Google 12.11.18 Hearing QFR Responses

QFRsfromChairmanBobGoodlatte 1. In terms of Google'ssearchresults,whatarethedifferencesbetween personalization,contextualization,customization,andlocalization? personalization,contextualizatio n,customization,andlocalization? GoogleSearchrespondstotrillionsofuserqueriesfromaroundtheworldeveryyear,driven byourcompanymissiontoorganizetheworld’sinformationandmakeituniversallyaccessible byourcompanymissiontoorganizetheworld’sinformationandmakeituniversallyaccessible anduseful.Wehavefoundthatthequeryitself—notanydataabouttheuser—isbyfarthe mostpowerfulsignaltoidentifythemostrelevantandusefulresults.Thus,weonly personalizeresultswhenthereisaclearbenefittotheuser,suchashelpingthemcontinue researchthey’vebegunoransweringquerieslike,“wheretodonateclothes.”Whilewedon’t refertoSearchintermsof“contextualization”and“customization,”wedodrawdistinctions amongafewformsofrankingsignals: ●

●

●

●

Personalization signalsrelatetoinformationabouttheparticularuser,suchastheir signalsrelatetoinformationabouttheparticularuser,suchastheir interestsandpastSearchhistory. Sessionsignalsidentifyprecedingactivityintheuser’ssessiontounderstandthe Sessionsignalsidentifyprecedingactivityintheuser’ssessiontounderstandthe contextoftheirnewquery.Forexample,ausersearchingfor“soccer”followedbya searchfor“Barcelona”maybemorelikelytogettheBarcelonasoccerteamasaresult fortheirsecondsearch.Howlongasessionlastsdependsonuserbehavior(e.g.,how longuserskeeptheirbrowserwindowsopenorhowlongusersaresignedin). Locationhelpsusunderstandthecontextoftheuser’squerytogivethemmore Locationhelpsusunderstandthecontextoftheuser’squerytogivethemmore relevantresults.Forexample,ausersearchingfor“Giants”inNewYorkCitymaybe morelikelytogetsearchresultsfortheNewYorkGiantsratherthantheSanFrancisco Giant Giants. s. Lastly,userscanaltertheirSearch Lastly,userscanaltertheir Search settings,suchaslanguage,andinsomecases, settings,suchaslanguage,andinsomecases, thosesettingsmaychangetheresultstheysee.

Weareanindustryleaderwhenitcomestochoice,transparency,andcontrolforusers.We remaincommittedtoensuringourusersunderstandhowSearchworksandthattheycan controlwhatSearchactivityweusetoimprovetheirSearchexperience.Thatincludes adjustingwhatdataissavedtotheirGoogleAccount(availableat https://myaccount.google.com),andourrecentlylaunched“YourDatainSearch,”acontrol https://myaccount.google.com ),andourrecentlylaunched“YourDatainSearch,”acontrol panelthatenablesuserstobrowseanddeletetheirSearchactivitydirectlyforthepasthour, orevenentirely,withoutnavigatingawayfromwhatthey’redoing.Youcanfindmoreonhow SearchworksandwhatdatapowersyourSearchresultsat https://www.google.com/search/howsearchworks/and https://www.google.com/search/howsearchworks/ and https://myactivity.google.com/privacyadvisor/search.. https://myactivity.google.com/privacyadvisor/search 2. How Howar arec econ onte text xtua ualiz lized edor orl loc ocali alized zeds sear earch chr resu esult lts s(s (suc uch has ass sho howi wing nga a differentsetoflinksperCongressionaldistrict)differentfrom"redlining"consumer classifications?

Trusted by over 1 million members

Try Scribd FREE for 30 days to access over 125 million titles without ads or interruptions! Start Free Trial Cancel Anytime.













