Join Date: Jun 2004
Location: New York, USA
Search Memories - Live from SES San Jose
Chris Sherman looks at his watch, looks up at the audience, looks back at his watch. The room is basically empty, could the Overture party last night have kept everyone sleeping this morning? Anyway, its 9am here and the show has to get started.
Chris introduced the speakers. Doug Cutting the senior engineer from Excite, Steve Kirsch the founder of InfoSeek, Louis Monier the primary 'instigator' of AltaVista. He then threw up that gimmick where someone typed out the Google homepage on a type writer and wrote it was up in 1960, by fury.com. He said its hard to know the exact dates of the new technologies. Steve Kirsh made an announcement on July 18, 1994 saying that InfoSeek is using Python. Steve said in his post that he is giving out 5,000 FREE accounts. So then Steve began speaking.
The original concept of InfoSeek came from a computer library (on a disk). He used it daily and then decided to make it available to more people at a lower price. He wanted to make this available at a very aggressive price for these 150 periodicals. He worked out a deal where he can give this away to 5,000 people for free. Then beyond that would be $10 per month for this information. They then added as an additional resource such as a Web index. Kind of moved from there. He wrote excite in Python, he likes Python (most people don't),
In September 27, 1995, there was a post to toutell.com on the announcement of Excite.com. They said its a free Web search with 1.3 million web pages. He clarified he was not an original founder, he came on a few months after this date. He said the original founders were roommates in college. He said that they were looking to work together so they decided to go into this text search software. They brought a crawler and demoed it off one PC to a VC. The VC pointed them in the direction of Web search. Excite did both Web search and Web directory. Excite is arguably the first portal. Doug started in January 1996. Then soon after AltaVista was started up, which launched with 16 million Web pages indexed and it was fast! So one of the VPs at Excite said, lets crawl 50 million. 4 - 5 months they relaunched with 50 indexed pages. They were the largest search engine for a period of time and soon were surpassed by many others. It was a good time.
AltaVista posted on December 15th 1995. This was the internet's first "super spider". Talked about the volume covered and the speed of the search. The one thing that really never came through from the press was that there was no support. AV was about 6 people, locked inside a dying software company. Louis convinced his company to invest (in one server) to let them do this and get good PR. He said AV is a story against all odds. He said how the company was like, "Digital planned this big thing out", when in fact it was Louis's idea one day. Louis's wife designed the logo, etc. He said AV never had any business development. Yahoo came to AV and AV didn't know what to do. He said it was a great technical rush but a business disaster.
Chris then asked them to touch on some of the new pioneering features they came up with.
Doug from Excite started off by saying they didn't necessarily invent them but made them popular. They said they used "discrimination analysis" or a big thesaurus. Like, words you would add to a query to refine the search. It was very popular and then it was not used later on. They also had a feature named "channelized search", basically if yo u typed in San Francisco baseball and it knew there was only one team. It would show you the score of the game in process, and a link to the site. It had maps, etc. A lot like what everyone is trying to do today. It died out at Excite but is now popular.
Steve from InfoSeek said they were one of the first to put up banner ads on search results pages. They were the first to come up with the CPM model (he believes). They had enhanced relevance because they looked at the links to your Web pages, much like Google today (poor man's google). One of the firs to have phrase search and then soon added a syntax (+, -, etc). The last innovation he is proud of is AltraSeek, which was a corporate search tool. Steve had a ton of opposition with this in his company, but he did it anyway. AltraSeek was sold to Inktomi and you can still find this feature in some companies.
Louis from AV said people wanted to make AV into a portal. Every attempt to make AV a portal was a disaster. He seems really ticked off about AV's business division (which makes a ton of sense to me). On the search side he feels pretty good, it was all about speed, simple GUI - you can see these blueprints at Google today. Universal access, he spent a week trying to work out a bug with a very small browser (used by 0.00023% of the Internet population). he made sure it worked on every single browser. He used to joke that AV would work on a washing machine. The languages worked well across some countries. He then went to a unicode index (Russian, Chinese, etc). He said the business took us down again. An other really good thing was the babelfish translation service. (I personally loved this service and I use it today when reading some non US SEM forums). AV launched something that was once called "Life Topics" that was a super techie thing, it was cluster your results a major refinement search. He said, when he left AV, they unplugged it.
