Letter to Jerry Yang, CEO Yahoo!
The Wall Street Journal article on Saturday, January 19, 2008 really struck me that you're having a hard time getting your company to gel and get back on track at a pace that investors will value.
While at IBM Lotusphere, I took the IBM Labs Tour and discussed with a few IBMers the technology of their research focus and saw where it could address the basic business need at the heart of your search woes. Of course, Yahoo! is many things, but foremost it must be a search platform. Here, you need to define a new category of search.
At Lotusphere I saw a blog crawler. Perhaps as a competitor to the Google Alert service, this tool captures RSS feeds from a universe of blogs, enabling the analysis of content where a marketer is instantly notified of changes in their brand's treatment and timely enough so that they could make effective, corrective and responsive changes to their press/blog outreach and program. Particularly important in minute-by-minute markets like the Presidential Primary campaign, this capability could change the dynamics of search and create a new approach to search. On an Internet scale operation, the service would have a different crawl profile.
Today, Google rents thousands of computers around the world every few days to visit and capture millions of links. Then they process these results and deliver a massive index of the web that is at the heart of Google.com searches.Blog searching needs to be done in much faster chunks of processing, probably with hourly index updates. The advertising algorithms that would complement blog searches are a little more substantial than Google Adwords. The opportunity to maximize advertising is when there are dramatic changes in the typical signal-to-noise ratio of a topic.
What is [[signal-to-noise]]?
In many communications systems, the background noise is the systematic error or random energy that occurs whether any useful information is being transmitted or not. The signal refers to the transmission of the information. A high signal-to-noise ratio refers to a quality communication, while a low signal-to-noise ratio means that the background noise is relatively high compared to the signal so some information may be lost.
In the context of the blogosphere, it refers to a change in the typical quantity and quality of blog content. For example, on any given day 100 bloggers write about Microsoft Office. Suddenly 1,000 bloggers wrote about the suite – it would be useful to know why, and it would be useful to know whether it was neutral, positive or negative commentary.
It is likely that Microsoft, for example, will want to advertise as a result of this change to make the peak happen higher if its a positive trend and will want to advertise even more aggressively if its a negative trend with the goal to turn the signal around. With a contrarian view, Microsoft's competitors may want to advertise more aggressively if its a positive view and less so if it's a negative trendline.How to measure positive compared to negative trend lines?
Jerry, let's get moving to defining that new category of search and making it relevant on the basis of time. The classic search paradigm is stagnating. It's time for a disruptive implementation and you could be it.
This post has already been read 0 times!