Ourbusinessmodelandcompanymissiondependonusbeingausefulandtrustworthy sourceofinformationforeveryone.Redlining—whichgenerallyreferstoidentifyingaclassof consumerswithwhomabusinessdeterminesnottotransact,includingbyplacinglimitations ongeographicareasthatcancorrespondwithincome,race,orotherfactors—isatoddswith ourmissionandbusinessmodel.Wedonotengageinredlining,nordowebelievethatthe signalsweusetoensureweprovidethemostrelevantresultstoourusershavethesame effect.Weonlypersonalizeresultswhenthereisaclearbenefittoourusers,suchashelping themcontinueresearchthey’vebegunoransweringquerieslike,“wheretodonateclothes.” Andweuselocationsothatouruserscangettheexactanswertotheirquestion.For example,whentheyaskGooglefortoday’sweather,theymostlikelywanttheweatherintheir area.Providinglocallyrelevantresultsisonewaywemeetourcommitmenttoourusers. 3. When WhenG Goo oogl glel eloc ocali alizes zess sear earch chr res esul ults ts, ,it iti ins nser erts tsl loc ocal ally lys spec pecifi ific cres resul ults tsl lik ike e restaurants.Doyoualsochangearoundthegeneral,non-localresultsbasedonlocation (suchasshowingageneral,non-localWikipediaarticlehigherintherankingsforsome communitiesthanothers)? Insomecases,auser’slocationcontextcanbehugelyimportanttothequalityoftheirresults. Forinstance,auserinVirginiasearchingfor“propertytax”likelywouldfindadifferentsetof pagesusefulthanwouldauserinCaliforniarunningthesamesearch.Ofcourse,notall queriesareparticularlylocation-sensitive,andGoogletakesaccountofthataswell.Wehave foundthatthequeryitself—notanydataabouttheuser—isbyfarthemostpowerfulsignalto identifythemostrelevantandusefulSearchresults. 4. How Howcan canc con onsu sume mers rsg get etan anu unf nfil ilter tered edsea searc rch hexp experi erien ence ceon onG Goo oogl glet etha hat t doesnotchangethesearchresultsbasedonanypersonalizedorcontextualized information?Forexample,iftwovoterssearchedatthesametime,besideslinksonly relevanttotheirlocation(e.g.,restaurants),howcouldtheyensuretheywouldseethe samesearchresults? UserssearchingonGooglegeneratedatasuchasthetermstheysearchfor,thelocationof thedevicetheysearchfrom,andthelinkstheyinteractwith.Google’sMyAccounttool (availableathttps://myaccount.google.com (availableathttps://myaccount.google.com)letsuserscontrolthedatathatissavedtotheir )letsuserscontrolthedatathatissavedtotheir GoogleAccount.Also,ourrecentlylaunched“YourDatainSearch,”enablesuserstobrowse anddeletetheirSearchactivitydirectlyforthepasthour,orevenentirely,withoutnavigating awayfromwhatthey’redoing.Buteventhen,twousersrunninganidenticalsearchmaysee differentresults.Theirqueriesmaybehandledbydifferentdatacenterseveniftheyarein thesamelocation,tospreadouttheload.Andwhilewetrytokeepourdatacenterscloselyin sync,theconstantrollingupdatesthatwemaketobothourSearchalgorithmandSearch















theirphone.Again,wehavefoundthatthequeryitselfisbyfarthemostpowerfulsignalto identifythemostrelevantandusefulresults. 5. InI InInc ncog ogni nito tom mod ode ean and dsig signe ned dou out tof ofGo Goog ogle le,i ,if fac acon onsu sume mer rdo does esa ase sear arch ch onGoogle,isthatsearchsavedwithanykindofuniqueidentifiersuchthatitmaybe associatedwithotherpersonaldatasuchasothersearches? WhenauserconductsasearchonGoogleinChromeIncognitoandsigned-outmodes,weset acookietocorrelatesearchesconductedinthesameIncognitowindowduringthesame browsingsession.Thissessionsignalallowsustousetheprecedingactivityintheuser’s sessiontounderstandthecontextoftheirnewquery.Forexample,ausersearchingfor “soccer”followedbyasearchfor“Barcelona”maybemorelikelytogettheBarcelonasoccer teamasaresultfortheirsecondsearch.Howlongasessionlastsdependsonuserbehavior (e.g.,howlonguserskeeptheirIncognitobrowserwindowsopen).However,information fromIncognitoandsigned-outsearcheswillnotbeassociatedwithauser’s signed-inuser datatoinfluencetheuser’ssearchresults.Finally,userscanstoptheuseofthesession signalsonChromealtogetherbygoingtotheirPrivacy&Securitysettingstoturnoffthe settingthatreads,“Allowsitestosaveandreadcookiedata.” 6. InI InInc ncog ogni nito tom mod ode ean and dsig signe ned dou out tof ofGo Goog ogle le,i ,if fac acon onsu sume mer rdo does esa ase sear arch ch onGoogle,whatareallthesignalsthatareusedto"contextualize"thatsearch?Areany ofthesesignalsbasedonadataprofiletiedtoanykindofuniqueidentifierthatmaybe associatedwiththedeviceusedforthesearch? Asdescribedabove,ifauserisinIncognitoModeandsignedoutofGoogle,Googlewillnot tietheSearchrankingstoanysigned-inuserdata.Wewill,however,usecertainfactorsnot tiedtotheuser’sidentity,suchasthebrowsertype,language,timeofsearch,location(oran estimationoflocation),andpriorbrowsersessionsearches,toimproveSearchranking relevancefortheuser’squery.MoreinformationonhowSearchworksandwhatdatapowers users’Searchresultsisavailableathttps://www.google.com/search/howsearchworks/ users’Searchresultsisavailableat https://www.google.com/search/howsearchworks/.And, .And, userscancontroltheirsettingsbygoingto https://myactivity.google.com/privacyadvisor/search.. https://myactivity.google.com/privacyadvisor/search 7. Goog Google leo ope perat rates esa asu suit ite eof ofser servi vices ces( (e. e.g. g., ,Go Goog ogle leDo Docs cs, ,Gm Gmail ail) )in inad addi diti tion on toGooglesearchthatcollectpersonalinformationonAmericans,whethertheyuse Googlesearchregularlyornot,oraresignedintoGoogleornot.Whenaconsumer comestoGoogle,inIncognitomodeandsignedoutofGoogle,isanyinformationfrom othersources(onlineoroffline,suchasadviews,clicks,browsinghistory,purchase history,locationhistory,etc.)usedtocontextualizehissearchresults?Whatifthe















