Arc90 Mobile NYT Reader | Home | RIA's & Revolutionary Design

Filed under Thoughts on January 18, 2007 by Tim Meaney

A Two-Way Conversation

[This is part one in a two post series on metadata]

It is widely regarded that the introduction of Google Maps was the tipping point for AJAX, or more generally, for Web-applications employing rich, desktop-like, interactions. Recently, Google introduced personal usage statistics and graphs to Google Reader (RSS / feed-reader) via Reader Trends, and after spending some time with it, I'm wondering if this will be regarded as the tipping point for the reporting of personal usage metadata.

Before elaborating on that though, I need to first admit something - I'm obsessed with my usage metadata. I love when applications give me information about my interactions, and wish that every application would be so thoughtful. I'd like to know which word I most often spell incorrectly (has to be thourough or seperate), would be interested to see a graph of the hours of the day and the quantity of email sent in each hour, or a simple report for the previous year of the how often I work on my laptop on the weekends. I even enjoy tracking the play-counts for my songs in iTunes, but that's another story altogether. I may be an extreme example, but personal usage metadata is a compelling and largely unexplored area, full of potential value in both personal and business applications.

Back when I was a kid, applications were an island, and communication across islands was difficult (as true usage tracking occurs at the person-level, not machine-level -- if you use Outlook or iTunes at home versus in the office, your true usage reporting would span across both.) Pre-Internet applications, however, could have tracked machine-specific personal usage data and avoided the three-hour tour of cross-island communication, but in that era of personal computing, the execution of tasks was paramount and programmers were kept busy long into the night tackling cognition tasks, meta-cognition just wasn't important enough. One can imagine the product team for Lotus 1-2-3 working to implement new features and functions, and might suppose that they would have had a tough time fighting for a graph of how many times you used VLOOKUP this year. Or on which days of the week you are most likely to SUM.

Going forward a bit, you can now notice some applications tracking your usage characteristics, even if they haven't yet taken the step of exposing this data back to you. Take iTunes for example, at minimum your play-count across devices is being tracked and synched back to iTunes (however I don't think it's tracking the date and time for each play, yet that is.)

Back to Google Reader and the new Trends features - here's a summary of the usage data they are reporting:

* A simple chart of what hours of the day I catch-up on reading
* Another simple chart of the days of the week with reading frequency
* Total # and % read by blog
* Items per day and % read of daily total by blog
* # Starred or Shared by blog

That's it - simple reporting of a blog's activity, my general usage, and the intersection of the two (how my usage relates to a blog's activity). This information is more than cute or interesting, it is useful. I can use this information to cut loose inactive (or worse, active sources that I regularly ignore) sources of information, or to provide insight into how I'm spending my days of the week or allocating time of the day. A personal goal of mine this year is to incorporate a period of "learning" after having just built software for 15 months straight and barely surfacing for air. Trends is a great way for me to see if I'm on track for that goal. Noticing that a vast majority of my reading continues to occur after midnight, for example, may prove that it's more of the same.

So regarding the tipping point comment, in conversations with others it seems to me that this may capture the attention of the technologists of the world about the compelling nature of usage data. I therefore should qualify my statement to say it may tip it only for technologists in the near-term, but how long will it take from that point for this to go mainstream? Once those building applications are hooked, you'll start to see this trend in more mainstream contexts (feed reading still only touches a small % of Internet users.) Software will then become a two-way conversation, beyond a mere execution of tasks.

I'll end with a prediction - iTunes will incorporate usage reporting in the near-term. Can't you already picture the graph of your favorite genres by month? Charts of number of plays among your top 10 artists (inferred by average song rating) by sliding timescale... Artists you burned-out on the fastest... Your favorite and most played songs (or artists / genres) in your twenties (or thirties / fourties.) Artists you've maintained interest in the longest... The % of total songs played that is from your top 10 or top 50 most played songs. Outliers such as highly-rated songs that you barely ever listen to, or low-rated songs that you play often...

C'mon, we all have them! I get to blame them on my wife though.

[Part two will cover aggregated metadata and describe the ways usage data across the masses, along with more overt forms of metadata, will be used to assist decision-making and influence behavior.]

Post a Comment Digg Del.icio.us

Comments

Post a Comment:

Name:

Email Address:

URL:

Comment:

Arc90 Mobile NYT Reader | Main | RIA's & Revolutionary Design