I wrote an essay back in 2006, after reading John Battelle’s book, The Search, and thought I’d revist it and take a look at how far we’ve come, and how far we still need to go. All that said, I’m still an idiot for not buying GOOG back in the day…
The unrelenting rise of search (and with it, Google’s share price) as chronicled in John Battelle’s excellent book, The Search, is the topic of much discussion. That said, Search still has a long way to go, and there is plenty of room for improvement, and I think there are real opportunities to unseat Google’s top spot by providing strong offerings in the 3 C’s outlined below.
Without a doubt one of the hardest things is to know the intent of what someone is looking for — while I get Battelle’s “database of intentions” concept — I just don’t think a text field can truly understand the context of what I’m looking for in an independent trial. That is, if I type in “dogs” what am I really looking for? to buy a dog? Dog allergies? Dog food? The Pets.com hand puppet dog? Natural language search engines (like AskJeeves) tried to infer context by enabling a consumer to type actual questions, and Yahoo’s original “curated” approach to the web enabled hierarchical category browsing; however, it’s hard to know what someone truly means based on keywords in a text entry field. Some approaches may include reworking the UI to include some basic parameters on the search to say it’s a “shopping query” or a “food query”, or other basic, global categories. I’m not sure this goes all the way, but clearly there is a better way to accomplish this.
One major issue with search engines is the time delay for new content to appear in the results. This is actually runs exactly counter to Google’s PageRank system. Since the number of incoming links is a big driver of relevance, something new will no doubt require more time to be viewed, and have others link to it. Thus, it would seem that more timely content is in effect penalized. Google News certainly helps to address this issue, but indexing the Associated Press wire barely scratches the richness of content out there.
[RD Update: This is where some of the bigget noise has been made recently with Facebook, Twitter, and the “real-time” web. Google itself has invested heavily into Google Caffeine]
The driving force of the Web 2.0 onslaught is its ability to enable rapid and relevant communication between groups of people, or communities. While this concept isn’t necessarily new (Geocities anyone?), social networking sites, like MySpace, Wikipedia, and LinkedIn, have made it easier for groups to interact in meaningful (and frivolous) ways. How can the classical search engine take advantage of this “wisdom of crowds”, as well as social recommendations. While social bookmarking has approached this from another perspective, and one can argue that PageRank is the ultimate crowdsourced tool, there seems to be something lacking.
[RD Update: Once again, Facebook and Twitter seem to be the leaders here, while Google did SearchWiki and Knol. Clearly there is more to happen to bring these all together into a more consolidated interface.]
Observability and Switching Costs create vulnerability
So, one could argue that now that “Google” is a common verb in every day life it has cemented its leadership position. However, it is very easy for people to use something else – we are always one click away from another site, that we can easily try, observe its result, and refresh our bookmarks should we so choose. If Search Engine X emerged today with the capabilities outlined above, why stay with Google? Personalization and customization certainly help (iGoogle), but with all the mashup tools out there, it’s not enough. The ideal solution would become more personal to me over time, more relevant, and offer me benefits in related areas over time.
Now of course, the folks at Google are much smarter than I am, so without a doubt none of these things has gone unnoticed. It will be exiting to see how this all shapes up, and what we, as consumers, can look forward to.
[RD Update: I had two other C”s to represent “cross-media”, the ability to search more than hypertext, and “cross-channel”, meaning access from desktop, mobile, and other devices. However, it seems that those areas are less relevant today, as they are very obvious]