Asdescribedabove,ifauserisinIncognitoModeandsignedoutofGoogle,Googlewillnot tietheSearchrankingstoanyoftheirsigned-inuserdata.Ifauserissignedoutbutin non-IncognitoMode,GooglelikewisewillnottietheSearchrankingstosigned-inuserdata, butmayusepriorsearchestiedtotheuser’ssamebrowsercookietoimprovetheirSearch results.Ifauserissignedin,GooglewillhonortheGoogleAccountsettingsthattheuserhas selectedwhenshowingthemSearchresults.Userscancheckandcontrolthosesettingswith Google’sPrivacyCheck-Up(availableathttps://myaccount.google.com/privacycheckup Google’sPrivacyCheck-Up(availableat https://myaccount.google.com/privacycheckup)). 8. Atw Atwhat hatd dis ista tanc nceb ebet etwee ween ntw two odev devic ices eswo woul uld dGo Goog ogle les sear earch chr res esul ults tsl like ikely ly changebasedonlocation(e.g.,0.01miles,0.1miles,1mile)?Thatis,wouldGoogle searchresultslikelychangebasedonlocationbecausepeopleliveinthehousenext door,thenextcityblock,thenextzipcode,theCongressionaldistrict,nextcity,orthe nextstate? Thisdependsontheuser’ssearchquery.Forexample,ausersearchingfor“restaurantsnear me,”whoprovidesGooglewithahighlyaccuratelocation,suchasareadingfromtheGPSon theirphone,islikelytogetdifferentresultsastheymovefromonecityblocktoanother.But forlesslocation-sensitivequeries,suchas“whereisthenearestDMV,”theuser’sresults wouldlikelynotchangeacrosssmalldistances. 9. Goog Google leh has asch char arac acter terize ized dth theD eDuc uckD kDuc uckG kGo ost stud udy yas asm met etho hodo dolo logi gical cally ly unsound;ifGoogleweretoconductthisstudy,listallthemethodologicalwaysyou wouldimproveitsothatyouwouldnotconsideritflawed. WecontinuetobelievethattheDuckDuckGoresearchisflawedforanumberofreasons, including: Size:Thestudyinvolvedfewerthan90users,whileweusethousandsofsearchraters Size:Thestudyinvolvedfewerthan90users,whileweusethousandsofsearchraters ● fromallovertheworldtotestourSearchresults. IncorrectAssumptions:Thestudyisbasedonincorrectassumptions—namelythat IncorrectAssumptions :Thestudyisbasedonincorrectassumptions—namelythat ● anydifferenceinsearchresultsisrelatedtopersonalizationderivedfromprevious searches.Instead,differencesinsearchresultsareattributabletomanyfactors outsideofSearchpersonalizationthattheDuckDuckGostudyfailstoconsider, including,timingofthequery,whetherthesearchisonmobileordesktop,and localizationofresultsforqueriessuchas“restaurantsnearme.” Ill-definedConclusions:Thestudyusedanidiosyncraticandill-defineddefinitionof Ill-definedConclusions :Thestudyusedanidiosyncraticandill-defineddefinitionof ● theterm“filterbubble.”IttreatedanyvariancebetweenSearchresultsastheresultof whatitdefinedasa“filterbubble.”Incontrast,weunderstandthroughourowntesting thatvariousfactorscanimpactSearchresults,including,thelocationoftheuser,the datacenterprocessingtherequest,theexperiment,andsoon.TheGerman















