We have a very clear mission-to help organize the world’s information

An interview with Krishna Bharat, creator of the revolutionary Google News.  Here he talks to Sunil Rajguru about the Bangalore R&D center, Google’s mission, mobile search plans and how search engines work.

What is happening in Bangalore?

Bangalore is an R&D center. It’s a peer center to our other offices in California, Seattle, New York, Santa Monica, and Internationally, Zurich and Tokyo. Our vision is that we are going to be a global campus with many offices. Every office is going to offer the same opportunities-which means that we will hire people in India at the same time as in Tokyo or anywhere else. They will work on projects that are relevant locally as well as globally. We spent some fraction of our time on the same kinds of projects that our Google marketing does. Roll up the same product, service or make some fundamental changes in the infrastructure that everyone will benefit from directly. So this includes figuring out how Google’s mission can be most effective in India.

By the way we are not just here for India but South of Asia. We feel our scope is just not in India but the entire neighborhood. It’s like Zurich is responsible for all of Europe. So we want to understand the problems people face in information: it might be performance, availability of content-like, say, why is this content missing or it’s much better developed in Europe or such thing. So we need to understand the issues in India and South Asia, and make appropriate adjustments to products or develop new products. The main challenge lies in the languages and encoding them.

Is Google getting into mobile search?

We are very interested in mobiles because it certainly makes sense in developing countries. There is a lot more of mobile usage here in India than in the US. At the same time the bandwidth and the screen is much smaller and, also, you don’t have the luxury of a keyboard. But not all applications make sense on a desktop. Finding information does make a lot of sense, especially in the sense of location. Like a person may say: Give me information based on where I am. I think part of our charter is to understand this. In India, the penetration of broadband is not as far as in Europe. Even the availability of network accessing again will influence the way that people access Google.

What about Ad Sense?

Google does a very good job of ad-based monetization. In ‘Ad Sense’ you can embed ads that are monetizing the content for you. So we can figure out which ads are most relevant, given the contents, especially for something this technical. When you key in “How to do ‘Wi-Fi’ on the notebook” you’ll get lots and lots of products there to be sold related to Wi-Fi. So ads are very relevant.

There are a lot of companies which believe in Google optimization. How does that work?

Google optimization is a good thing because in a market where there are lots of inexperienced websites, it’s important to have someone come in and say useful things: it has an educational aspect to it. But sometimes they go a little beyond that. They pretend to be a flower site while selling tea. That’s not good. It makes users waste their time. This is called spam. Search engine optimizers sometimes bring in spam to an otherwise spam free area. Search engines don’t like spam because they feel it’s an inconvenience for our users. The way by which Google fights spam is by looking for organic growth from the web. It’s a natural number of times a search engine/search site will put up its name. We have invested a fair amount of effort in research. We have a large team working on it.

It looks as if you’re into practically everything…web search, question search, video search, archives. Where does it all end?
We have a very clear mission-to help organize the world’s information and make it accessible. So, information is the only thing that you may need to know to make decisions. Our role is very clear, we are not trying to hold the information, we are trying to mediate. So we want to be friends with both consumers and producers and we want to save time for the consumers. Like if you’re trying to make a decision and you need information, Google helps in finding that information quickly. Information may come in many forms but the way it is accessed and served up has many things in common.

Is there a limit to how much you can store?

No. The limit is only based on the availability of the events. Google has very large spaces around the world to maintain data. As the requirements come up, we increase the space. We constantly make the process more efficient, make all the failures more manageable, optimize storage and availability of content.

How do search engines search?

We have trained computers to look for information or content given a large amount of data. Historically, web search came out of library science. Library science is a small collection of documents in a library that helps in finding information. But that had limited options and the quality wasn’t that good. This then migrated to the web, with the same parameters. But the web has lots and lots of content with very highly variable quality. But people wanted only say the top ten results. So the stakes are much higher.

We have to understand from a number of signals, which are indicators of quality and relevance. Google has a fine track record of finding these signals for research and integrating them. Our ability is to find the content of the web and make it available very quickly using specialized infrastructure. There are hundreds of searches happening simultaneously. For all searches we need to run through the index rapidly and find all potential cases. So we have to go beyond the superficial signal and move into the linkage and what is the text saying. The other consideration is the language.

People say Google’s beta versions drag on and on. Comments?

That’s just a way of Google saying there are men at work. There are people inside the team and the team is still thinking of ways to make it better. So Google spends a lot of time improving the products it has. We keep rolling out new updates too.

(This interview appeared in Dataquest magazine in June 2005)