I enjoyed an engaging presentation by Garrett Camp, founder of StumbleUpon, who was the speaker at this week’s PARC Forum … the first PARC Forum I’ve attended, despite having been working at Nokia Research Center Palo Alto, only about a mile away, for the past 14 months (!).

I’d signed up for StumbleUpon in August of last year, a few weeks before I started at Nokia (when I was working on a white paper on Web 2.0 (in one of the few income-producing activities I engaged in during my entrepreneurial phase)), but I don’t remember using it much … and, indeed, can’t find any evidence that I rated any pages in my StumbleUpon profile.

Garrett offered a number of interesting insights and experiences that rekindled my interest in this web-based tool that facilitates "channelsurfing the web". Rather than regurgitate them all here, I’ll start off with the description from the StumbleUpon web site, and then note a few gems of particular interest (to me).

About StumbleUpon

StumbleUpon helps you discover and share great websites. As you click
Stumble!, we deliver high-quality
pages matched to your personal preferences. These pages have been explicitly recommended
by your friends or one of 3,961,071 other
websurfers with interests similar to you. Rating these sites you like
() automatically shares them with like-minded people
– and helps you discover great sites your friends recommend.

How Does it Work?

StumbleUpon uses /
ratings to form
collaborative opinions on website quality. When you
stumble, you will only see pages which friends and like-minded
stumblers () have recommended. This helps
you discover great content you probably wouldn’t find using a search engine.

Garrett started off distinguishing between search and discovery, noting that search engines (e.g., Google) help you find pages when you know what you want, but recommender systems (e.g., StumbleUpon) help you discover pages you didn’t know you wanted (or, at least, would be interested in) – a form of computer-mediated serendipity.

He shared some impressive statistics, e.g., 4M users, 3B stumbles (web pages rated visited – via the "Stumble!" button – by StumbleUpon users) total, growing by 10M stumbles per day. StumbleUpon users appear to be Internet explorers, visiting 170 domains per month (vs. the general average of 70 domains per month), yielding a total of 13M unique URLs thus far, and adding 25K new URLs per day (5 times more than Digg users … despite the fact that the StumbleUpon system does not have any kind of leaderboard that might provide the kind of extrinsic motivation that may have led to influence peddling among Digg users). The relative absence of gaming incentives may also help explain the relative absence of lurkers: over 90% of people rate recommendations and 50% of people submit new content [though with three orders of magnitude difference between the number of users and the number of stumbles, I wonder how many stumbles are rated, or how many ratings there are (Netflix has over 2 billion ratings from 7M customers, and relative newcomer Flixter has 1 billion ratings)].

Garrett enumerated a few of their recent innovations, e.g.,

• StumbleVideo: "the closest online experience to traditional channel surfing"
• StumbleThru: facilitating discovery within a single domain (e.g., Wikipedia)
• StumbleAds: targeted paid placements displayed between organic stumbles (1 in 20), with a CPV (click per visitor) rate of $0.05 – no click through required. Users can actually rate the ads (using the StumbleUpon toolbar), and the feedback is then offered to advertisers, giving them additional consumer insight. [This seems somewhat less intrusive, or at least less objectionable, than the recently announced social advertising "features" in Facebook.]

He also enumerated a number of lesssons learned, but I’ll restrict my focus to some important and interrelated insights into costs, benefits and incentive schemes. StumbleUpon succeeds, in part, because it leverages one interaction (clicking "I like it") to trigger multiple actions (and thus offers multiple incentives for clicking):

• adds to personal profile
• adds to search reviews page
• implicitly approves the page for distribution to friends and peers (social benefit)
• improves quality of personal recommendation for the user (personal benefit)

Thus, there are 3 reasons to rate:

• recommend to friends / community (social/altruistic motives)
• keep a lightweight blog (memory/self-expression)
• improve recommendations (self-interest)

After the talk, the first thing I did when I next accessed the web was to login again to StumbleUpon. I installed the StumbleUpon toolbar in Firefox and started stumbling. My first click brought me to Squashed Philosophers, which included the following among its philosophical gems of wisdom (from William James):

"If merely ‘feeling good’ could decide, drunkenness would be the supremely valid human experience".

This triggered a vague recollection of Dan Gilbert’s book, Stumbling on Happiness (which, in turn, motivated the title of this post), but I let it pass.

My second click brought me to a page on how to people tick people off. I already have a pretty good idea of how to tick people off (though I like to think I don’t practice this [often]), but I "liked" it anyway.

On my third click, I stumbled upon Amazing Dots … which is part of a collection of optical illusions … which reminded me of one of the many things that fascinated me about the Stumbling on Happiness book (Gilbert included a number of optical illusions in the book). The fourth click brought me to a page describing how to detect lies … which also seemed related (though I’m not sure).

The fifth click brought me to the Essentials of Buddhism, which starts out with the Four Noble Truths:

   1. Suffering exists
   2. Suffering arises from attachment to desires
   3. Suffering ceases when attachment to desire ceases
   4. Freedom from suffering is possible by practicing the Eightfold Path

I had to stop, because this was getting too serendipitous for comfort – late last night, I sent one of my colleagues an excerpt from the blog post I’d written when I was joining Nokia, in which I’d referenced Stumbling on Happiness, and then further noted:

In yet another odd instance of irony, preaching what I want to practice, reading what I’ve written and/or being reminded of what I didn’t know I knew, I read further in the post:

I’m not sure what to expect, so it’s relatively easy to let go of expectations — and embrace working without attachments.
NRCPA is very young, and I believe that the experimentation and
innovation in our research will be complimented by experimentation and
innovation in our model(s) of research.  In fact, this multidimensional openness is part of the appeal for me.

Speaking of attachments, I’m reminded of Steve Pavlina’s post on self-acceptance vs. personal growth (or, perhaps more properly, self-acceptance vs. positional growth):

The
underlying problem is that by rooting your sense of self in something
that will fluctuate, like the current position of any measurable part
of your life, you’re going to suffer in one way or another.
…
Instead
of rooting your sense of self in your position, which is changeable,
what would happen if you rooted your sense of self in something
permanent and unchangeable?  Stop identifying yourself with any form of
positional status, and pick something invulnerable instead… like a pure
concept that nothing in this world can touch.  Examples include
unconditional love, service to humanity, faith in a higher power,
compassion, nonviolence, and so on.

[Other examples may include helping people relate … or connecting people.]

I’m further reminded that the root of passion is suffering … and that (for Buddhists, at least) the root of suffering is attachment.

I think there’s another whole blog post I could write about the idea that blogging (for me) is often a serendipitous re-discovery process of reading what I’ve written … and being reminded of what I didn’t know I knew. But I’ll leave that for another post … and leave StumbleUpon for another time when I"m looking for some additional computer-mediated serendipity.