studyisavailableat https://algorithmwatch.org/de/bei-der-google-Suche-Personalisierunggeringer-als-gedacht-hauptsaechlich-regionale-effekte/.. geringer-als-gedacht-hauptsaechlich-regionale-effekte/ 10. Youro Youroffic fficial ialres respon ponses sesto toth theDu eDuckDu ckDuckGo ckGost study udyse seemt emtod odeny enyth that at GoogleusespersonaldatatoinfluencesearchresultsinIncognitomodeandwhen signedoutofGoogle.Iffalse,wouldthisconstituteaviolationofyourConsentOrder withtheFederalTradeCommission? WhenauserissignedoutofGoogle(whetherusingIncognitoModeornot),wedonotuse accountinformation(profileinformationstoredwithauser’saccount)toinfluenceSearch results.Thetrustofourusersisofutmostimportancetous.Wecontinuetocomplywiththe termsofourFederalTradeCommissionConsentOrder. 11. Whatd Whatdoes oesGo Googl ogledo edoto tose search archre resul sults, ts,if ifany anythi thing, ng,to tore respe spect ctDo DoNot Not Trackheaderssentbysomeconsumers'browsers? DoNotTrackrequestsmaynotaffectGoogle’scollectionofinformationbecauseGoogle,like mostwebsitesandwebservices,collectsandusesbrowsingdatatoimprovesecurity,aswell astoprovidecontent,services,ads,recommendationsonitswebsites,andtogenerate reportingstatistics.WeprovideuserswithtoolstomanageGoogle'scollectionofinformation abouttheirsearchorbrowsinghistorythroughGoogle’sWebandAppActivitytool(available athttps://myaccount.google.com/activitycontrols/search?pli=1 athttps://myaccount.google.com/activitycontrols/search?pli=1)). 12. Accordi According ngto toGoo Google gle's 'scalc calculat ulation ions, s,what whatis isyo your ursha share reof ofthe thege gener neral al searchenginemarketintheUnitedStatesondesktop?AccordingtoGoogle's calculations,whatisyourshareofthegeneralsearchenginemarketintheUnitedStates onmobiledevices?WhodoesGoogleconsidertobeitscompetitorsinthegeneral searchenginemarketintheUnitedStates? Weworkinahighlycompetitiveenvironment.“General”searchenginesarejustonepartof thevastarrayofcompetitorsweface.Thetermdoesnotdescribethemanywaysthatusers searchforinformationontheInternet.Somesearchengines,suchasGoogle,are“general,”in thesensethattheytrytoprovidearelevantanswernomatterwhattheuserisseeking.Other sites,oftencalled“vertical,”orspecializedsearchengines,assumethattheuserislookingfor aparticularkindofinformationandprovideonlythatkindofresult.Foranygivencategoryof query,wecompeteagainstawiderangeofcompaniesthatgoeswellbeyondother“general” searchengine searchengineslike slikeBinga BingandYah ndYahoo!. oo!.















Sterling,Report:AmazonGrowsLea Report:AmazonGrowsLeadasProductSea dasProductSearchEngineo rchEngineofChoice fChoice,SearchEngine Land(Sept.28,2016),availableathttp://searchengineland.com/report-amazon-grows- Land(Sept.28,2016),availableat http://searchengineland.com/report-amazon-grows- lead-product-search-engine-choice-259985(showingthatinSeptember2016,55%ofusers lead-product-search-engine-choice-259985 (showingthatinSeptember2016,55%ofusers begantheirshoppingtasksatAmazon,comparedwithjust28%forallsearchengines combined).Similarly,Expediaprovidessearchresultsforuserssearchingforflights,hotels, andcarrentals.Wealsocompetewithsocialnetworks,mobileapps,anddirectnavigationto websites—andnewformsofaccessinginformationarebeingdevelopedeveryday.Insofaras wehaveaneedformarketshareinformation,wetypicallyconsultthird-partysourcessuchas Comscore. 13. Whatp Whatperce ercenta ntage geof ofque queries rieson onGo Googl oglein einth theUn eUnited itedSt States atesar areco econdu nducted cted bypeopleloggedintoGoogle? 14. Whatp Whatperce ercenta ntage geof ofque queries rieson onGo Googl oglein einth theUn eUnited itedSt States atesfo forp rpeop eople le loggedintoGooglearepersonalizedorcontextualizedinsomeway(includingchanging general,non-locallinksbasedonlocation)? 15. Whatp Whatperce ercenta ntage geof ofque queries rieson onGo Googl oglein einth theUn eUnited itedSt States atesfo forp rpeop eople le loggedoutofGooglearepersonalizedorcontextualizedinsomeway(including changinggeneral,non-locallinksbasedonlocation)? Becausetheanswertothesequestionsisrelated,we’vegroupedtogetherourresponseto questions13t questions13through hrough15. 15. GoogleSearchrespondstotrillionsofuserqueriesfromaroundtheworldeveryyear,driven byourcompanymissiontoorganizetheworld’sinformationandmakeituniversallyaccessible anduseful.OfthetrillionsofsearchesGooglerespondstoeveryyear,morethan50% happenonmobile.Furthermore,morethan50%ofoursearchescomefromoutsidethe UnitedStates.Andeveryday,15%ofthequeriesweprocessareoneswe’veneverseen before before.. Indeterminingwhatresultstoshowinresponsetoauserquery,wemayusefactorssuchas thebrowsertype,language,timeofsearch,location,andpriorsessionsearches.Andwhen usersuseSearch,theygeneratedata—likethetermstheysearchfor,thelocationofthe devicetheysearchfrom,andthelinkstheyinteractwith.Google’sMyAccount(availableat https://myaccount.google.com)letsuserscontrolthedatathatissavedtotheirGoogle https://myaccount.google.com )letsuserscontrolthedatathatissavedtotheirGoogle Account. Account.

















TheDataandPersonalizationsectionofMyAccount(availableat https://myaccount.google.com/data-and-personalization)allowsuserstoturnonoroff https://myaccount.google.com/data-and-personalization )allowsuserstoturnonoroff featureslikeWeb&AppActivity,andtooptinoroutofadspersonalization.Inaddition,the PrivacyCheckuptool(availableathttps://myaccount.google.com/privacycheckup PrivacyCheckuptool(availableat https://myaccount.google.com/privacycheckup)makesit )makesit easyforuserstoreviewandchangetheirprivacysettings.Thesetoolsenableuserstomake informeddecisionsabouttheirdata.

















