Michal Migurski's notebook, listening post, and soapbox. Subscribe to this blog. Check out the rest of my site as well.

Feb 28, 2006 7:41am

death to search engines

David Hawk writes:

As Google and Yahoo like to proudly proclaim, a PPC advertisement on one of their networks reaches "98% of active Internet users" with well over half of that metric coming straight from a search engine. So . . . with such astounding success, there's no better time than to predict the inevitable demise of the search engine.

...immediately prior to predicting the inevitable death of search engines, because he believes they will be disintermediated like travel agents and classified ads (and bars, and the phone company ...huh?). He believes that they will be replaced by smart sites that learn about your own habits as you browse to give you a more personal experience, imagining that typing "Los Angeles travel" into a search box should yield personalized travel results for you. Aside from the obvious holes (who's in the best position to deliver this personalization besides Google or Yahoo?), I think Hawk has too much faith in the power of personalization algorithms, too little faith in the strength & cheapness of recommendations from other people, and not a lot of imagination to see the ways in which internet search makes agents and brokers more valuable.

Personalization algorithms are pretty terrible. Amazon is the classic example, because their "People Who Bought..." feature is actually good, but it's working with a strongly-controlled universe of information and is under little consumer pressure to kick ass. If Amazon's recommendations aren't exactly what I want, the worst-case scenario is that I move on to somthing else, while Amazon loses a sale. The majority of the cost is borne by the service, not the user. If my hypothetical search results for "Los Angeles travel" are bad, I'm going to be pissed because I was actually looking for some help. Imagine the misery if they proved worthless after recommended hotels and flights are booked. Imagine the annoyance if I'm not even looking to travel to Los Angeles. Leaving this up to machines will be like the Smart Tags debacle all over again:

As Paul Thurrott of Paul Thurrott's WinInfo found out, this sort of thing can get old in a hurry. While typing up his review of Office XP (using Word 2002, of course), Mr. Thurrott typed the word "nice." Up popped a smart tag offering to book a flight to Nice, France using Microsoft's Expedia website. When he typed the word "long," up popped a smart tag from ESPN offering more information on Oakland Athletics centerfielder Terrence Long. As Thurrott put it, "Folks, this is lame."

Besides, there are millions of people out there perfectly willing to lend a hand, give advice, throw in their two cents, or start a business selling expertise. When I need to something, I usually contact one of these people, because I know that they know what I don't know. This used to also work on a slightly less-personal level with e-mail lists, though there's apparently a website called "Craig's List" which hosts these kinds of conversations for local interests.

A lot of the early noise about Web 2.0 focused on how a new breed of web technologies helped to focus interpersonal relationships around shared interests or goals, and how there was a lot of value to be found providing forums where people could communicate with each other more easily. I don't know why the bandwagon took a wrong turn into a wasteland of frameworks, collaborative filtering algorithms, and ajax portals, but I'd very much like to see it return to the main road. If it's going to be necessary to spend weeks or months waiting for some PhD thesis project to assemble a mathematical model of my wants and needs so I can conduct searches online, why not focus on solving the much simpler problem of connecting people to other people who are born experts at this? The "stars" of Web 2.0 excel at this now. The common thread running through projects like Delicious, Flickr, Digg and others is that they focus on this need first and worry about the crazy math shit later.

October 2021
Su M Tu W Th F Sa

Recent Entries

  1. Mapping Remote Roads with OpenStreetMap, RapiD, and QGIS
  2. How It’s Made: A PlanScore Predictive Model for Partisan Elections
  3. Micromobility Data Policies: A Survey of City Needs
  4. Open Precinct Data
  5. Scoring Pennsylvania
  6. Coming To A Street Near You: Help Remix Create a New Tool for Street Designers
  7. planscore: a project to score gerrymandered district plans
  8. blog all dog-eared pages: human transit
  9. the levity of serverlessness
  10. three open data projects: openstreetmap, openaddresses, and who’s on first
  11. building up redistricting data for North Carolina
  12. district plans by the hundredweight
  13. baby steps towards measuring the efficiency gap
  14. things I’ve recently learned about legislative redistricting
  15. oh no
  16. landsat satellite imagery is easy to use
  17. openstreetmap: robots, crisis, and craft mappers
  18. quoted in the news
  19. dockering address data
  20. blog all dog-eared pages: the best and the brightest