Last week, right after I released a column on "The Fine Line Between Search and Discovery," three reports came out from Radar Networks, StumbleUpon, and ClipBlast that offer more clues on how search and discovery are converging and diverging.
The Decline of Search?
The most provocative discussion comes from a TechCrunch post, "Is Keyword Search About to Hit Its Breaking Point?" The post features analysis from Nova Spivack, CEO and founder of Radar Networks, which created Twine, featured in Kaila Colbin's column last week. Spivack says that as we move from Web 2.0 (2000-2010) to Web 3.0 and beyond and the volume of data keeps climbing, the productivity of search starts to decline.
The solution, according to Spivack, is to make the data smarter. While TechCrunch's Erick Schonfeld notes that making the data more intelligent can scale, the reliance on search is still going to increase, and the engines themselves will need to keep evolving. It's not an either/or proposition. A further point comes from the white paper Google scientists presented in China last week about new image search technology that doesn't rely on keywords; Google may move beyond keywords to index images, but that will lead consumers to search more for images as the technology becomes more reliable.
Stumbling upon Happiness
TechCrunch's Schonfeld added to his discovery coverage with his report on StumbleUpon registering its five millionth member. StumbleUpon allows users to discover new sites, vote on them, and share them, largely through the use of a toolbar. It's also a favorite of search engine optimization professionals, as it was a hot topic at SMX Social Media last week. StumbleUpon has even spawned imitators, including X-rated content discovery site StumblePorn (the homepage is safe for work, though the rest probably isn't).
While TechCrunch cites comScore metrics saying unique visitors have fluctuated wildly over the past six months, the number of times users have "stumbled" links has continued to steadily climb over the past two years. In the first quarter of 2008, the number of stumbles reached 974 million, 160% more than Q1 2007's 375 million. Soon StumbleUpon will notch its cumulative five billionth stumble.
Schonfeld quotes StumbleUpon founder Garrett Camp's speech at the Next Web conference, where Camp said, "Personalized search is just getting started. I think personalized crawling will start too. Crawlers now are trying to create the biggest map of the web, but implicit filtering and intelligent agents-that is where search and discovery will meet." What's most important here is that search and discovery need to meet.
What you search for is generally different than what you discover. At StumbleUpon's video section, "cats" is one of the most popular video categories. I checked Google Trends, and it shows that searches for "cat videos" rank on par with "horse videos" and consistently surpass searches for "Pauly Shore," but cat video searches trail "chewing gum" and "euthanasia." Personalizing the Web for users based on their cat video viewership isn't going to be that meaningful. But, if you can factor in the stumbled cat videos along with their blog and news readership, their purchase history, and their search behavior, then you accumulate a thorough portrait of that user. Given Google's forays into personalized search and how it lists queries, pages, and videos related to your Web history if you're logged in, Google has a tremendous advantage here.
X Marks the Video
Finally, we turn to the results of a ClipBlast study that enlisted Synovate to survey 1,000 online consumers about how they access video online. Of those who had a preference, "discovery" ranked first (28%) followed by recommendations from friends (27%), then search engines (22%), and then recommendations from people they only know online (10%). The word-of-mouth distinction seems academic, as a recommendation from someone I interact with daily on Twitter holds more weight than one from a college friend I haven't spoken to in years. In light of that, I'd count word of mouth collectively trumping discovery, yet both still handily top search.
This may not be too surprising to long-time readers of this column, as I discussed this in referencing a Digital Hollywood panel last July. I noted, "With video search, searchers may be more valuable than browsers or discoverers, but there won't be as much of a bounty on consumers who search. With video, publishers want consumers to gorge on as much content as possible, and automated discovery is the most efficient way for publishers to hog consumers' attention."
More interesting still is that of the 1,000 people surveyed by ClipBlast, 47% did not express any preference at all for finding videos. In a press release, ClipBlast founder and CEO Gary Baker said, "Even with millions of new clips coming online each day, online video is still in its infancy -- and as the survey shows, habits have yet to be fully formed."
That's the fun of exploring the search and discovery interactions now, when we're still discovering -- or searching for -- what resonates most with consumers.