Andourrecentlylaunched“YourDatainSearch”enablesuserstobrowseanddeletetheir Searchactivitydirectlyforthepasthour,orevenentirely,withoutnavigatingawayfromwhat they’redoi they’redoing. ng.

Dependingonthesearchquery,someresultspagesmaychangerapidly,whileothersare morestable.Forexample,whenausersearchesforthelatestscoreofasportsgame,we havetoperformup-to-the-secondupdates.Butresultsaboutahistoricalfiguremayremain staticforyearsatatime.Thesemetricsremainfluid,however,andwehavelimitedpublic metrics.Thatsaid,totheextentwerequiremarketshareinformation,wetypicallyconsult third-partysourceslikeComscore.

















intheU.S.,providealistofthetopthreeorderingsandtheirfrequencyof occurrence. b. Howm Howman any yvar variat iatio ions nso of fth then enews ewsi inf nfob obox ox( (als also okn known ownas as" "To Top p Stories")wouldGoogleshowinaggregate,includingdifferencesintheordering ofthoselinks?Foreachlinkshown,whatisthefrequencyofitsplacement(e.g., nytimesshown50%aslink#1,40%aslink#2,10%notshownatall).Foreachzip codeintheU.S.,providealistofthetopthreeorderingsandtheirfrequency. c. Howm Howman any ydif differ feren ent tpl place aceme ment nts sof ofth then enews ewsin info fobo boxw xwou ould ldGo Goog ogle le showwithinthesearchresultspage,andwhatisthefrequencyofthose placements(e.g.,20%ofthetimethenewsisshownabovelink1,50%abovelink 2,10%notshownatall,etc.)? d. Howm Howman any yvar variat iatio ions nso of fth thev evide ideos osi inf nfob obox oxwo woul uld dGo Goog ogle lesh show ow, , includingdifferencesintheorderingofthoselinks?Foreachlinkshown,whatis thefrequencyofitsplacement(e.g.,CNNvideoshown50%aslink#1,40%aslink #2,10%notshownatall).ForeachzipcodeintheU.S.,providealistofthetop threeorderingsandtheirfrequency. e. Howm Howman any ydif differ feren ent tpl place aceme ment nts sof ofth thev evid ideo eos sin info fobo boxw xwou ould ld Googleshowwithinthesearchresultspage,andwhatisthefrequencyofthose placements(e.g.,20%ofthetimevideosareshownabovelinkI,50%abovelink2, 10%notshownatall,etc.)? f. Ofal Of all lth the ecli clicks cksju just stc com omin ing gfr from omt the heo org rgan anic icli link nks son ont the heG Goo oogl gle e searchresultspage,whatpercentagecomefromclickingonthefirstlink,the secondlink,andsoforth? g.

Ofal Of all lth the ecli clicks cksju just stc com omin ing gfr from omt the hen news ewsin info fobo boxo xon nth the eGo Goog ogle le

















querythatisprocessedwouldhavethepotentialtoimpactlaterqueries.Forinstance,Google usescachestospeedupresponsestofast-spikingqueriesbynothavingtocompletely re-processtheuser’squery.Theanswerstothesequestionswouldalsodependonthe distributionandtimingoftheserequestsacrossourdatacenters.Withthatsaid,we continuouslyworkonensuringthatthemostrelevantresultsaresurfaced.Weare transparentaboutourranking,includingwithrigoroustestingoftherelevanceofour results—weranover270,000testsandupdateslastyearalone,resultingin2,400 improvements—andwithpublicguidelinesdescribinghowweassesstopresults. 17. Suppos Supposea eaper person sonsea search rches esfor for"p "pedia ediatri trician ciansw swash ashing ington tondc dc"o "onG nGoog oogle. le. The"answerbox"providedaGooglemapwiththreepinneddoctors'offices,three linkeddoctors'offices,andatthebottomoftheanswerboxisa"Moreplaces"button. Belowtheanswerboxareanumberofbluelinksto,amongotherthings,Yelp.com, Zocdoc.com,Vitals.com,Healthgrades.com,doctor.webmd.comandspecific pediatricians'offices.Thethreelinkeddoctors'officesintheanswerboxonlyhave ratingsfromGooglereviews.Afterclickingonthe"Moreplaces"button,oneistakento amorecomprehensivemapandalistof22pediatricians'offices.Thoseofficeshave eithernoreviewsoronlyGooglereviews. a. Does DoesG Goo oogl gle's e'sa ans nswer werb box oxal alway ways spr prov ovid ide eth theu euser sert the heb best estr resu esult lts, s, inthiscaseresultsforpediatricians? Ourbusinessmodelandcompanymissiondependonusbeingausefulandtrustworthy sourceofinformationforeveryone.Thequestionreferstoresultsthatreflectanimportant innovationGoogleintroducedin2007called“UniversalSearch.”UniversalSearchresults integrateinformationfromavarietyofspecializedformats—suchasnewsarticles,images, videos,products,localplaces,andmore—sothatusersnolongerhavetoconductmultiple searchesbasedondifferentinformationtypes.Instead,wecompiletheinformationand makeitaccessibletousersenteringasinglesearch.

















