John Battelle's Searchblog

I worked at Realnames back in 1999 and we had two search-related businesses -

1) Internet Keywords, which were based on the idea that, like the URL replaced the IP address, natural language would ultimately replace URLs as the way people navigated the web from within the browser URL line. We had a system that mapped natural language terms to URLs such that someone looking for information on 'Audi A6' could simply type 'Audi A6' into the URL line rather than www.audi.com, then 2-3 clicks into the site to get to Audi's page for the A6. We had a distribution deal with Microsoft whereby our system worked in IE, and tried our best to get AOL/Netscape as well, at which point we would've been able to place a natural language navigation layer on top of URLs and simplify the process of navigating to a site.

2) We also had a paid search model where we sold keywords to advertisers and had distribution relationships with AltaVista, MSN, Google and a number of other search engines such that our paid search results would be displayed at the top of those search engines. Much like Overture, and in fact RealNames' founder (Keith Teare) claims to this day (and I've been able to corroborate this) that he actually gave the idea of paid search distribution to Bill Gross at Idealab back when GoTo didn't yet have a revenue model.

Suffice to say that advertisers were lining up all day long back in 1999 and 2000 to buy the search distribution rather than Keith's strategic vision of replacing the URL with natural language and creating an offline-to-online marketing enhancement/measurement service in Internet Keywords. Whereas Keith thought big offline advertisers would love to replace http://www-1.ibm.com/services/us/index.wss/in/imc/a1000038 with 'Internet Keyword: IBM Aeraspace Solutions', the reality was that since consumers didn't yet know about this functionality, no one would use it.

HOWEVER, our paid search distribution business was selling like hotcakes, and the few of us who were capable of standing up to Keith's vision-at-all-cost, domineering personna, tried to refocus the company on the huge paid search opportunity we saw. But to no avail; Keith and the oblivious board refused to accept the reality of the situation (namely that neither MSFT nor NSCP were willing to give RealNames 100% control of the URL line) and refocus on the huge paid search opportunity.

So that's one story of bosses missing the boat over [paid] search in the late 90's, and it's one that literally cost RealNames investors billions and billons of dollars. I've since used my options paperwork to start a barbeque...

I used to work at Looksmart in the late 90's, and at one point I moved into the "Search" group during one of the periodic reorgs. The task in our first meeting as a group was to come up with a mission statement, so I raised my hand and said we should build "the best &^% search on the internet". I was flattered that everyone agreed except the group manager, who had apparently gotten some winks and nods from higher-ups about what the mission was, so she took it under advisement.

The management buzzword at that time escapes me, but it was something like monetization, or "sponsored search", or even something oxymoronic like "paid relevancy". It came down to finding any way possible to sell credibility for cash.

A week later, the final mission statement came out. Unfortunately, I don't even remember the exact wording, but the word "%^&" had been replaced by the buzzword, so it came out something like "the best monetized search on the web". Needless to say there wasn't a wholesale rush to improve our search technology.

If I hadn't been there so long I would have been shocked, but this sort of thing happened every day.

Here's a related question: what is broken and ignored right now that needs to fixed in the same way search was back then?

Great blog.

Noticed today that I went to a9.com for the first time to search it did a quick redirect into www.amazon.com before going back to a9.com. Then...a9.com had grabbed my cookies from amazon.com and knew who I was.

FYI, but I'm sure you already knew.

I remember around that time AltaVista introduced pop-up banners. In 1998 you didn't even have to look into results to get the spam -- they were right on the search engine front page. The first AltaVista pop-up was the time I switched to Google, and I stuck to it ever since... like millions of others.

I don't recall search being all that bad, only the delivery of it. Focus remains unver-rated.

what is broken and ignored right now that needs to fixed in the same way search was back then?

My personal gripe is that a few areas -- off the top of my head, hotels and car rentals -- simply can't be searched: it's reminds me of the days in which everything was touched by the hand of smut. Stepping back, the problem is that where there are well-established sites devoted to particular commercial comparisons, search is often incapable of providing the big picture alternative that one usually encounters on a Google results page. I'd like something more 'meta' than having to throw my hat in with Orbitz, Expedia or Travelocity. (Or Citysearch -- are they still going?)

My experience was on the webmonkey side, being asked to lard pages with meta tags. Something I will never do again.

It's funny, though, to recall RealNames, given that Google has, to some extent, circumvented the doman-name space. (And to some extent, reinvented it: one wonders whether the motivation behind [movietitle]movie.com or [movietitle]themovie.com is through analysis of search queries as much as a reflection that most movie-title domain names have long been taken.)