Discovery: The Anti-Search

This column picks up from a post mentioned on this blog about two weeks back on how this is the Year of Discovery. The idea’s fleshed out more in the MediaPost column below, which continues in the extended entry.

Discovery: The Anti-Search
By David Berkowitz

If some visions for the future of search are realized,
then we’ll be surfing a Web where search doesn’t matter. The emphasis will
instead shift to discovery.

Through discovery, when you read your favorite newspaper online,
you’re presented with a wealth of links from around the Web that should be of
interest to you, including other articles, related books or products, or video
clips, whether or not you’d expect them to be directly relevant. Amazon.com does
this regularly, such as when it told me that customers who bought the Black
& Decker 3.4 PS550B Handsaw also bought a 5-pound bag of Haribo Gummi
Bears and the movie "Borat." It’s through the power of discovery that a site
like Home Depot could suggest, "After busting your butt with this handsaw, why
don’t you unwind with five pounds of Gummi Bears and a satire about a Kazakh
journalist touring America?"

The better discovery gets, the less search should matter, at least
in theory. Discovery is the lazy man’s search, and I mean that in a positive
way. If you can bring someone what they want "serendipitously" (really, it’s
done through sophisticated algorithms and targeting), then you save him the
trouble of ever searching for it.

Here are four examples of how discovery plays a role online and
who’s making it happen:

1) Aggregate Knowledge: This company,
which recently raised a $20 million round of funding, is most directly applying
the Amazon-style recommendation engine for retailers and online publishers.
Aggregate’s CEO, Paul Martino, helped shape my perspective on discovery when I
visited his office last month, and I’m excited to hear more at the Search
Insider Summit next week.

2) WorldBeam: In a recent edition of Forbes
magazine, David Gelernter described his new vision for the Web called the
WorldBeam, which is designed to give every computer user his own personal,
secure access to the Internet and all his personal digital files. One of the
many pieces Gelernter aims to reinvent is search: "Our first versions use an
ordinary keyword search, but we’ve developed intelligent search procedures that
find relevant stream-elements even if they make no explicit mention of the
search terms you typed. We can use on-beam documents themselves to help generate
clusters of words with similar meanings." Combining this with Gelernter’s
description of how people will use WorldBeam to quickly flip through large
volumes of documents, it’s easy to envision a minimized role for search in this
iteration.

3) David Siegel: I met David Siegel recently at
a DoubleClick-hosted luncheon (pre-Google acquisition), and he shared an article
he penned called "Defining Web 3.0" that offers a roadmap for the Semantic Web.
He says that people search far more than they should have to because everything
we want to find is so unstructured. He writes, "Where would I look online for a
mountain bike? Today, there are a hundred sites… I have to find bikes that
match my criteria by searching on keywords — good luck!… Searching by keyword
is full of false positives (finding the wrong thing) and false negatives (what
you’re really looking for doesn’t show up on search results)… Why can’t search
be smart?" It becomes smarter in the age of discovery.

4) Google: This search engine is betting heavily on discovery. On
one level, discovery can be achieved through better ad targeting, where the ads
help consumers discover products and services relevant to them. Then there’s
personalization, which Google applies not just to search but to content
recommendations on users’ personalized homepages. My personal favorite example
comes from a report in MIT’s Technology Review from August 2006, "Googling
Your TV": "[Google’s] prototype software… uses a computer’s built-in
microphone to listen to the sounds in a room. It then filters each five-second
snippet of sound to pick out audio from a TV, reduces the snippet to a digital
‘fingerprint,’ searches an Internet server for a matching fingerprint from a
pre-recorded show, and, if it finds a match, displays ads, chat rooms, or other
information related to that snippet on the user’s computer." That Google
prototype would integrate online discovery with other media, which is how it
could ultimately provide the most value.

Even with the innovation around discovery, dismissing the
importance of search would be foolish for two reasons:

1) Search is an integral part of the online experience, one that
won’t fade overnight. It will also take a while for anyone to improve on
search’s model of targeting content and ads to consumers’ explicit queries.
Contextual and behavioral targeting can be considered forms of  "discovery
marketing," but they only grow the pie of online ad spending instead of eating
into search marketing dollars.

2) Discovery is a form of search, except search here is outsourced
to the back-end technology. In David Siegel’s examples, the volume of searches
keeps rising, except people manually search less while their programmed bots
search more. With Aggregate Knowledge, its algorithms search its partners’
inventory to come up with the best matches. Search and discovery can’t be neatly
separated.

Consider search and discovery as akin to hunting and gathering.
With the former, you’re actively out seeking your prey and you get fed when you
grab it. With the latter, if you wander into a spot that’s fertile enough,
you’ll be able to glean what looks ripe. Protein and produce together keep you
both full and healthy.

Since we need both to survive, expect search to share the
limelight with its counterpart as we find out how much more about discovery we
need to discover.