In a World Filled with Data, What's missing?

Date:   Wednesday , January 02, 2013

Headquartered in Irvine, CA, SRCH2 offers search software built from the ground up to address the needs of a data-driven world.

We live in a data-driven world. New apps, devices and platforms — from social to mobile — have enabled end-users to connect, interact and share. We are building new businesses, and creating new ideas, which would not have been possible even five years ago.

All of this has been enabled by a technology revolution, which has happened largely outside the purview of the average person. In the last five years, the traditional data stack has been revised, piece by piece. Enterprises are tapping unstructured and semi-structured data, in robust environments at ever lower costs, and the benefits have been endless.
Not just technologies, but everyday activities have been upended. Buying a car is not what it used to be. Or dating, or dining, or banking, or watching a movie or basketball game. Data is everywhere.

Yet the data explosion has come at a price. We are overburdened with data, both as enterprises and as humans. In many ways, the data explosion has placed a premium on the role of the editor — an intelligent filters which turns data into something useful.

Surprisingly, the data revolution has put new stress on a very old idea: search. When people think of search, they naturally think of Google. But it’s not just web site search we are talking about. We are talking about every search box you see, whether on a phone, site, or app. We are talking about search boxes on, or TicketMaster, or on your TV. The explosion of textual data (unstructured, or at best, semi-structured) has often meant that you can only access it through a search box. And the software that enables the search box has been far outpaced by development.

While the entire data stack has been rewritten, the one area that is largely unchanged is search. And there is good reason for this. Even the best developers are reluctant to touch the “black box” of search, at an algorithmic level.
Recently, the smart engineers at Twitter took on the task of updating their search capabilities. They toyed with everything, from how new search queries are managed and queued, to how their server loads are balanced. They basically played with and optimized the whole stack. Except for one thing. They did not futz with their search algorithms. That was a black box, which survived the overhaul untouched.
At my company, we have opened that black box.

I go to work every day with search experts from Google and Stanford. PhDs who have quite literally written books on the subject. They have spent the last few years developing new full text search software, from the ground up, to solve the unmet needs of a data-driven world.

Here is an example. Say you are searching on your phone for sushi in Laguna Beach. That search requires a full text index of restaurants and dining categories, overlaid with geo-search. And you do not know the name of any one place. You are searching in a category. This simple task is actually a very complex search. And if you try it, the results you get will be pretty slow, uneven, and probably weak.

Engineers at SRCH2 have rebuilt search, using a hybrid collection of indexing methods which are very different than those commonly used elsewhere. We are able to provide relevant type-forward results, with geo-overlay, instantly. Up to 10x better than today’s most commonly used search software. We are able to offer error-correction with fuzzy search. We are able to do real-time updates, with simultaneous concurrent inbound queries. Search for sushi in Laguna Beach on a site powered by us, and you get highly relevant results, concentrated in your geographical area, every time.

In a data-driven world, good search makes the difference between noise and utility. In a very practical way, good search means you find what you are looking for faster at, or TicketMaster, or on your TV. It means less time spent looking, and more time enjoying.