T H E T R A D E O F F F A L L A C Y H o wm a r k e e r s A r e M s r e p r e s e n n g A m e r c a n C o n s u m e r s A n d O p e n n g T h e mu p o E x p l o a o n Ourfind ngs s uppor ane we x pl a na on:ama j or yof Ame r c a nsa r er e s gne d og v ngup he rda a a nd ha s whyma nya ppe a r obee nga ge d n r a de offs. ( Se epa ge3 ) JosephTur ow Annenber gschoolf orcommun ca on Un ver s yofpennsyl van a M chaelhennessy Annenber gpubl cpol cycen er Un ver s yofpennsyl van a Nor adr aper Depar menofcommun ca on Un ver s yofnew Hampsh r e ARepor f r om heannenber gsc hoolf orcommun c a on Un ver s yofpenns yl van a
Joseph Turow, Ph.D., s Rober Lews Shayon Professor of Communcaon a he Annenberg School for Communcaon, Unversy of Pennsylvana. Among hs several books are The Daly You: How he New Adversng Indusry s Defnng Your Ideny and Your Worh (Yale Unversy Press, 2011), Nche Envy: Markeng Dscrmnaon n he Dgal Age (MIT Press, 2006) and Breakng Up Amerca: Adversers and he New Meda World (Unversy of Chcago Press, 1997). Snce 1999 he has conduced naonal elephone surveys ha have moved forward publc dscourse on dgal meda, markeng, and prvacy. See hp://papers.ssrn.com/sol3/papers.cfm?absrac_d=2423753 Mchael Hennessy, Ph.D., s projec manager and sascan n he Healh Communcaon Group of he Annenberg Publc Polcy Cener. He has publshed over 120 peer-revewed arcles n journals as AIDS & Behavor, Evaluaon Revew, Healh Educaon and Behavor, Journal of Sex Research, Srucural Equaon Modelng, Meda Psychology, Psychology Healh & Medcne, and he Amercan Journal of Publc Healh. Dr. Hennessy was prevously employed by he Ceners for Dsease Conrol, has held faculy posons a he Unversy of Hawa and Emory Unversy, and has worked n he prvae secor dong evaluaon research. Nora Draper, Ph.D., s an Asssan Professor n he Deparmen of Communcaon a he Unversy of New Hampshre. Her research examnes he complexes of auhency, prvacy, deny, and repuaon n he dgal era hrough frames of culural heory, crcal nsuonalsm, and publc polcy. Her work explores how deny, parcularly gender, race, class, and sexualy, shape experences of prvacy, survellance, and vsbly n a dgal envronmen. Her work has been publshed n he Inernaonal Journal of Communcaon, Crcal Sudes n Meda Communcaon, he Journal of Chldren and Meda, and Survellance & Socey. June, 2015 Thanks o Penn-Annenberg docoral sudens Ope Akanb, Emly Hund, Elena Mars, Lee McGugan, and Sephen Schrag for her help durng he quesonnare-consrucon phase of hs projec. Ths survey was suppored by The Annenberg School for Communcaon Mchael Dell Carpn, Dean. Turow, Hennessy, Draper Page 2 of 24
Overvew New Annenberg survey resuls ndcae ha markeers are msrepresenng a large majory of Amercans by clamng ha Amercans gve ou nformaon abou hemselves as a radeoff for benefs hey receve. To he conrary, he survey reveals mos Amercans do no beleve ha daa for dscouns s a square deal. The fndngs also sugges, n conras o oher academcs clams, ha Amercans wllngness o provde personal nformaon o markeers canno be explaned by he publc s poor knowledge of he ns and ous of dgal commerce. In fac, people who know more abou ways markeers can use her personal nformaon are more lkely raher han less lkely o accep dscouns n exchange for daa when presened wh a real-lfe scenaro. Our fndngs, nsead, suppor a new explanaon: a majory of Amercans are resgned o gvng up her daa and ha s why many appear o be engagng n radeoffs. Resgnaon occurs when a person beleves an undesrable oucome s nevable and feels powerless o sop. Raher han feelng able o make choces, Amercans beleve s fule o manage wha companes can learn abou hem. Our sudy reveals ha more han half do no wan o lose conrol over her nformaon bu also beleve hs loss of conrol has already happened. By msrepresenng he Amercan people and champonng he radeoff argumen, markeers gve polcymakers false jusfcaons for allowng he collecon and use of all knds of consumer daa ofen n ways ha he publc fnd objeconable. Moreover, he fuly we found, combned wh a broad publc fear abou wha companes can do wh he daa, porends serous dffcules no jus for ndvduals bu also over me for he nsuon of consumer commerce. Markeers jusfy her daa-collecon pracces wh he noon of radeoffs, depcng an nformed publc ha undersands he opporunes and coss of gvng up s daa and makes he posve decson o do so. A 2014 Yahoo repor, for example, concluded ha onlne Amercans demonsrae a wllngness o share nformaon, as more consumers begn o recognze he value and self-benef of allowng adversers o use her daa n he rgh way. 1 Ths mage of a powerful consumer has become a way o clam o polcymakers and he meda ha Amercans accep wdespread rackng of her backgrounds, behavors, and lfesyles across devces, even hough surveys repeaedly show hey objec o hese acves. Our sudy challenges he asseron ha radeoffs explans wha mos Amercans are dong. Wh he help of Prnceon Survey Research Assocaes Inernaonal, we conduced a represenave naonal cell phone and wrelne phone survey of 1,506 Amercans age 18 and older who use he nerne or emal a leas occasonally. We presened hem wh everyday crcumsances where markeers collec people s daa, phrasng he suaons as radeoffs and learned ha very many feel hose radeoffs are unfar. 91% dsagree (77% of hem srongly) ha If companes gve me a dscoun, s a far exchange for hem o collec nformaon abou me whou my knowng. Turow, Hennessy, Draper Page 3 of 24