Chris then asks them to look into the future and touch on what you feel are miss opportunities for their company. Chris read off an article where he said InfoSeek missed the opp to buy Yahoo! for 20 million because he felt it was over valued. Also he turned down eBay as well.
Larry Page came to Steve from InfoSeek, asking them if they want to license PageRank. So did Doug and Louis claimed he was impressed with PageRank but didn't have the power to sign the check. They all believe that was a big mistake. 1997 Google was asking to license it, then in 1998 -99 they had their own site and began taking on all the users. Doug said that there was little innovation since 1998.
Steve really regrets not just focusing on search. He said he was the only one pushing the search size index. The business people said that people don't want so much, they just want the top pages. The management and VCs wanted to move it to a portal. He gets credit for turning down Larry, and help inventing Google.
Louis said he missed the Google opportunity as well, but he had no power to do anything. His own personal regret, a few days after they launched AV, Excite called him. He turned down a job opportunity at InfoSeek, where he could of had purchased two jets. AV missed the aspect of link popularity to improve the relevancy of the Web. AV used it to map the Web in size but not to improve relevancy.
Doug adds that Google also innovated the ability to search on one word queries. Everyone else had the default operator was OR. But Google used AND as the default (which was simple but genius). Focusing on precision and recall was also an area where Excite did not focus on.
Louis said Google built a system that scales very easily, the hardware at Google is very impressive. He bets Google can double its index with no problem, because of their hardware architecture. Doug adds that he was talking to Larry Page, and Doug told Larry that Google has an excellent brand but the search quality will not always be best. Larry said that people do not use Google because of its brand but because of its operations. He said just like with Coke, Coke's brand is because of its operations. Louis said Google does their stuff on the cheapest possible platform they can.
Chris then gets into valuation and they all start laughing at the dollar signs. Chris asks if we are going through an other bubble or???
Doug said he wont comment and won't speculate on the market.
Steve said the difference here is that Google has real revenue and real profits. The other companies did not have this.
Louis agrees with Steve but he is not buying any stock.
Doug adds he feels that the IPO is a bad thing for Google and they were smart to push it off so far.
Chris then asked them to look into the future.
Doug started talking about his Nutch, an open source Google. People find it useful to build and search niche search engines. He said its hard to know where it will go. Its open source, good enough quality and major commercial search engines will use. That is predicated on that search technology is becoming a commodity and is not getting much better.
Steve doesn't like to predict in the future, but if an other Stamford student comes to him, he will listen.
Louis said search is not sticky. Google is now adding elements to keep people there (gmail, photos, etc.).
Join Date: Jun 2004
Location: New York, USA
Q & A:
Q: Mike Grehan asks what they think of PageRank, Teoma's (HITS) answer and when did you realize spam is an issue?
A: Doug said spam was thought of as search quality, improve the quality. You can also build spam filters as well. Google probably does both. He feels PageRank is overrated.
Steve said PageRank is underrated, if it wasnt really that important then why is Google where they are? The fundamental difference was PageRank. If you don't have the precision that Google has, then none of that matters. Steve doesn't know much about Teoma.
Louis said pure PageRank is a joke, Google clearly uses a mix of things. Teoma he feels is a good idea, looking at subjects but it will be great to see it bigger, faster, personalized, etc. Louis said spam in Av was a nightmare. He said one day they turned something on where you can add pages to the index in real time. Louis was so excited because the Web index will increase. He didn't think for a second about spam. One guy wrote a script to pound spam into the index and break the system. AV tried everything and they never had a good answer. They tried using a duplicate content filter and keyword frequency. He said he feels for Google.
Join Date: Jun 2004
Location: Washington state
Great recap, thanks RB. I would've loved this one. Shame it wasn't too popular.
And so true about the OR / AND issue. That's when SERP quality really started to improve. Such a little thing ... and such a big thing, too.
|Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)|