Fast For­ward: How Google got schooled at the game it mas­tered.

The Washington Post Sunday - - BUSINESS - ROB PEGORARO robp@wash­post.com

Google can give you free longdis­tance call­ing and pro­vide driv­ing, walk­ing, tran­sit or bi­cy­cling di­rec­tions to al­most any­where in the world. But can it find in­for­ma­tion on theWeb when you ask?

The­Moun­tain View, Calif., com­pany has made an un­usual con­fes­sion: It’s hav­ing some trou­ble with its orig­i­nal and pri­mary task. Asmy col­league Michael Rosen­wald

writes, Google’s rep­u­ta­tion for un­canny ac­cu­racy has been dulled by “con­tent farm” sites that game its search sys­tem to boost the vis­i­bil­ity of pages many read­ers say they don’t want. Google search en­gi­neerMatt Cutts’s

Jan. 21 blog post ac­knowl­edged their com­plaints: “we hear the feed­back from the web loud and clear: peo­ple are ask­ing for even stronger ac­tion on con­tent farms and sites that con­sist pri­mar­ily of spammy or low-qual­ity con­tent.”

It’s eas­i­est to see this prob­lem if you search for a re­view of a prod­uct or in­struc­tions on how to do some­thing and find your­self look­ing at dozens of ir­rel­e­vant re­sults that don’t an­swer your ques­tion or that rip off an­other site’s work.

There’s not much point in get­ting too mad at these sites. They’re sim­ply fol­low­ing a prime di­rec­tive of the com­mer­cialWeb: Get peo­ple look­ing at your site, then use ad­ver­tis­ing — of­ten placed through Google’s ser­vices — to trans­mute that traf­fic into money.

( We in the me­dia play by these rules, too. Build­ingWeb traf­fic through “search

en­gine op­ti­miza­tion” has be­come a ma­jor

part of a jour­nal­ist’s job.)

Mean­while, plenty of writ­ers, pho­tog­ra­phers, videog­ra­phers and editors are will­ing to ac­cept min­i­mal per­prod­uct pay­ments to crank out a large vol­ume of posts that­match up with com­mon Google searches.

As one re­sult of this dy­namic, Ya­hoo paid $100 mil­lion for a con­tent mill named As­so­ci­ated Con­tent in­May. The best-known com­pany in this cat­e­gory, De­mandMe­dia, staged its ini­tial pub­lic of­fer­ing Wed­nes­day and closed the day with a higher mar­ket value than the New York Times Co.

Both sites would dis­pute the con­tent­farm cat­e­go­riza­tion, and I’ve seen each pub­lish use­ful in­for­ma­tion. But I’ve also

read plenty of dreck at these sites, and there’s no dis­put­ing their busi­ness model

of mass-pro­duc­ing con­tent to fit, Le­go­like, with search terms.

In essence, Google has un­in­ten­tion­ally been teach­ing to the test — and some of its stu­dents have learned all too well.

As Rosen­wald’s ar­ti­cle sug­gests, this opens up an op­por­tu­nity for so­cial­net­work­ing sites that con­nect peo­ple with trusted, knowl­edgable friends to beat Google at its own game. But by re­quir­ing your iden­tity to work, they in­cor­po­rate pri­vacy and se­cu­rity risks.

It will not be a healthy devel­op­ment for theWeb if find­ing use­ful data on­line re­quires a user­name and pass­word. Google com­peti­tors such asMi­crosoft’s

Bing and smaller ri­vals such as the com­mu­nity-cu­rated Blekko or the data ori­ented Wol­fram Al­pha have an ex­cel­lent op­por­tu­nity to do a bet­ter job of con­nect­ing peo­ple to in­for­ma­tion they need — not just pages that try to look use­ful to a search en­gine’s au­to­mated in­dex­ing sys­tem.

But Google will have to be­come a lit­tle pick­ier too, as Cutts’s blog post sug­gests it will. And as that hap­pens, Google may run into a sec­ond prob­lem: “search neu­tral­ity.”

The term sur­faced a few years ago, in part be­cause op­po­nents of net-neu­tral­ity reg­u­la­tions be­gan talk­ing up that an­gle. (In ret­ro­spect, that PR tac­tic looks like a clever ex­ploita­tion of the me­dia’s weak­ness for he-said/she-said sto­ries that look “ bal­anced” by quot­ing each side at­tack­ing the other.)

The no­tion car­ries a lot of built-in ridiculousness. Web search is in­her­ently an ed­i­to­rial act — and not an easy one, ei­ther. You’re ask­ing a site to sift through about a tril­lion pages and find the ones most rel­e­vant in less than a sec­ond. The whole point of search — as in jour­nal­ism — is to ex­er­cise bias against things judged to be less rel­e­vant.

But when ad­just­ing search al­go­rithms to de­value mass-pro­duced, low-qual­ity con­tent can knock the legs out from un­der­neath a con­tent farm’s busi­ness model, you don’t need a search en­gine to pre­dict that calls for “search neu­tral­ity” will in­crease.

Google and other search sites can in­su­late them­selves from some of this pres­sure by be­ing more trans­par­ent about how they re­vise their search sys­tems.

Google in par­tic­u­lar should take one ex­tra step: Al­low users of An­droid phones to change the de­fault search en­gine in its mo­bile op­er­at­ing sys­tem as eas­ily as they can in Google’s Chrome browser.

But trans­parency and open­ness — not to men­tion the fact that no­body is be­ing forced to use Google — may not be enough to fend off calls for reg­u­la­tion. The Euro­pean Union’s an­titrust reg­u­la­tors launched an in­ves­ti­ga­tion of Google last year and have be­gun ask­ing sites whether they think Google

ma­nip­u­lates search re­sults. Google thinks enough of this pos­si­bil­ity that it sent Cutts toWash­ing­ton two weeks ago to talk to pol­i­cy­mak­ers and jour­nal­ists,

my­self in­cluded, about these is­sues. He and Google don’t have an easy task in store. They’re go­ing to need more than a con­tent farm’s how-to write-up to get through it.

JEF­FREY COOLIDGE/GETTY IM­AGES

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.