71% dsagree (53% of hem srongly) ha I s far for an onlne or physcal sore o monor wha I m dong onlne when I m here, n exchange for leng me use he sore s wreless nerne, or W-F, whou charge. 55% dsagree (38% of hem srongly) ha I s okay f a sore where I shop uses nformaon has abou me o creae a pcure of me ha mproves he servces hey provde for me. Furher analyss of hese responses ndcae only a very small percenage of Amercans agree wh he overall concep of radeoffs. In fac, only abou 4% agree or agree srongly wh all hree proposons. If we use a broader defnon of a belef n radeoffs he average value of all hree saemens even hen only 21% of he respondens accep he dea. Ye, when we presen a real-lfe radeoff case askng Amercans wheher hey would ake dscouns n exchange for allowng her supermarke o collec nformaon abou her grocery purchases 43%, or more han wce as many as n he broader defnon of radeoff supporers, say yes o radeoffs. Underscorng he nconssency, we found ha 40% of he people who sad hey would accep he grocery-dscoun deal do no agree wh he hrd radeoff saemen above even hough he ype of exchange suggess s smlar. Wha s gong on? Conrary o he clam ha a majory of Amercans consen o dscouns because he commercal benefs are worh he coss, our sudy suggess a new explanaon for wha has hus far been msconsrued as radeoff behavor n he dgal world: a large pool of Amercans feel resgned o he nevably of survellance and he power of markeers o harves her daa. People who are resgned do no predcably decde o gve up her daa. We acually found no sascal relaonshp beween beng resgned o markeers use of daa and accepng or rejecng varous knds of supermarke dscouns. Raher, he larger percenages of people n he populaon who are resgned compared o people who beleve n prncple ha radeoffs are a far deal ndcae ha n he real world people who gve up her daa are more lkely o do whle resgned raher han as he resul of cos-benef analyss. Markeers would have all hese behavors caegorzed as a raonal accepance of radeoffs. Bu when we looked n our scenaro a he people who agreed o gve up her daa for supermarke dscouns, we found ha 57% of hose who ook he deal were resgned, whle even usng he broader undersandng of radeoff suppor only 32% were radeoff supporers. To furher queson markeers emphass on Amercans use of cos-benef calculaons, we found ha large percenages of Amercans ofen don have he basc knowledge o make nformed cos-benef choces abou ways markeers use her nformaon. Renforcng hese naccuraces, large percenages of Amercans overesmae he exen o whch he governmen proecs hem from dscrmnaory prcng companes changng prces from person o person based on hose ndvduals consumer profles. For example: 49% of Amercan aduls who use he nerne beleve (ncorrecly) ha by law a supermarke mus oban a person s permsson before sellng nformaon abou ha person s food purchases o oher companes. 69% do no know ha a pharmacy does no legally need a person s permsson o sell nformaon abou he over-he-couner drugs ha person buys. 65% do no know ha he saemen When a webse has a prvacy polcy, means he se wll no share my nformaon wh oher webses and companes whou my permsson s false. Turow, Hennessy, Draper Page 4 of 24
55% do no know s legal for an onlne sore o charge dfferen people dfferen prces a he same me of day. 62% do no know s legal for an offlne or physcal sore o charge dfferen people dfferen prces a he same me of day. 62% do no know ha prce-comparson ses lke Expeda or Orbz are no legally requred o nclude he lowes ravel prces. These msconcepons sugges ha even when Amercans do wegh he coss and benefs of gvng up her daa, hey frequenly base hose choces on ncorrec nformaon. Ironcally, and conrary o many academc clams abou he reason people gve up her nformaon, hose who know he mos abou hese markeng pracces are more lkely o be resgned. Moreover, resgned people s decsons o accep supermarke dscouns even when he supermarke collecs ncreasngly personal nformaon are also posvely relaed o knowledge. When comes o proecng her personal daa, our survey fnds hose wh he wherewhal o accuraely calculae he coss and benefs of prvacy are lkely o consder her effors fule. As U.S. socey moves furher no he weny-frs cenury, personalzed deals and prces wll become common boh onlne and n physcal sores. These sors of alored offers wll undoubedly be ncreasngly roublng o Amercans who beleve hey are on he losng end of ofen-hdden consumer profles and argeng formulas. Suspcon already abounds: 72% of Amercans rejec he dea ha wha companes know abou me from my behavor onlne canno hur me. I s no dffcul o predc wdespread socal ensons, and concerns abou democrac access o he markeplace, f Amercans connue o be resgned o a lack of conrol over how, when, and wha markeers learn abou hem. A he end of hs repor, we offer suggesons for urnng hs rend around and movng from a socey where people have lle hope for conrol over her everyday daa o one n whch he expecaon and realy of conrol s par of ransparen and responsble commercal ransacons. Background Publc opnon polls repeaedly fnd ha consumers are concerned abou ways markeers access and use her daa onlne and offlne. A 2005 naonal publc opnon elephone survey from he Unversy of Pennsylvana s Annenberg School found 66% of nerne-usng aduls dd no agree I s okay wh me f supermarke I shop a keeps dealed records of my buyng behavor. 2 Close o a decade laer, a 2014 Pew sudy found 80% of socal nework se users n a naonally represenave onlne poll of adul Amercans answered hey were concerned ha some of he nformaon you share on socal neworkng ses mgh be accessed by hrd pares lke adversers or busnesses, whou your knowledge. 3 And n 2015 he Ban & Company consulancy repored on a survey conducng of more han 900 consumers ha found ha hey are ofen uncomforable wh how her daa s used and shared. A full wo-hrds feel ha should be llegal for companes o collec or use such daa whou geng pror consen. 4 A he same me, academc and busness observers concur ha people ofen release nformaon abou hemselves n ways ha sugges much less concern abou dsclosure and collecon of her personal daa han he polls depc. Some have ermed hs gap he prvacy paradox. 5 I s he dea ha, as reporer Brad Sone wroe n The New York Tmes, normally sane people have nconssen and conradcory Turow, Hennessy, Draper Page 5 of 24
