Security and Storage are the current buzzwords in the world of IT. Add a
third S to them. Search is emerging as the next killer app. Over the next couple
of years it is going to get many different manifestations and explode as a
revenue stream for the IT industry. One reason is what I wrote about a few weeks
back. Search is going to be paid for by the users and the advertisers on
search-related services. Another reason is that it is taking on many more
manifestations—mostly related to intelligence. So to put it more accurately,
Intelligent Search is the next big wave. The power of current search engines and
associated services is awesome. But it is still like doing surgery with hammers
and chisels. The need is for finer tools, which are just about developing. Once
these fall in place, watch the fireworks. There will be hundreds of companies
whose primary business will be to provide search-related services.
At the base level, there are two directions in which the intelligence is
increasing. One is to understand the user better and therefore give him relevant
results. The second is to make the search itself more powerful and better. The
merger of these two is leading to the emergence of many new search services.
|
The first approach essentially tracks individual customer preferences and
provides 'better' results the next time. Nothing very spectacular as far as
the concept is concerned. Humans have been doing that for years. If managing
directors ask for financial details of major IT companies, the assistant
provides the company accounts and if the sales manager asks for them, the
assistant provides the sales figures. If they ask Google for financial details
they may get everything else, but these. Google simply does not understand the
difference between the managing director and the sales manager. For that matter,
no other search site understands the difference either. Not only that, they
cannot decide whether the data is required for the quarter or the year? Is it
required in rupees or in dollars? And many more such questions, which have a
specific context associated with them? In order to be able to read minds search
facilities have to first understand the managing director and the sales manager.
So they are attempting to keep a track of what the visitor did on the last site
visit and which were the links that were accessed. They could also keep a track
of their individual profiles, what kind of an industry they work in, what is the
country of residence and other such stuff. Based on this, search facility can
give more intelligent responses. But in the process it also collects megabits of
unrelated data. A variation is to clarify the nature of the query by posing
counter questions. So when a user fills the search box, the engine fires back
with a string of questions designed to clarify the query. In the rudimentary
form-a bit like the Twenty Questions game. Both the approaches are
technologically challenging and suffer because many searchers want to give the
minimum information but expect the most relevant results.
The other approach is to keep the query simple and the understanding of the
customer limited to what is reluctantly provided but enhance the search itself.
So if the question is about financial details of IT companies, the results are
sorted chronologically, by country and are presented graphically with color
codes. Which is what a reasonably intelligent human colleague would do — if
presented with the same question. There are many examples of such enhancements
that are visible today.
Obviously there are a number of quantum leaps still imaginable in technology
development and deployment. And as long as they are imaginable they will happen
— sooner or later. And each of these would morph into distinct service
segments. Internet search is a badly organized departmental store with all goods
piled on to the floor in one heap.
Which brings me to the question—why is there no visible Indian effort in
this international departmental store? No search algorithms, no India specific
content search, no Indian language interfaces, no search services started by an
Indian company. Yes, this is an expensive business but the prospects are equally
high. Google this year is expected to rake in something like $600 million in
operating income on revenues of $1.6 billion. That is a 40% margin.
Are we waiting for the Big Bang to take place?
The author is Editor-in-Chief of CyberMedia, the publishers of Dataquest.He can
be reached at Shayam Malhotra