ofplace/business,openinghours,precisegeographicallocation,phonenumber,reviews, ratings,andwebsite.Google’sspecializedlocalSearchresultsmakethatinformationeasily accessible.Providingtheseresultsisonewaywemeetourcommitmenttoourusers,and we’llcontinuetoworkhardtoimproveourSearchresults. b. Why Whydo doesn esn't 'tG Goo oogl gle's e'sa ans nswer werb box, ox,o or rth the eres resul ults tsr reac eache hed daft after er clickingthe"Moreplaces"buttonintheanswerbox,containanyinformation poweredbyYelp.com,Zocdoc.com,Vitals.com,Healthgrades.com,or doctor.webmd.com?

Google’sanswerbox,or“LocalUniversal,”useslocalsearchalgorithmsthatdrawresults fromanindexofplaces,notanindexofwebsites(manyofthesmallbusinessesthat appearinGoogle’slocalsearchresultsdonotevenh appearinGoogle’slocalse archresultsdonotevenhaveawebsite).If,withintheanswer aveawebsite).If,withintheanswer boxorLocalUniversal,youclickon“MorePlaces,”andselect,forexample,arestaurant nearyou,thatresultmightcontainreservationlinksfromOpe nearyou,thatresultmightcon tainreservationlinksfromOpenTable,deliverylinks,and nTable,deliverylinks,and third-partyreviewslistedin“Criticreviews.”Forexample,ifauserentersthequery“hot dogsDC”andclicksontheresultforBen’sChiliBowl,theresultco dogsDC”andclicksontheresu ltforBen’sChiliBowl,theresultcontainsorderlinksfrom ntainsorderlinksfrom fivedifferentsitesandthird-partyCriticreviews. fivedifferentsitesandthird-partyCriticrev iews.



















Webmd,andHealthgrades.AndasearchfortheDCseafoodrestaurant“Oceanaire” includesOpenTable,Foursquare,Facebookuserreviews,andZag includesOpenTable,Foursquare ,Facebookuserreviews,andZagatcriticreviews. atcriticreviews.



















helpful,theycan(andwill)useanothergeneralsearchengineorgotoaspecializedvertical searchsite. searchsite. d. Can Canbe bett tter erco cont nten ent texi exist sto out utsi side deof oft the hea ans nswer werbo boxo xor rth the epa page geth that at followsafterclickingthe"Moreplaces"button? Asdiscussedinanswers(a)and(c),usersmayfindwhattheyarelookingforinmanyplaceson GoogleSearch,includingtheanswerbox,byclickingthe"moreplaces"button,orbyrefining theirsearch theirsearch.. e. years?

Howm Howman any ylo local cals sear earch ches eswe were reco cond nduc ucte tedf dfor ort the helas last tfi five vecal calen enda dar r

Asdiscussedinanswers(a)and(c),ourspecializedlocalSearchalgorithmsattemptto determineresultsthataremostlikelytoberelevanttotheuser’squery,andnumerous variablesfactorintothisprocess.Wedonot,however,haveastandarddefinitionforwhat searchesareconsideredlocalsearches,norareweawareofanygenerallyaccepted definitionsthatcanbeeasilyappliedacrossusersearches. f. Does DoesG Goo oogl glep epri rior orit itize izei its tso own wnc con onte tent nto ove ver rot othe her rco cont nten ent twh when en userssearchforlocalgoodsandservices?



















https://techcrunch.com/2 015/02/05/twitter-confirms-new-google-firehose-deal-to-distribute-traffic-to-logged -out-users/ Ourcompanymissionistoorganizetheworld’sinformationandmakeituniversallyaccessible anduseful.GooglecurrentlyhasapartnershipwithTwitterwherebyTwittermakesTweets availabletoGoogle.ThisallowsouruserstohaveaccesstocontentfromTwitterinnearlyreal time,whichwebelievemakesourresultsmoreuseful.Nothingaboutthedealimpacts whetherornottweetswillsurfaceinresponsetoaspecificquery.Wearealwayslookingto giveourusersthemostusefulandaccurateinformationavailable,butdonotcurrentlyhave agreementsliketheonewithTwitterwithothercompanies. AccordingtoGoogle'sTransparencyreports,itsremoved10itemsattherequest oftheIraniangovernmentsince2013.Couldyoupleaselisteachpiecesofcontentthat wasremovedandthereasonfordoingso. Googlereceivescontentremovalrequestsfrommanygovernmentsaroundtheworldevery yearandreviewsthecontentforpotentialproductpolicyviolations.OurTransparencyReport (availableathttps://transparencyreport.google.com/government-removals/overview (availableathttps://transparencyreport.google.com/government-removals/overview), ), shows atotalof3requestsfromIranandremovalofatotalof10individualpiecesofcontentacross Googleproductsandservices.Wereceivedtheserequestsin2012and2016abouttwo BloggerblogpostsandeightYouTubevideos,allofwhichwereviewedagainsteachproduct’s



