mpulses and opnons when comes o her safeguardng her own prvae nformaon. 6 So, for example, a 2015 Accenure onlne survey found 60 percen of shoppers wan real-me promoons, whch ofen depend on sophscaed commercal survellance. Ye, he same survey found, only 20 percen wan realers o know her locaon and jus 14 percen are comforable sharng her browsng hsory. 7 Some observers read hs paradox as evdence ha people don acually care abou prvacy or, alernavely, ha hey prvlege oher hngs convenence, dscouns, socal connecons above prvacy. 8 Markeers have embraced hs seemng conradcon as a way o argue n favor of wderangng daa collecon. The McCann Worldwde adversng nework s Truh Cenral projec derved hs concluson from a global research sudy surveyng over 10,000 people n eleven counres, ncludng he U.S. No breakng down he resuls by counry or dealng he survey mehod, McCann saed ha whle 71% worry abou he amoun onlne sores know abou hem, 65% are wllng o share her daa as long as hey undersand he benefs for hem. 9 The edor for mcommercedaly nerpreed he fndngs o mean ha he rackng of consumers all comes down o he rade-off n value. 10 Along he same lnes, he presden and chef sraegy offcer of Mobquy, a moble-sraegy consulancy, wroe n 2012 ha he average person s more han wllng o share her nformaon wh companes f hese organzaons see he overall gan for end-users as a goal, no jus for hemselves. 11 A May 2014 repor by Yahoo Adversng followed hs logc n nerpreng s survey of 6,000 respondens ages 13-64, a represenave sample of he U.S. onlne populaon. 12 I hghlghed he fndng ha Roughly wo-hrds of consumers fnd accepable or are neural o markeers usng onlne behavor or nformaon o craf beer ads. Dgally conneced Amercans, he sudy concluded, demonsrae a wllngness o share nformaon, as more consumers begn o recognze he value and self-benef of allowng adversers o use her daa n he rgh way. A few corporae voces have pu cauons around such generalzaons. An Accenure execuve sad If realers approach and marke personalzaon as a value exchange, and are ransparen n how he daa wll be used, consumers wll lkely be more wllng o engage and rade her personal daa. 13 Ban nerpreed he resuls of s survey o mean ha cusomers rus canno be bough by companes offerng compensaon n exchange for sellng or sharng personal daa. And he Brand Bond Loyaly consulng frm suggesed he mporance of greaer ransparency of daa polces and pracces ha are communcaed clearly and concsely o consumers and [loyaly] program members. 14 Ye none of hese companes explaned wha exacly hey mean by ransparency; nor dd hey gve examples of how ransparency could be bul no adversng sysems ha rely on he capure and analyss of personal nformaon. Ban answered s clens crucal queson how can companes acqure cusomer daa whle buldng cusomer loyaly a he same me? by answerng vaguely Ask permsson. Brand Bond focused s repor on s more cenral opc: emphaszng how knowledge of personal nformaon abou cusomers allows for personalzed messages and offers whch, n urn, are lkely o acheve ncremenal busness resuls. These resuls nclude, clamed, cusomers beng hree mes as lkely o spend more wh he brand, less prce-sensve, and four mes as lkely o buy somehng hey don wan or need n order o earn pons or manan her program saus. 15 Turow, Hennessy, Draper Page 6 of 24
Lke he Brand Bond repor, Yahoo Adversng s underscored ha Ths concep of value exchange for personal daa s sarng o come o lfe hrough personalzaon. 16 In fac, for an ncreasng number of markeers, from fnance companes o realers, personalzaon s he nex sep n sophscaed argeng. Yahoo noes s personalzaon echnology racks ndvduals onlne behavors (her clck hsory ) and hen facors n onlne acvy ypcal of [ha person s] demographc cohor. 17 Sascal consulans, meda-buyng agences, major adversers, and major real esablshmens have acceped hs basc approach and adaped o her needs. The common goal nvolves predcve analycs he manpulaon of huge amouns of daa abou ndvduals o assess he lkelhood ha a parcular knd of conen or produc wll lead hem ac n a drecon desred by he markeer. The daa for hese new forms of argeng can come from sources ha may be obvous o he reader, vewer, or cusomer: regsraon on a se, requess for nformaon, or perhaps he use of coupons. Increasngly, hough, markeers are collecng hundreds of daa pons, accurae and no, abou ndvduals n ways ha are far more hdden from vew. Read a ypcal real chan s prvacy polcy, and you may well fnd ha n he neres of personalzaon he merchan buys nformaon from daabase frms o add o s daa abou ndvdual cusomers; uses cookes and smlar elemens o follow people s behavors on ses and n emals as well as o nfer her neress across ses; allows adversng neworks works wh o collec daa abou ndvduals comng o he se; racks people s locaon on her moble devces whou explc permsson unless hey urn ha feaure off; collecs a person s commens from a produc revew, queson, or oher nformaon on s webse, app or socal-nework pages; and allows buons from Facebook, Twer, Pneres, Youube, FourSquare and oher socal-meda ses ha collec nformaon abou ndvduals and place rackers on her compuers. To add furher o her sores of knowledge, ncreasng numbers of physcal sores nsall ools ha use W-F, Blueooh Low Energy, ulrasonc and oher echnologes o denfy ndvdual cusomers presence, rack her movemens, offer hem personalzed coupons or oher messages, and lnk ha nformaon o everyhng else hey know abou hem. Markeers enhuse over he dea ha people s accepance of he general dea of radeoffs jusfes markeers collecon of enough daa pons abou consumers o lead o he knd of personalzaon Yahoo calls he pahway o adversng nrvana. 18 Prvacy advocaes and researchers, by conras, have challenged he asseron ha people make raonal calculaons when hey op-n o prvacyarophyng servces. These groups argue ha decsons abou nformaon dsclosure are no as prone o conradcons beween her expressed audes and her acons as he prvacy paradox would sugges. Insead hey have argued ha ndvduals wan o manage her daa bu lack he necessary knowledge o make nformed decsons abou he consequences of enerng no neracons wh markeers. 19 From surveys, conen analyses, and expermens academc researchers have argued he publc s largely unable o undersand wha markeers are dong wh her daa behnd he compuer screen. For example, egh Annenberg School surveys from 1999 hrough 2012 have conssenly shown ha whle Amercans are que aware ha markeers are collecng daa abou hem and are wary of ha, hey have lle undersandng abou how ha akes place or abou he laws ha govern hose acves. 20 In response, Turow, Hennessy, Draper Page 7 of 24
