Skip to content →

The Power of Track: Top Down, Bottom Up

Pyramid with eye on the dollar bill

Any discussion of ‘Track’ seems to require a brief definition. By now we’re familiar with ‘Search,’ it’s the process of examining everything in an index to find keyword matches– and then ranking the results in the most useful order. ‘Track’ is a filter on the firehose of information generated by the microblogging space. It’s stuff that’s not in the search engine index yet, it’s what people are saying and thinking right now. The ideal time parameters around track are the rhythms of the conversation. Tracked results should reach one quickly enough to be able to respond to a question or statement without the thread of the conversation being lost. Track expands the public instant message dialogue beyond a personal directed social graph. When I track a topic, I’ll see messages from anyone who uses the words I’m tracking.

Our recent Presidential campaign featured diametrically opposed approaches to organization. One campaign employed a bottom-up strategy; the other preferred the top-down approach. This got me thinking about track and the difference between top-down and bottom-up strategies. A top-down strategy is easier to support in that the tracked keywords are very limited and they don’t change unless consciously changed. The creation of an intentional scarcity of tracked keywords also lays the foundation for an economic model.

Here are two approaches to a limited track keyword set:

A bottom-up strategy is harder to support, every user in the system could have a separate list of keyword filters for the full stream. That’s a lot of computing that needs to be done very quickly. As events unfold around us in the world, each user’s set of keywords would change to aid in the discovery around the new topics emerging. The full set of track filters would be constantly shifting and morphing.

Search uses a top-down strategy to prioritize keyword results. You probably want to see the results that most people want to see or think are important. The brilliance of the modern search engine is that it makes that calculation for everything it indexes. It arrives at top-down through bottom-up voting on importance. Services like Mahalo attempt to only provide extended results for popular searches– converging to the top of the pyramid.

Here are some examples of services trying to create lists of what’s popular:

If you track what’s popular, depending on the community providing the attention data, you quickly converge on the least interesting topics.  You’re only seeing what everybody knows.

And the broader the community, the more quickly you end up with The National Enquirer or the World Weekly News. Now that the election is over, take a look at the trending topics on Occasionally you’ll see a micro-community poke through with a hashtag for a particular conference or event, but mostly there’s little of interest.

New York Public Library Reading Room

Each of us is a unique nexus of connections tracing our path through the world. Our individual curiosity provides the fuel, the focus, the set of keywords we carry with us to make sense of, and create, the future. The difference between all of us having the power to track the stream and only the most popular keywords being tracked is the difference between all the books in the world and the magazines available at an airport bookstore. For a short period of time we all had track, now track is the privilege of the few.

For some reason, I’m reminded of a song by Laurie Anderson for tape-bow violin based on a quote by Lenin:

Published in economics politics real time web social graph zettel