YoucanreadourYouTubeCommunityGuidelinesat https://www.youtube.com/yt/about YoucanreadourYouTubeCommunityGuidelinesathttps://www.youtube.com/yt/about /policies/#community-guidelines.. /policies/#community-guidelines Google'stransparencyreportliststhirdpartyorganization,whichthecompany fund funds. s. a) Jona Jonah hGo Gold ldber berg gre recen centl tly ydis discl clos osed edth that atth theN eNat atio iona nal lRev Review iew InstitutereceivedfundsfromGoogle,buttheInstituteisnotlistedonthereport. Why? Googlehasalonghistoryofsupportingorganizationsonallsidesofthepoliticalspectrum. GooglewasamongseveralcorporatesponsorsoftheNationalReviewInstitute’sWilliamF. BuckleyPrizedinnerin2018,whichisscheduledtobereportedinourupcomingtransparency report(availableathttps://www.google.com/publicpolicy/transparency.html report(availableathttps://www.google.com/publicpolicy/transparency.html). ). b) Does Doest the hed dis isclo closu sure rein inclu clude deor orga gani nizat zatio ions nsw whic hich hre recei ceive vefu fund ndin ing g fromGoogle'slobbyists,publicaffairsconsultants,tradeassociationsorother intermediaries? Google’sPublicPolicyTransparencyReport,(availableat https://www.google.com Google’sPublicPolicyTransparencyReport,(availableathttps://www.google.com /publicpolicy/transparency.html),makesquarterlydisclosuresregardingourfederallobbying, /publicpolicy/transparency.html ),makesquarterlydisclosuresregardingourfederallobbying,



















BroadbandnetworksandInternetapplicationshavedifferentroles,andconsumersand businessesusethemfordifferentpurposes.Fortheirpart,networksaregatekeepersforthe entireonlineexperience.GeneralaccesstotheInternetiswhatISPspromiseandwhat consumerspayfor. Incontrast,theroleofinternetapplicationsandwebsitesisverydifferenttothatof broadbandnetworks.WithInternetapplications,consumersmoveeasilybetweenvarious onlineapplicationsandservices,eachofwhichidentifiesandorganizesinformationtomakeit morerelevantandusefulfortheparticularuser.Theusefulnessofthesesites—including GoogleSearch—dependsontheirabilitytoreturndifferentiatedresultsthatrespondtothe specificuser’spreferences.Rankinghigherquality,morerelevantsitesoverlowerquality,less relevantsitesiswhatmakesSearchresultsuseful.Andunlikenetworktrafficneutrality,there isnoobjectivetechnicalcriteriaoragreementonwhat"neutral"searchresultswouldlooklike (orwhatismostrelevanttoausers'query). Wewillcontinuetocompetevigorouslytoprovideouruserswiththemostaccurateand usefulinformation,withtheknowledgethatadifferentwebsiteorapplicationisonlyoneclick away. "Google'ssearchrankingalgorithmsmaywellsimplyreflectthisalready-existing mediabias.WouldyouconsiderchangingGoogle'salgorithmstocountertheeffectsof theideologicalbiasalreadyinherentinthenewsmediaindustry?"





















1) Whati Whatist sthea heaver verage agenum number berof ofsear search cheng engine ineres result ultsp sper erday dayfo fort rthe helas last t tencalendaryears2008-2018? Whilewedonothavemoredetailedpublicmetrics,wecanconfirmthateveryyear,thereare trillionsofsearchesonGoogle.Overthepast20years,Googlehashelpedtobringmoreof theworld’sinformationonlineandmakeitaccessiblewithSearch,images,videos,local information,andmore.We’vecomealongwaysincetheearlydaysofGoogle,butwenever thinkofsearchasasolvedproblem.PeoplehavegreaterexpectationsforSearchtodaythan theydid20yearsago,andwewelcomethat.Itpushesustoimaginewhatwecandonext, andhowwecanbetterconnectpeoplewithinformationabouttheworldaroundthem. 2)

What Whatp per ercen centa tage geo of fth thos ose esea searc rch hen engi gine ner resu esult ltp pag ages esr res esul ulte tedi din nth the eus user er: :



















thatdataprocessingoccursinamannerthatiscompatiblewithindividuals’interestsand socialbenefits,andminimizestheriskofharmtousers.Wedescribeourframeworkfor responsibledataprotectionathttps://services.google.com/fh/files/blogs/google_framework_ responsibledataprotectionat https://services.google.com/fh/files/blogs/google_framework_ responsible_data_protection_regulation.pdf.. responsible_data_protection_regulation.pdf b. Pleas Pleased edet etail aili in npa part rtic icul ular arh how owt thi his spo polic licy yaff affect ects sbo both thy you our r collectionandyourretentionofdata,regardlessofthechoiceofsettingsmade bytheuserabouteithercollectionorretention. AtGoogle,wedesignourproductstocollectandusedatacommensuratewiththeirpurpose andtokeepthatdatanolongerthannecessary.Theseprinciplesarereflectedinallofour products.Forexample,mostofourlogsareonlykeptforaperiodofdaysorweeks,andall ourlogsaresubjecttodatadeletionandanonymizationpolicies.Bywayoffurtherexample,



















