A New Dimension? (continued)

« Innovating Once Again | Main | Feed Nonsense »

A New Dimension? (continued)

Maybe I am falling into the classic entrepreneur move of seeing everything as validation of my ideas, but this post by Danny Sullivan on Search Engine Watch (via John Battelle) seems like validation of my relevancy rant.

Size doesn't matter, relevancy does.

In fact, size is the problem.

The Gotham Gal hates Google now because she gets too much garbage and not enough good results.

I don't feel the same way, but I understand why she feels that way.

Size of index is clearly the wrong vector. Relevancy is the way to go.

Comments (9) | | TrackBack (2)

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d83451b2c969e200e550367f7b8834

Listed below are links to weblogs that reference A New Dimension? (continued):

» Asking the right questions from The Daily Peg: Pegasus News, Inc. Founder's Blog
Relevance has become the new theme of all of our propaganda (and hopefully our product). It's on other minds too. In the midst of the pissing match between Yahoo! and Google over search size, Danny Sullivan at Search Engine Watch [Read More]

Tracked on Aug 18, 2005 5:20:32 PM

» Validation from Darby Wong
Fred Wilson alludes to “the classic entrepreneur move of seeing everything as validation of [his/her] ideas”. I think I tend to do that… for example, I see Trumba and Plaxo and I’m reassured that Daymill was a good idea. And ... [Read More]

Tracked on Aug 18, 2005 7:38:34 PM

Posted August 18, 2005 in Venture Capital and Technology

Comments

Google is fast becomming the Paris Hilton of search engines: all form and little substance. It's obsession with the number of incoming links works to the detriment of its search results. You can read more of my comments on it here http://smartstartup.typepad.com/my_weblog/2005/08/memo_to_yahoo.html

Posted by: Peter | Aug 18, 2005 3:28:33 PM

You're just focusing on one aspect that does matter and neglecting another one that matters. Both size and relevancy matter.

For a search on "Diamonds", with 8 million results, you can have an index of half or a quarter the size, do a great job on relevance and be the best engine out there.

But if you do a search on "weather podunk iowa 1924" you only get 55 results. If you're writing a novel about Podunk (assuming there is one http://www.straightdope.com/classics/a3_132.html ) in 1924 and need to know what the weather was like that day and there's only one webpage out there that provides the answer, you definitely want a big index.

Now I happen to feel that Google's results are more relevant and wonder if the extra pages Yahoo has at this stage are overly burdened by scraper sites and dupe content. In which case Google's index size is larger in terms of quality pages.

Did you catch Jeremey Zawodny's take on this btw?

Posted by: GilbertZ | Aug 18, 2005 3:50:50 PM

This is why I have believed that Google is highly vulnerable. Scraping is a commodity skill. Microformats, tagging, and other developments will allow specialized search engines to take high-value searches and ads away from Google ... and Google's hubris will let them.

Posted by: Jeff Jarvis | Aug 18, 2005 6:04:13 PM

I definitely agree that the problem is relevance. I think a small and innovative company is going to need to combine cookies with something similar to the recommendation feature of Amazon and Stumble to develop profiles of individual search users and tailor their results for relevence.

A basic model is not entirely out of reach with currently available technology.

Posted by: Dan Marques | Aug 18, 2005 8:12:28 PM

ah... good ole recall vs. precision. . . its easy to use index size as a proxy for recall (but very wrong), precision on the other hand is even harder to find a metric and measure.

But instead of being de-constructive, lets try to be constructive and suggest a few metrics for Yahoo and Google to release to the public. (maybe they need a little help:)) I suggest Pagination/Search Attempt, Natural Click Throughs/Results Page Shown, and Sponsored Click Throughs/Results Page Shown. Any other ideas?

Posted by: will | Aug 18, 2005 10:20:58 PM

Leave Google alone.

They're slowing growing up and leaving the nest.

We're going to have to learn to live with it.

GOOG is no longer a pure search play.

When they buy out Skype! inside of 6-12 months, you'll see what I mean.

There's a whole bevy of newer, fresher search engines out there anyway.

I like to beat up on Arfie over at Dogpile or get down and dirty with Clusty when I do my searching...

Posted by: Catablast! Media Group | Aug 19, 2005 3:39:42 AM

Yahoo is what it is today because they used to be the most relevant index on the Internet.

Way Back When the Internet reached only a couple percent of the population, Yahoo was an index. Lots of people (myself included) made it their home page, just as Google is now the home page of choice, becuase you could be reasonably confident of finding what you were looking for on Yahoo's index.

But then the dot-com bubble hit in full force, and suddenly everyone had to be a "Portal." Yahoo (and AltaVista and most of the other early search/index plays) cluttered up their homepages with everything from free e-mail services to dating services to credit card offers. I actually carried a Yahoo-branded Visa card until just a few months ago.

The other problem with Yahoo in particular is that the very fact that made its index the most relevant--being compiled by human editors--meant that it got far too expensive to maintain as the Internet grew. It didn't scale, to use a VC buzzword.

In fact, if you go to Yahoo.com today, you have to scroll all the way down to the bottom of the page to find the original Yahoo index, the thing that gave Yahoo its seat at the table.

Google became popular because they found a way to make search more relevant. If you think Google has relevancy problems today, think back on what other search engines were like a few years back.

So really, the whole breadth vs. relevance question is nothing new, and history shows that the winner at any given time is the company best able to provide relevance while still maximizing breadth.

Posted by: Shivering Timbers | Aug 19, 2005 7:44:24 AM

I’ve been toying with the idea of web searching for a specific item, say “sailing”. Now, I’ve been a web user for a decade now like the rest of the world and I’ve already spent hundreds of hours in the Lycos’, Googles and Yahoo’s of the world. You can say that I’ve a pretty good feel for what’s out there on the www in the way of “sailing” sites, posts or entries; yet I inevitably find myself BACK in the search engine typing in sailing and having to crawl through the 7,968,542 hits – not to mention the fact that this number grows every single day.

What if I could search by upload date? When I type in “sailing” I’m now only searching the “new” web for sailing hits. Granted this won’t work as well for “sailing” because it’s just too broad; but we’ll stick with our example. However, imagine entering the web-elusive musician M. Ward being the "searn term". I can either get notified (maybe via RSS) of new sailing entries. I envision this working in much the same way the “find” function works in Word, Adobe or Internet Explorer (Safari for Mac users like Kelly); BUT for the WEB and by DATE of upload. I’m only crawling for NEW stuff.

You can say that tagging and delicious sort of do this; but I’m not interested in OTHER people’s tags. What if 14 year old Joey Calucci takes a picture of his first zit which he got on a sailboat while at sleep away camp, tags in in delicious under “sailing” – I’m no interested. Granted of course “sailing” will yield noise and LOTS of it; but you can sot of manage the “new” web entries, rather than keep on crawling around for yesterday’s news…

Posted by: Marc | Sep 29, 2005 10:44:58 AM

Posted by: ad aware | Nov 17, 2005 11:07:37 AM

A VC

Musings of a VC in NYC