markeers as well as some polcymakers have exhored people o read he prvacy polces o fnd ou wha parcular ses and apps do wh her daa. Ye even some lawyers who wre he polces for large companes have acknowledged o Joseph Turow ha he documens are legal ender no desgned o be undersood by ordnary people. Indeed, n 2012 Lorre Fah Cranor and Aleeca McDonald deermned ha f nerne users read every prvacy polcy hey encounered onlne hey would spend 25 days a year engaged n hs acvy. 21 Bu before ryng o read prvacy polces, Amercans mus have a realsc reason for clckng on hem. Unforunaely, Annenberg surveys have repeaedly shown ha large percenages of he populaon consumers are msled by he very erm self. Illusrave s he fndng of a 2009 Annenberg sudy ha found 78 percen of respondens dd no know he correc response false o he saemen If a webse has a prvacy polcy, means ha he se canno share nformaon abou you wh oher companes, unless you gve he webse your permsson. 22 I s clear ha he very erm prvacy polcy gves people an mpresson of markeers daa sharng ha expers agree s ofen que unwarraned by he acual exs hemselves. Some crcs ress he cos-benef explanaon wh explanaons of he publc s gnorance ha emphasze he dffculy of undersandng he echnologcal and nsuonal sysem. Allessandro Acqus, Curs Taylor, and Lad Wagman, for example, noe ha new search engnes, socal neworks, e-commerce webses, web browsers, and ndvdualzed conrols for prvacy-conscous consumers have emerged. 23 They conclude, however, ha Whle overall, hese echnologes seemngly leave prvacy choces n he hands of consumers, many (f no mos) consumers, n pracce, lack he awareness and echncal sophscaon requred o proec and regulae he mulple dmensons of her personal nformaon. Smlarly, Gordon Hull emphaszes ha users do no and canno plausbly be expeced o know enough neher abou he uses o whch her nformaon mgh be pu, nor abou he specfc benefs and harms ha mgh resul from hose uses, nor abou he lkelhood ha such harms mgh resul for consen o be meanngful, especally f one makes he assumpon ha hose users are followng a rsk/benef model of economc raonaly. 24 Noe ha he noon ha people s lack of knowledge leaves hem vulnerable o gvng up her nformaon every day o adversers and realers has no moved regulaors n he Uned Saes o acon. Wh he excepon of ceran forms of healh and fnancal nformaon where poenal harms can be denfed and quanfed, and barrng chldren under 13 years old, companes connue o be free o collec huge sreams of ndvduals daa. They are also free o consruc profles of ndvduals wh he daa, o share he daa, and o rea people dfferenly n he markeplace based on conclusons drawn from hose hdden survellance acves. Markeers, for her par, have no been dssuaded by he knowledge-falure argumen from connung o sress he cos-benef approach as he reason people parcpae n neracons ha nvolve he dsclosure of her personal nformaon. Boh posons rase mporan quesons for nvesgaon. The research on how lle people know abou wha happens behnd he screen leads us o dsagree ha he publc s accepance of raonal cos-benef calculaons explans he ease wh whch daa realers and adversers rereve nformaon from ndvduals hough hs proposon has never been drecly esed. A he same me, we fear ha he mos common alernave argumen ha he publc allows markeers s daa ou of knowledge falures ges markeers off he hook oo easly. I perms hem o blame schools and he meda for no eachng people wha hey need o know abou wha happens behnd her compuer screens. I les hem Turow, Hennessy, Draper Page 8 of 24
pon o onlne vdeos explanng wha hey do as well as o programs such as Ad Choces ha allow people o op ou of argeed (or neres based ) adversng. And also allows hem o sound more opmsc abou he publc, o arbue o more awareness and power, han hose who allegedly dengrae he value-exchange prncple markeers celebrae: gve he consumer more of wha hey care abou and less of wha hey don. I s abou surfacng value for users and adversers alke. 25 Concerned abou hese drawbacks, we began o consder an alernave undersandng of he prvacy paradox ha s abou far more han people s cos-benef analyss or her lack of knowledge. I s abou czens belef hey have no agency n a cenral area of democrac socey: commerce. More han a lle anecdoal evdence suggesed o us ha people are makng conscous decsons ha s fule o even ry o cope wh he daa de around hem. People we mee have decded hey serously canno learn hngs, change governmen or busness polcy, or do anyhng else ha wll allow hem o manage her personal nformaon he way hey wan. Moreover, hey feel ha hey would face sgnfcan socal and economc penales f hey were o op-ou of all he servces of a modern economy ha rely on an exchange of conen for daa. They have sld no resgnaon a sense ha ha whle hey wan conrol over her daa world hey wll never acheve. Frusraon over lack of knowledge may have spurred hs sense of resgnaon n many cases. Ye we have noed ha even people que knowledgeable abou how he daa-world works beleve hey have lle realsc alernave o leng her daa flow. They