AGILEVC My idle thoughts on tech startups

June 26, 2009

There is an incredible amount of buzz about “real-time” search in the web ecosystem these days. And if the availability of real-time data makes possible more accurate prediction or inference of future events, than that will almost certainly be a catalyst for dramatic things.

Real-time search, now possible thanks to the continuously updated flow of information from Twitter and others, tends to be the focus of discussion. Venturebeat has a good analysis of some of the startups in this space, and it’s an area big players like Google and Microsoft are looking at closely.
But is what we’re calling “real-time search”in fact search as we typically know it? Many of the use cases folks cite are things like finding out about breaking news (Michael Jackson’s death or US Air #1549 landing in the Hudson), identifying popular trends, or delving into the discussion about topical events (what are Iranian protesters saying right now). I’d argue that most of these use cases are really more “discovery” than “search”.
Search is when you know basically what you’re looking for in advance, discovery is when you want things to be revealed to you. Search experts talk about lots of different forms of search… recall or recovery search (when you know a name or a brand and are trying to locate it), research search (trying to learn more about a topic or concept), competitive or comparative search (seeking similar concepts or objects to one already known), etc. Most of the products we call real-time search today involve consumers and companies wanting to find out what’s happening right now and perhaps explore the conversation or draw high level conclusions from it. That’s discovery.
So is this different just semantic? The correct answer is probably it’s too soon to say, but the differences might have implications for the business models around real-time data. Paid search advertising as we know it is valuable in large part because direct response marketers can see and measure the purchase intent embedded in many types of search. It’s unclear if or how this might translate to real-time discovery of information. Perhaps sentiment analysis or trend mapping tools will prove better businesses from the rise of real-time data than the “search” engines themselves, at least for companies. Perhaps consumers will be monetized through display or other forms of non-search advertising, as they come to rely on real-time search products for their news.
I have very little doubt that the proliferation and availability of real-time data will have meaningful impact on the web. But it’s entirely possible real-time will have a lot less to do with “search” than the current buzz suggests.