across all Intranet content, then emergence —  the appearance of high-level patterns and structure as the result of many low-level interactions —  is limited.  

On the Internet if we can see it we can link to it, and that link is useful to everyone thanks to the PageRank algorithm and its cousins.  It’s the same with tagging, thanks to Del.icio.us.  These low level interactions give rise to an elaborate and emergent Web-wide structure. People realize how important these structuring interactions are, as evidenced by the recent controversy over Wikipedia’s decision to make its internal links ‘no follow.’  

If an Intranet consists largely of mutually inaccessible E2.0 environments, how does the cream of the content rise to the top?  How does any employee know where the best stuff is?  Even if the Intranet’s search capabilities were advanced enough to keep track of all employees’ access rights and show them results from all the content they were allowed to see, and only that content, the quality of these results would still be impaired.  It would be impaired because there are by definition fewer low-level interactions (links, tags, etc.) across several walled gardens than within a single one; the whole point of the wall is to keep such interactions from happening. So I don’t see how a Balkanized Intranet can display as much emergence as a single E2.0 platform can.  Is this a big deal?  We’ll have to stay tuned.

The second reason that walled gardens are sub-optimal has to do with an phenomenon that my colleague Karim Lakhani calls ‘broadcast search.’  This is simply the process of asking the world to help you find a solution to your problem.  It’s an old technique; the English government used broadcast search early in the 18th century to solve the problem of finding a ship’s longitude at sea.  Parliament offered the equivalent of about $12 million to anyone who could ‘find longitude.’  The prize was eventually won not by a famed astronomer or scientist of the day, but by the self-taught Yorkshire clockmaker John Harrison.1

As this example illustrates, it’s advantageous to broadcast your search as broadly as possible, because you don’t know in advance who’s got the relevant knowledge to help you solve your problem. So assuming there are no confidentiality issues, why would you post your problem only on the lab wiki when you could post it company-wide?  

And why wouldn’t you do it as visibly as possible so that others could benefit from the knowledge generated? My students this semester have started using our course wiki to ask and answer tech questions.  So far, these range from somewhat geeky —  how do I keep a bot from harvesting email addresses from websites I maintain —  to the fairly basic —  what is RSS, and how do I use it?  So far, every one of them has been answered quickly and well.  What’s more, now everyone who cares to look knows what RSS is and where to download a reader.  So a ‘normal’ search process for answering this question would have made one person smarter; the broadcast search process made as many as 80 people smarter.

None of the above, of course, implies that there are no good reasons to have a walled garden.  Sometimes a team wants to collaborate in private, for very good reasons.  JP Rangaswami, when he was the CIO of DrKW, put in place what I thought was a very smart policy:  he deployed a single, public, bank-wide E2.0 environment, then built private environments by request.  This probably isn’t the universally optimal policy, but it does illustrate that there are ways to address the tensions between the benefits of commonality and the desire for privacy.

How else is this tension being managed?  How is your organization grappling with the issues of E2.0 Balkanization and walled gardens?


1Dava Sobel‘s wonderful book Longitude tells the story of Parliament’s broadcast search and Harrison’s improbable solution.   Harrison’s original timepieces have been restored and are on display and running in the National Maritime Museum outside London in Greenwich.  Many readers of this blog will be enthralled by them.  In addition to being fundamentally important pieces of technology, they are also beautiful.