feel caugh n he daa-gaherng markeng sysem for economc, socal, and me-relaed reasons. They know hey wll ge good prces only f hey carry her supermarke s loyaly card; hey beleve hey don have he me, knowledge, or energy o learn how o use anonymzers; her frends are neracng on Facebook and Insagram; and hey know ha erasng cookes ncreasngly doesn sop adversers from rackng hem, anyway. The suaon echoes wha Mark Andrejevc descrbes as he dgal enclosure: he crcumsance when he conemporary dgal real economy compels consumers o dsclose personal nformaon or suffer negave consequences he slower servce, lmed access o dscouns, and conen rrelevan o her neress ha resul from opng ou. 26 Someone observng he behavors of consumers mgh beleve people are ndulgng n radeoffs when hey are acually resgned o gvng up daa. There need no be any dfferences beween how ndvduals who are resgned and hose who beleve n radeoffs ac n markeng suaons. Our goal n conducng hs survey was o look behnd he acves o examne people s belefs abou wha hey are dong and why. How broadly does a sense of resgnaon exs among Amercans? How broadly does a belef n he appropraeness of radeoffs exs among Amercans? How does people s knowledge abou everyday daa gaherng acves of markeers and he laws ha govern hem relae o her sense of resgnaon or belef n radeoffs? And how do knowledge and resgnaon relae o people s decsons n a realsc supermarke scenaro ha offers dscoun coupons for daa? The Sudy and Is Populaon We explored hese quesons as par of a survey of Amercans opnons abou and undersandng of a varey of onlne and offlne prvacy ssues. We cas our populaon ne broadly. We ncluded people n our sudy f hey were 18 years or older and sad yes o one of he followng quesons: Do you go onlne or use he nerne, a leas occasonally? and Do you send or receve emal, a leas Turow, Hennessy, Draper Page 9 of 24
occasonally? The survey quesons ncluded n hs repor focus on four areas. One explores Amercans audes regardng he farness of hree forms of markeng radeoffs. Anoher explores Amercans belefs regardng her ably o know abou and exer conrol over he daa markeers collec abou hem onlne. A hrd nvesgaes people s knowledge of rules of he markeplace when comes o sharng nformaon n he onlne and offlne world as well as o change prces for ndvduals based on locaon, me of day, and an ndvdual s characerscs. A fourh explores wheher Amercans would respond o a specfc scenaro n whch he supermarke hey go o mos ofen wll gve hem dscouns n exchange for s collecng nformaon abou all your grocery purchases. The survey was conduced from February 23 o March 15, 2015 by Prnceon Survey Research Assocaes Inernaonal. PSRAI conduced elephone nervews wh a naonally represenave, Englsh or Spansh speakng sample of 1,506 adul nerne users lvng n he connenal Uned Saes. A combnaon of landlne (n=750) and wreless (n=756, ncludng 412 whou a landlne phone) random dg dal (RDD) samples was used o represen all aduls n he connenal Uned Saes who have access o eher a landlne or cellular elephone. The nervews averaged 20 mnues. Based on a 7- callback procedure and usng he Amercan Assocaon of Publc Opnon research (AAPOR) RR3 mehod, a sandard for hs ype of survey, he overall response raes were a raher ypcal 10 percen for he landlne sample and 10 percen for he cellular sample. Sascal resuls are weghed o correc known demographc dscrepances. 1 The margn of samplng error for he complee se of weghed daa s ±2.9 percen a he 95% confdence level. The margn of error s hgher for smaller subgroups whn he sample. Table 1 provdes an nroducory snapsho of he populaon we nervewed. As Table 1 ndcaes, women slghly ounumber men; 76% desgnae hemselves as Whe, 11% denfy hemselves as Blacks or Afrcan Amercan, Asans or Pacfc Islanders comprse 4%, Nave Amercans comprse abou 1%, and mxed race makes up 3%. Hspancs (Whe and Black) comprse abou 11% of he sample. Abou 54% are under age 45. Mos have a leas some hgher educaon, and 30% repor over $75,000 household ncome whle 25% ls as below $30,000; 14% refused o reveal her household ncome. 1 A wo-sage procedure was used o wegh hs dual-frame sample. The frs sage of weghng correced for dfferen probables of selecon assocaed wh he number of aduls n each household and each responden s elephone usage paerns. Ths weghng also adjuss for he overlappng landlne and cell sample frames and he relave szes of each frame and each sample. The second sage of weghng balanced oal sample demographcs o populaon parameers. The oal sample was balanced o mach naonal populaon parameers for sex, age, educaon, race, Hspanc orgn, regon (U.S. Census defnons), populaon densy, and elephone usage. The basc weghng parameers came from he U.S. Census Bureau s 2013 Amercan Communy Survey daa. The populaon densy parameer was derved from Census 2010 daa. The elephone usage parameer came from he analyss of he NHIS daa (Blumberg SJ, Luke JV. Wreless subsuon: Early release of esmaes from he Naonal Healh Inervew Survey, January-June, 2014. Naonal Cener for Healh Sascs. Dec 2014). Turow, Hennessy, Draper Page 10 of 24
Consumers Rejec Tradeoffs as Unfar To assess markeers clam ha Amercans are generally posve oward allowng he collecon of daa abou n reurn for some advanage, we asked our respondens wheher hey agree or dsagree hree ypes of radeoffs are far or OK. We presened he radeoffs ogeher wh fve oher aude saemens n random order so as no o call aenon o our parcular neres n he cos-benef Table 1: Characerscs of U.S. Aduls n Sample (N=1,506)* % Sex Male 48 Female 52 Age 18-24 14 25-34 21 35-49 30 50-64 26 65-89 9 Race Whe 76 Black or Afrcan Amercan 11 Asan or Pacfc Islander 4 Amercan Indan or Alaskan Nave 2 Mxed Race 3 Oher/Don Know/Refused 3 Hspanc or Lano Background? Yes 11 No 88 Don Know/Refused 1 Household Income Las Year Under $30,000 25 $30,000 o under $50,000 19 $50,000 o under $75,000 12 $75,000 and Over 30 Don Know/Refused 14 Hghes Educaon Level Less han hgh school graduae 7 Hgh school/ech school graduae Some college College graduae or more No answer 1 * In hs and all oher ables, when he numbers don add o 100% s because of a roundng error. 25 33 34 Turow, Hennessy, Draper Page 11 of 24
