Thinking about Complex Searches

Google and Yahoo! are part of a thriving industry that provides billions of dollars in real value. Yet in The Search John Batelle and Udi Manber suggest that Search is only about 5% invented. Bill Gross, inventor of the Pay-Per-Click business model (among other accomplishments), follows that thinking, saying in an interview with John Battelle:

I do think pure search sites will continue to prosper, but I also think that there will be many new kinds of specialized search that continue to surprise us. I just made up a little table of the searches I do per day over the last 20 years, looking at some key milestones, like when I started using email heavily, and then when Netscape took off, and then when the first search engine companies went public, and then again when new tools came out, like X1 for searching email, iTunes for searching music, my TomTom for searching for locations.

In other words, current Search solutions do ok, but there is lots of room for innovation. In February, I started working on SwitchBook Software to address a core problem that seems to have been left out of the conversation, called “Complex Search.” (My apologies… the website doesn’t have much there at the time I wrote this blog entry.)

In a sentence, Complex Searches require more than a single result to resolve.

Simple Searches—like finding how to replace the toner in your printer—can resolve quickly once an appropriate online resource is found. When you find the instructions, you are done Searching.

In contrast, Complex Searches—like finding a job or buying a home—require multiple queries, using multiple Search Providers, to discover various aspects of the “result” at multiple final destinations, often Searching for fundamentally different things at different parts of the Search.

For example, when Searching for a home one might begin by asking “What homes are available that fit my criteria?” Then, for each candidate home you find: Is it a good price? What’s crime like in that neighborhood? What kind of financing can I get? What are the schools like? Complex Searches require answering complicated, evolving questions by exploring the web through various Search Providers and websites. Current tools make this difficult.

As Bill Gross’s comments suggest, users today will use many different Search engines to resolve these types of questions, visiting dozens or even hundreds of individual websites over the course of weeks or even months. With today’s tools, this is a pain.

What we need is a tool that helps users keep track of their Search across their entire interaction with the web. Something that not only logs their Attention, it also provides an interface for expressing Intent, transforming it into better results for their Search. It should be nearly invisible, unobtrusively augmenting current Search behavior. It should help sellers as much as buyers. It should both scale well and improve with widespread use. And it should be capable of extensive user customization and open integration with existing search silos, so power-users can surprise us with new functionality and Expedia, CheapTickets, and United.com can not only interoperate, they can individually create new forms of value by providing their own customization.

It should be something like the AttentionTrust‘s GestureBank, but with more focus, structure, and explicit user direction. It should be something like Doc Searl‘s Vendor Relationship Management, but perhaps with more focus on the Search before users are committed to a purchase. It should work seamlessly with Microformats, OpenID, and OpenURL to link user-driven activities to the semantic web through authenticated but privacy-supporting mechanisms. And perhaps most importantly, it should make the most of open standards and open source to reach as many people as quickly as possible.

So, that’s why I’m here. Drop me a line if this is a topic of your interest.