The Next Generation Search Engine…Isn’t A Search Engine

Wolfram Alpha and Microsoft Bing were both unveiled last month. Each provides a glimpse into the future of how we’ll use the Internet, and shows that Google has been sleeping on the job. Read on to imagine the future of search on the web.

Home Page Evolution

When I started using the web in the early nineties, search engines were in their infancy. The page you marked as your home page was a link directory. Eventually, the link directory that came to dominate was Yahoo.

And then the paradigm shifted. The web became too big to manually index and organize. The only way to access all the information the Internet had to offer was to use a search engine. And so came Alta Vista, and then Google.

Today I use Google for everything. To find links to sites containing information I need to know, to answer simple questions, to price compare and to find funny videos. Yet while Google sometimes provides answers, it usually only provides links to answers. Obtaining the answer requires me to scour other web sites to extract the information I need.

And yet, with Google at it’s peak, the paradigm is shifting once again. Google’s mission is to “organize the world’s information and make it universally accessible and useful”. The assumption is that the information already exists, and Google’s mission is just to find it and organize it. With the launch of Wolfram Alpha and Bing, it’s clear that this is no longer true.

Both Bing and Wolfram Alpha use the user interface metaphor of a search engine. Yet both are clearly working under different assumptions. While Google helps you find information, Bing and Wolfram Alpha help you answer questions. That difference is crucial.

Wolfram Alpha: A Computational Knowledge Engine

Wolfram Alpha bills itself as a computational knowledge engine. It answers questions, not by searching the web, but by retrieving information from its databases. Actually, this is a simplification. Wolfram Alpha actually makes an attempt to understand your question, to parse it as a human would. It then uses information in its databases to answer your question. Answers are displayed not only in text, but charts, diagrams and pictures.

For example, type in “what is the current temperature in asheville?” and you’ll get the current temperature in Asheville, a chart of the temperature for the past seven days and the predicted temperature for the next three days, and the minimum and maximum temperatures for today.

But Wolfram Alpha goes one step even further. If it doesn’t have the answer in its database, it can calculate the answer for you. Type in “population boston / population san francisco” and you’ll get back the population for both Boston and San Francisco, along with a calculation indicating Boston has 75.6% of the population that San Francisco does. You’ll also get a history of the population ratio since 1860 so you can find out that San Francisco exceeded the population of Boston by 1960.

While Wolfram Alpha has only a limited amount of information in its databases today, it’s coverage is still impressive. And as it grows, it will change our expectations of how to answer questions using the Internet.

Bing: A Decision Engine

Bing is Microsoft’s next generation search engine, being billed as a decision engine. It combines traditional search capabilities, with specialized applications and computed knowledge, similar to Wolfram Alpha.

While Bing has many improvements in search, it’s the core paradigm of decision-making that gives Bing its greatest promise. In four key areas, local businesses, shopping, travel and health, Bing provides mini-applications that not only return information, but provide intelligence around that information. Bing does this by providing predictive and calculated information based on your search and other information you provide.

For instance, Bing Travel provides an application for locating and booking flight and hotels. The application appears quite similar to Kayak, with a couple exceptions. Bing Travel not only provides you the price of a flight, it tells you whether that price is expected to go up or down in the coming days. It also provides charts showing how the price would differ if you chose a different flight date, a different length of stay or a different destination.

Similarly, Bing Shopping provides much the same functionality that Google Shopping does, but with a twist. Bing Shopping analyzes the language of reviews and rates the characteristics of a product. So for cameras, it calculates ratings for the speed, resolution and photo quality of each camera model. Clicking on “speed” when viewing a camera model brings back only those reviews related to speed, separated out by positive versus negative reviews. Bing calls this sentiment analysis, and if it works as advertised, it could become extremely powerful.

As a search engine, Bing improves the search experience by analyzing the semantic meaning of your query and attempting to understand your search as a human would. Search results are then broken down into categories dynamically determined based on your search, allowing you to quickly refine your search. This type of clustering isn’t new. Clusty has been doing this for years. But this is the first time we’ve seen this functionality in a mainstream search engine.

Finally, Bing adds lots of small usability features. Similar to Google, Bing provides information related to your search inline with the results. So if you enter a flight number, you’ll get the flight information without having to visit the airline web site. But Bing takes this one step further and provides interactive widgets. So searching for “fedex” brings up a box that let’s you enter your tracking number and track your package directly from Bing. Bing also provides pop-up information boxes for links that provide text excerpts and additional links without leaving the search results page.

The Future of Search

While Bing and Wolfram Alpha provide interesting functionality, I think the best is yet to come. The paradigm shift is apparent, yet not fulfilled.

Imagine a world where search engines don’t merely refer you to information, but collate, analyze and present that information to you in a decision-making tool. Imagine a world where queries are parsed semantically and matched to, not four, but hundreds of mini-applications specific to the query type. Imagine a world where these mini-applications are developed by developers all around the world, just like Facebook applications are developed today, with an open pluggable architecture, allowing the power of this new search engine to expand exponentially.

Imagine the power of Bing, Wolfram Alpha and Google, all in one place, where a question asked instantly receives an answer without needing to read through a bunch of links to find the right one. I, for one, can’t wait.