calculaons. Lsed n Table 2, each of he radeoff proposons depcs an everyday daa-collecon approach as well as a common prvlege (dscoun, mproved servce or use of a sore s wreless nerne) markeers clam o presen n reurn. Noe ha despe hese acves happenng all around hem, large majores of Amercans beleve each form of exchange s no accepable. Moreover, as Table 2 ndcaes, her rejecon s frm: by far he larges majory of naysayers say hey dsagree srongly. Table 2: Amercans Responses o Tradeoff-Aude Saemens (N=1,506) Srongly Agree Dsagree Agree (%) (%) (%) If companes gve me a dscoun, s a far exchange for hem o collec nformaon abou me whou my knowng. (91% dsagree) I s far for an onlne of physcal sore o monor wha I m dong onlne when I m here, n exchange for leng me use he sore s wreless nerne, or W-F, whou charge. (71% dsagree) I s OK f a sore where I shop uses nformaon has abou me o creae a pcure of me ha mproves he servces hey provde for me. (55% dsagree) *Neher was a voluneered answer. Srongly Dsagree (%) Neher* (%) 3 5 14 77 1 9 18 18 53 2 12 30 17 38 3 Furher analyss of hese responses ndcaes only a very small percenage of Amercans oally agree wh he overall concep of markeng radeoffs: abou 4% agree or agree srongly wh all hree proposons. To allow for a broader nerpreaon of a belef n radeoffs we creaed a radeoffaccepance ndex ha gave a value o each answer 5 o srongly agree, 4 o agree, 3 o neher agree nor dsagree, 2 o dsagree and 1 o srongly dsagree. We hen calculaed he average scores respondens had on he hree saemens o see how many averaged dsagreemen or agreemen wh he saemens. We found ha even usng from hs broader measure a small proporon 21% beleves common radeoffs wh markeers amoun o a far deal n conras o wha markeers pos as he major reason for people gvng up her daa. The Supermarke Scenaro and Cos-Benef Analyss We hen evaluaed a form of he prvacy paradox. Afer a block of aude quesons unrelaed o radeoffs, we presened our respondens wh a more concree daa-exchange suaon han he ones we used n he radeoff-aude saemens. For he nex few quesons, he nervewer sad, please hnk abou he supermarke you go o mos ofen. Le s say hs supermarke says wll gve you dscouns n exchange for s collecng nformaon abou all your grocery purchases. Would you accep he offer or no? Turow, Hennessy, Draper Page 12 of 24
A majory, 52%, say hey wll no, bu 43% reply hey wll accep supermarke scenaro, and 4% voluneer depends more han wce as many as n even he broad defnon of radeoff supporers. On he surface, he percenage accepng s que smlar o he 45% who agree wh he hrd saemen n Table 2. Furher analyss, hough, ndcaes he overlap of ndvduals s far from complee: 40% of hose who accep he supermarke dscoun do no concur ha I s OK f a sore where I shop uses nformaon has abou me o creae a pcure of me ha mproves he servces hey provde for me. Ths lack of correspondence even when he scenaros appear smlar underscores ha a small percenage conssenly accep he dea of radeoffs. We waned o know wheher people who say hey wll accep he supermarke dscoun would sll do when presened wh specfc assumpons a supermarke mgh make abou hem from analyzng her grocery-purchasng habs. Ths s knowledge Amercans almos never receve drecly bu may nu n real lfe as a resul of ads and coupons hey beleve are argeed o hem. The caegores he nervewers presened ran he gamu from generalzaons ha focus drecly on buyng habs ( wheher you end o buy low-fa foods ) o lfesyle nferences (for example, when you ake vacaons ) o demographc nferences ( your racal or ehnc background ) o medcal generalzaons ( he healh saus of you or someone n your famly ). For each one, he nervewers asked wheher respondens would accep he general dscoun offer under he new specfc condon. Table 3 presens her answers. I shows he lms of cos-benef analyss (and he prvacy paradox) as a raonale for markeers clam ha mos people wll provde personal daa n exchange for sore deals. There s a subsanal drop n he percenage of people who say hey wll accep he dscoun when confroned wh wha he supermarke wll nfer abou hem. Moreover, we fnd ha he caegores make up an ordered dffculy ndex ndcang sensvy: ndvduals who say hey won exchange daa for a ceran rade-off opc (for example, when you ake vacaons ) also won exchange daa for he opcs below n he ndex. 2 Clearly, he percenage of Amercans who seem even on he surface o accep a cos-benef analyss does no exceed 43%. And when we ask people drecly o hnk abou he mplcaons of wha hey provde o markeers, hey balk, and he accepance level goes down o around 20%. Ths declne s no conssen wh markeers asserons ha people are gvng up her personal nformaon because of cos-benef analyss. In he supermarke scenaros, hey are dong jus he oppose: ressng he dea of gvng daa for dscouns. Mos Amercans are resgned and hey far ounumber radeoff supporers Here s where we begn o explore an alernave vew, our resgnaon hypohess. The meanng of resgnaon we nend s, o quoe a Google dconary enry, he accepance of somehng undesrable bu nevable. 27 Askng our respondens wheher hey agree or dsagree wh wo saemens, we nvesgaed how much of he populaon can be descrbed as resgned when comes o markeers 2 Usng hese daa combned wh hose 52% who sad hey would no accep any knd of supermarke dscoun we creaed a supermarke-radeoff ndex measure whch goes from 0 f a person sad hey would no accep any supermarke dscoun o 9 f a responden sad hey would accep a supermarke dscoun and also acceped all specfc dscouns based on shoppng and demographc characerscs lsed n Table 3. Noe ha nne percen of he respondens had he hghes possble value of hs ndex. Turow, Hennessy, Draper Page 13 of 24