GoogleoffersproductsthatuserscansignintowiththeirGoogleAccountidentification (oftenaGmailemailaddressorphonenumber)andauthenticationcredentials(password). Whenusersaresignedin,theyenjoythebenefitofaccessingandcontrollinginformationin theiraccount.Forexample,onefeatureofGoogleAccountsisacontrolnamed“Download YourData”(formerlyknownas“Takeout”),whichenablessigned-inuserstodownloador exporttheirdatafromtheGoogleproductstheyuse.Becausewesupportuserchoice,we believeit’simportanttogiveusersthefreedomtoporttheirdatatocompetingservicesif theysochoose.DownloadYourDataisavailabletouserswhosignintotheirGoogleAccount. Userswhoarenotsignedinhaveotherwaystocontrolthedatawecollectabouttheiruseof ourservicesanddevices—forexample,byclearingbrowsingdataorcookiedata.Finally,and asdescribedabove,webelievethatreasonablelimitationsshouldbeplacedonthemanner andmeansofcollecting,using,anddisclosinginformation.Andweworkhardtoensurethat dataprocessingoccursinamannerthatiscompatiblewithindividuals’interestsandsocial



















troubleshootbugsordeveloperissuesorinvestigateabuseandscraping.Additionally,when non-GoogleappsuseAndroid’slocationAPIs,andtheusergrantspermission,theinformation alsogoestothenon-Googledeveloperpursuanttoitsownuserprivacypoliciesand functionality. Google’sLocationHistoryfeatureisadistinctGoogleproductthatusersmustoptintoturn onanduse.ForuserswhohaveenabledLocationHistoryfortheirGoogleAccount,andhave theirdevicereportingtoLocationHistory(userscanblockanydevicefromreportingto LocationHistory),thedevicesendslocationdatabacktoGoogletostoreinLocationHistory. GoogleLocationHistoryisnotdesignedtocollectorstorelocationinformationfromany non-Googleapps.Wearealwaysworkingtoprovideusersgreaterchoice,transparency,and control,andwillcontinuetodothat.



















AsdescribedinourTransparencyReport(availableat https://transparencyreport.google AsdescribedinourTransparencyReport(availableathttps://transparencyreport.google .com/user-data/overview?hl=en),Googleroutinelyrespondstorequestsfromlaw .com/user-data/overview?hl=en ),Googleroutinelyrespondstorequestsfromlaw enforcementagencies.Weareabletodosonotwithstandingencryptioninvirtuallyallcases. Weunderstandtheimportanceoftheworkoflawenforcementandwillcontinuetopromptly respondtolawfulprocess.Asweevolve,wearecontinuingtolookhardattheseissuesto ensurewearestrikingtheappropriatebalance. b. Woul Would dGo Goog ogle lebe bewi will llin ing gto toco come meto tot the heta tabl blea eand ndw wor orkw kwit ith hlaw law enforcementonaworkablesolutiontothischallenge? Googlehasengagedwithmanystakeholders,includinglawenforcement,onthisissueandis happytocont happytocontinueo inueourprodu urproductive ctivedialog dialogue. ue.





















TheupdatedUserDataPolicypermitsaccesstousers’Gmailaccountsforappsdirectly enhancingemailfunctionality—forexampleemailclients,emailbackupservices,and productivityserviceslikeCRMandmailmergeservices.Thisdecisionisinlinewithconsumer expectations.Wefoundinourreviewofusers’expectationsthatthevastmajorityofusers



















harmsusersandiscontrarytothecompetitiveprocess.Weaimtoprovideourusersabroad rangeofinformationandchoicesandtofosteradiverseInternetecosystem. 4. The TheCh Chin ines eseg egov over ernm nment ento oft ften enex expl ploi oits tsfo forei reign gnc cor orpo pora rati tion ons sby bym maki aking ng accesstotheChinesemarketcontingentonforcedtransfersofitsproprietary



















QFRsfromRepresentativeKenBuck Q:Mr.Pichai,itisclearthatcounteringfalseinformation,misinformation,and disinformationontheinternetcan'tbedonebyartificialintelligence-it'sjusttooeasy



















Althoughweinvestheavilytoaddressmisinformation,wealsowelcometheeffortsofothers seekingtocreatesolutions.Wewillcontinuetoevaluateourefforts,andtheeffortsofother organizations,tohelpimproveourproducts.



















ThetypesofdataGooglecollectsorstoresmaybedifferentforusersbasedonvarious settingstheuserhasselectedandwhatproductstheyuse.Butwemakeiteasyforusersto reviewtheirdata,changetheirsettings,anddeletedata.Googlewasoneofthefirst

























































































































































































































































Google 12.11.18 Hearing QFR Responses

Recommend Documents