Shared navigation of online bureaucracies

In my talk on Friday at the GOVIS (government information systems) conference in Wellington, I wasn’t the only one to suggest that web 2.0 attitudes will change the relationship between governments and citizens. That notion now seems to be pretty firmly established, and the question is not whether citizens will collaborate with their governments, but rather how.

Among other developments, I think we’ll soon see a refreshing new approach to the consumption of government services. A couple of weeks ago at Berkeley’s school of information I met Anna Kartavenko, one of Bob Glushko’s graduate students, She’s working on ways to make the byzantine California regulatory apparatus more accessible. If you’re starting a business in that state, it’s really hard to figure out which licenses you need to apply for, as well as how (and in what order) to apply for them.

The problem is universal, of course, and folks at GOVIS were wrestling with it too. When you’re providing the information systems that both document and implement government services, you certainly want to do everything right in terms of system and information architecture. But I suspect there’s about to be a new force in the world that will work toward the same ends — easy discovery and effective use of services — by very different means. That force is shared experiential knowledge.

Yes, search should give the right answer, and the systems that search points you to should work well. No, these things don’t always happen. But even if they do, you’d still like to plug into somebody who’s been down the same path you are traveling. A formal description of a procedure is never enough. If possible, we’d always like to hear from somebody who’s been there, done that, knows the drill, and can point out the pitfalls.

What we loosely call social media are beginning to create that possibility. For a variety of reasons, people are beginning to document and share what they know. If you write it down, you’ll be able to remember it yourself in case you have to replay the steps. And writing it down in a shared information system in the cloud is becoming a more reliable way to assure your own future access to this documentation than writing it down locally.

To the extent your knowledge is a source of competitive advantage, you’ll want to be cautious about how much of it you publish. But then again, the reputation you establish by publishing some of your knowledge may lead to new opportunities to use that knowledge for your own gain.

Along with these incentives, which I classify as examples of enlightened self interest, there are also purely altruistic motives, and I don’t discount those. But let’s just stick with enlightened self interest for now. Given those incentives to share knowledge, how can we lower the activation threshold for sharing?

I think one answer will emerge from the intersection of social bookmarking and clickstream logging. Suppose that instead of bookmarking and tagging a single URL, you could bookmark and tag a sequence of page-visiting and form-filling events. The sequence corresponds to some complex multi-step task. The performance of the task crosses several (or many) online jurisdictions. The outcome might be successful or not: “Yes, I got the license,” or “No I didn’t.” But in either case, it would be qualified by an anecdotal report: “Yes, I got the license, but I found out that if you’re in my category you need an import license and you have to meet the following insurance requirement.”

You couldn’t reasonably expect very many people to reflect on their encounters with online bureaucracy and take time to write reports like that. But what if it were a much more lightweight activity, like the difference between writing a blog entry and tossing off a bookmark or a Twitter message? Then participation becomes much more likely.

The key ingredient here is identifying a sequence of events in the browser (or rich client), and enabling people to visualize and then categorize and describe that sequence. And that seems eminently doable.


  1. re: “The key ingredient here is identifying a sequence of events in the browser (or rich client), and enabling people to visualize and then categorize and describe that sequence. And that seems eminently doable.”

    It is being done.

    You need to have a chat with Tessa Lau and company at IBM research and ask her about Koala (there’s a little screencast and a pdf: Koala: Capture, Share, Automate, Personalize Business Processes on the Web)

    The ingredients are all there: a firefox extension, sidebar UI, capture sequence of repetitive tasks in browser, pipe, filter and customize, store resultant script, optionally share said script tagging it for later discovery by community, promote and foster an ecosystem of popular scripts, watch how it is used.

    The initial users will be folks wanting recipes for the boring bureaucratic tasks but this is a very general framework.

    Let’s just say this: I’ve actually ordered business cards using a Koala script braving Big Blue’s ordering process.

    The hard problems will still be there: for certain processes you need to deal with authentication. You need to factor out personal credentials so that the scripts can be reused by others…

    Anyway Tessa has her hands full but its a very promising start. This shouldn’t stay as research for long; Real People (TM) are using it.

    There’s another aspect that bears pondering. A lot of these processes are simply form applications (sometimes multi-step). What would happen if you didn’t have to spend so much time inferring structure e.g. navigate to the 3rd text field on second on page? What if the form’s structure was explicitly externalized? Such things would be quite easy to pull together. Forms sadly continue to be neglected on the web. For now we have the messiness of deciphering intent and programming by demonstration is fair starting point.

    It’s funny, it’s all about providing the same empowerment of the Excel or Applescript user for the web user. I think we have the data exchange in hand, we have the transport the transfer and the data definition. We have uris, we have feeds and even publish protocols.

    What remains are the challenges of the glue layer:
    1. how to effectively wire these things together
    2. how the social issues are addressed

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s