learnng abou her behavor onlne. One saemen was I wan o have conrol over wha markeers can learn abou me onlne. The oher was I ve come o accep ha I have lle conrol over wha markeers can learn abou me. We presened hese saemens n random order among a seres of sx agree/dsagree proposons so ha he respondens wouldn see he relaonshp beween he wo or suspec our nenon. To be denfed as resgned, a person had o agree wh boh of hem. Table 3: Percenage of people who accep he supermarke offer and know supermarke wll analyze grocery her purchases o make parcular assumpons abou hem N=1506 Yes (%) Accep he dscouns whou any specfc assumpons added 43 Accep he dscoun knowng supermarke wll make assumpons abou Wheher you end o buy low-fa foods 33 Wheher you have chldren and how old hey are 27 Wha acves you mgh do ousde of work 25 When you ake vacaons 22 The healh saus of you or someone n your famly 21 How much money you make 21 Wheher you are gong hrough a major lfe even 19 Your racal or ehnc background 19 Table 4 presens her answers o he saemens. The percenage of people agreeng wh he frs one (84%) s hgher han he second, where agreemen s neverheless que hgh (65%). When we nvesgaed he overlap ha desgnaes resgnaon, we found ha a majory of he populaon 58% s resgned. We also found wdespread suspcon based on anoher saemen we gave hem: 72% of Amercans rejec he dea ha wha companes know abou me from my behavor onlne canno hur me. If we combne he people who are resgned wh hose who beleve wha frms know can hur hem, we fnd ha 41% of Amercans are no only resgned, hey hold a dark concern ha he basc dynamcs of he emergng markeplace wll cause hem njury and ha hey canno conrol. Table 4: Amercans Responses o he Resgnaon Aude Saemens (N=1,506) Srongly Srongly Agree Dsagree Agree Dsagree Neher* DK (%) (%) (%) (%) (%) (%) I wan o have conrol over wha markeers can learn abou me 61 23 8 7 1 1 onlne. (84% agree) I ve come o accep ha I have lle conrol over wha markeers can 1 31 34 16 18 1 learn abou me onlne. (65% agree) * Neher was a voluneered answer. Turow, Hennessy, Draper Page 14 of 24
In he absrac, behavors drven by cos-benef analyses and hose drven by resgnaon may look he same. When we looked n our scenaro a he people who agreed o gve up her daa for supermarke dscouns, we found mporan dfferences beween radeoff supporers and people who are resgned. There s a srong posve sascal relaonshp (.56) beween he radeoff ndex (based on he hree saemens n Table 2) and accepng or rejecng varous knds of supermarke dscouns. By conras, here s no sascal relaonshp (-.01) beween beng resgned o markeers use of daa and accepng or rejecng varous knds of supermarke dscouns. Pu anoher way, people who beleve n radeoffs gve up her daa predcably, whle people who are resgned don do n a predcable manner. They do gve up her daa, hough. In fac, we found ha 57% of hose who ook he supermarke deal were resgned. By conras, a much smaller 32% were radeoff supporers even usng he broader measure of radeoff suppor. The larger percenages of people n he populaon who are resgned compared o people who beleve n radeoffs ndcae ha n he real world people who exchange her daa for benefs are more lkely o do whle resgned raher han as he resul of cos-benef analyss. Moreover, resgnaon s wdespread across he U.S. populaon. As Fgure 1 ndcaes, age and gender show no dfferences n resgnaon. When comes o educaon and race, sascally sgnfcan dfferences do show up: hgher resgnaon percenages for Whes compared o non-whes and for he more educaed people compared wh respondens who have a hgh school educaon or less. A he same me, hose comparson show ha 50% or more of he ndvduals n mos caegores of respondens are resgned. Fgure 1: Percenage of Resgned Respondens by Demographc Groups Resgnaon and Age Caegory Resgnaon and Race.6.6 % Resgned.55.5 % Resgned.55.5.45 18-25 26-40 41-55 56-66 67+ Age Caegory No Sgnfcan Dfference.45 Whe Non Hspanc Black Non Hspanc All Ohers Ehncy Sgnfcan Dfference.6 Resgnaon and Gender.6 Resgnaon and Educaon % Resgned.55.5 % Resgned.55.5.45.45 Female No Sgnfcan Dfference Gender Male HS or Less Some College College+ Educaon Group Sgnfcan Dfference Turow, Hennessy, Draper Page 15 of 24
Amercans ofen don have he basc knowledge o make nformed cos-benef choces To furher queson markeers emphass ha Amercans use cos-benef calculaons, we found ha large percenages of Amercans ofen don have he basc knowledge o make nformed cos-benef choces abou ways markeers use her nformaon. Renforcng hese naccuraces, large percenages of Amercans overesmae he exen o whch he governmen proecs hem from dscrmnaory prcng he ac of changng prces from person o person based on hose ndvduals consumer profles. Table 5 presens he answers our respondens gave o 13 saemens abou how markeers use her nformaon and he regulaons relang o personalzed prcng. For each saemen, we asked he person wheher you hnk s rue or false. If you are unsure, jus say so. The bolded number Table 5: True-False saemens abou how markeers use nformaon and regulaons relaed o personalzed prcng (N=1,506) True (%) False (%) DK (%) Wrong (%) I s possble for Facebook o lnk wha people do when hey access he nerne on a lapop compuer wh wha hey do on her cell phones or ables apps. 74 12 14 26 I s legal for an onlne sore o charge people dfferen prces dependng on where hey are locaed. 63 26 10 36 When I allow a moble app o know my locaon, ha app s legally allowed o share ha locaon wh anoher company. 62 26 12 38 A company can ell ha I have opened s emal even f I don respond. 61 27 13 40 By law, sores ha provde free wreless nerne, or W-F, have o secure he servce so nobody can rack wha cusomers who are usng are dong onlne. 