Socializing Search

I can sing the alphabet backward. I learned it as a child and it sticks with me, just like its more popular cousin, the alphabet song. In his keynote at Enterprise Search Summit West, held Sept. 23–24 in San Jose, Calif., Gene Smith said that social discovery isn’t new. It’s been around a lot longer than I’ve known any alphabet songs—not to mention the web, much less its 2.0 iteration.

Smith referenced Daniel Levitin’s work, including his book The World in Six Songs, which discusses how preliterate cultures used music to transmit information. These people also used music to create social bonds, allowing them to develop commonalities that paved the way for civilizations. We sang to teach and tell tales and then we began to sing together.

Smith extends this logic to say that social discovery—be it gossip, word-of-mouth, blogs, tagging, Twittering—is our natural way of finding things. However, this age-old tactic is not only misconstrued as the latest findability trend, but it’s underappreciated, "because we focus on algorithms, technology, and how many million documents we index on our Google appliance."

At the summit, people gather to share their problems, processes, and successes in enterprise search. Attendees hear real-world experiences in selecting, deploying, and optimizing solutions. Yet the event is not of hyperbole. Buzzwords abound, and those tasked with solving search problems sometimes bear the added burden of a vague corporate mandate to "Get us up to speed on this 2.0 stuff. You know: social search, networks, and tagging."

Without a doubt, some vendors capitalize on this mentality, repositioning tools to focus on "community," "the wisdom of the crowds," or "insert 2.0 buzzword du jour here." In the summit’s closing keynote, Stephen Arnold said he’s seen vendors appear one year selling an enterprise search solution then at the next show they’ve respun the same offering with trendy marketing. He pointed out that many don’t make it to the show a third time, having failed to produce a sustainable business model while frantically following the latest trends.

Arnold also said social search isn’t new, though his reasoning is more prosaic than Smith’s. According to Arnold, "Social ranking isn’t new. Google is inherently a social search engine because it counts clicks."

Popularity, and near ubiquity in the consumer space, has also generated one of the most dreaded phrases managers hear when commissioned to solve enterprise search problems: "We want it to work like Google." What users really want is a clean interface and split-second result times. However, no manager wants employees wading through masses of irrelevant results. Selecting enterprise solutions cannot be reduced to a popularity contest.

Of course Google uses more than popularity to rank its results. The company constantly tweaks its algorithms in an effort to improve its engine. It also periodically adjusts the way results appear—to provide premium placement for advertisers or to make clearer the range of results possible (such as text, images, or maps).

Recently, the company’s profoundly simple interface made a subtle shift: Users may have noticed that when they type in a single keyword, the engine offers to enhance it with a variety of more complex strings. If you type enterprise, Google suggests enterprise car rental, for example. If you type enterprise search, it suggests adding the word tools, or summit, and so on.

According to the company, "Google Suggest uses a combination of many signals, for example the overall popularity of various searches, to help rank the refinements it offers. As the user types into the search box, we then provide suggestions to the user based on these neutral algorithms to help them formulate the query."

While this type of query enhancement isn’t new ( has done it for ages), Google may be the tipping point in bringing it to the masses. As IDC VP Sue Feldman put it in the Day 2 keynote, "One of the problems we have with search is that people ask such lousy questions … anytime tools hand people clues, it helps."

During his endnote, Arnold spent a good deal of his time essentially reducing Google to "the devil you know" status. Google may not be the best solution, according to Arnold, but like it or not, the company is the preeminent leader in search and will likely dominate the enterprise market. He said smart companies need to stop trying to prove their solutions are best and get on the Google-enhancement gravy train. "Google will win. Build solutions that sit on top of Google and make it work better," Arnold commanded the vendors in attendance.

I believe this vibrant market will produce more interesting options than just myriad Google ad-ons. I also think Feldman may well have predicted the future of search in her keynote of the same title. She said that search will become a commodity. "Going forward," according to Feldman, "we will see ubiquitous search; it will be part of larger processes. It will be embedded everywhere to provide interaction." Search, it seems, will be more popular than ever.