29 59 11 40 By law, my supermarke mus oban my permsson frs before sellng nformaon abou my food purchases o oher companes. 38 51 11 49 Banks ofen send her cusomers emals ha ask hem o clck on a lnk wanng hem o verfy her accoun. 43 49 8 51 By law, a ravel se such as Expeda or Orbz ha compares prces on dfferen arlnes mus nclude he lowes arlne prces. 38 47 15 53 I s legal for an ONLINE sore o charge dfferen people dfferen prces a he same me of day. 44 42 13 55 I s legal for an OFFLINE or PHYSICAL sore o charge dfferen people dfferen prces a he same me of day. 38 50 12 62 When a webse has a prvacy polcy, means he se wll no share my nformaon wh oher webses or companes whou my permsson. 58 35 7 65 If I wan o be sure no o be racked on my cell phone, I should clear he phone s cookes. 46 35 18 64 By law, my pharmacy mus oban my permsson frs before sellng nformaon abou he over-he-couner drugs I buy o oher companes. 56 31 13 69 Turow, Hennessy, Draper Page 16 of 24
ndcaes he percenage ha go he correc answer; he las column ndcaes he percenage who ddn ge he answer correc (ha s, he percen wrong plus he percenage who don know or are unsure). The able reveals a range of dffculy, wh 5 saemens where 40% or less ddn know he correc answer and 8 answers where from 49% o 69% ddn know he correc answer. A separae analyss revealed ha half of he populaon go only sx or fewer answers correc, whle 32% go seven o nne quesons rgh, and 17% were correc on en o hreen of he saemens. Only 3% excelled a knowng all bu wo of he saemens. Overall, he able ndcaes large gaps n knowledge abou basc daa-markeng rules among large percenages of Amercans. In addon o legal acves Amercans canno evaluae correcly, 51% of Amercans canno recognze he possbly of phshng, an llegal acvy where heves posng as banks or oher organzaons wh mporan daa ask cusomers o clck on a lnk wanng hem o verfy her accoun. Moreover, large percenages of he populaon beleve ncorrecly ha governmen laws proec hem from prce dscrmnaon and ceran forms of daa collecon. These wdespread msconcepons sugges ha even when Amercans do wegh he coss and benefs of gvng up her daa, hey frequenly base hose choces on ncorrec nformaon. More knowledge predcs beng resgned and gvng up daa n he supermarke scenaro Whle he low knowledge level of Amercans abou how markeers use her nformaon makes unlkely hey could carry ou correc coss-benef calculaons n he dgal-markeng envronmen, havng more knowledge s no he proecve feaure ha academcs have suggesed. We found ha hose who know more abou he markeng laws and pracces are more lkely o be resgned. Moreover, resgned people s decsons o accep supermarke dscouns even when he supermarke collecs ncreasngly personal nformaon s also posvely relaed o knowledge. To demonsrae, we use our knowledge ndex based on he 13 ems presened n Table 5. 3 When we explore how beng resgned s assocaed wh he knowledge ndex, we ge a plo wh a sgnfcan posve slope, shown n Fgure 2. The plo shows ha he more a person knows abou nformaon collecon n he markeng world, he more lkely he resgnaon. In hs conex, knowledge defnely does no go along wh a sense of power. To he conrary, our survey fnds ha when comes o proecng her personal daa, hose wh he wherewhal o accuraely calculae he coss and benefs of prvacy are lkely o consder her effors fule. Resgnaon and ncreasng knowledge also work ogeher when comes o consumers exchangng her grocery-buyng daa for dscouns n he supermarke scenaro. Fgure 3 plos ha neracon based on mulple regresson analyss predcng he supermarke rade-off ndex (he number of daa opcs a person s wllng o exchange) conrollng for age, race, and gender. I shows ha he lkelhood of a person gvng more daa, and more sensve daa (because s an ordered-dffculy ndex), goes up more sharply wh ncreasng knowledge when he person s no resgned. The fgure does ndcae, 3 For hs analyss, we combned he respondens who go all he quesons correc (a score of 13) wh hose who go 12 correc because he number of respondens wh all correc was que low (N = 9). Turow, Hennessy, Draper Page 17 of 24
hough, ha resgnaon assocaes wh more raher han less knowledge when comes o predcng ha a person wll choose o exchange daa for supermarke dscouns. Fgure 2: Resgnaon and Knowledge (N = 1,506; brackes defne 95% confdence nervals).7 Percen Resgned.6.5.4 0 2 4 6 8 10 12 Recoded Knowledge Sgnfcan Posve Slope Fgure 3: Ineracon Plo of Knowledge and Resgnaon Predcng Supermarke Tradeoff Index (N = 1,416) 4 Predced Value of Trade Off Index 3 2 1 0 0 1 2 3 4 5 6 7 8 9 10 11 12 Recoded Knowledge No Resgned Adjused for age, race, and gender Sgnfcan Ineracon Resgned Turow, Hennessy, Draper Page 18 of 24
Concludng Remarks Our hope s ha he resuls from our naonally represenave survey of Amercans wll encourage new perspecves on he problems swrlng around consumer markeers relaonshps wh he publc. One ssue rased by he fndngs s ha Amercans are resgned regardng her ably o conrol he nformaon markeers can collec and use abou hem. Resgned ndvduals may behave n ways ha allow markeers o clam hey are unconcerned or accep he economc logc ha nsss people rade her daa for benefs. We found nsead ha mos Amercans rejec hs logc. They also express concerns abou how her daa are beng used and don beleve markeers gve hem npu no. A consequence of hs broad sense of fuly, perhaps even dsrespec, s ha over he nex decades Amercans mgh rejec he legmacy of a cenral nsuon: markeng and consumer commerce. Ths wll have a desablzng rpple effec on many aspecs of Amercan lfe. The fuly over nformaon conrol we are seeng n he publc sphere s dsrupng a compac ha commercal markeers made wh Amercans hrough he pas cenury. As hsoran Wllam Leach noes, a key developmen of he pas one hundred and ffy years was he democrazaon of desre. I nvolved he remagnng of democracy beyond polcs o Amercans equal rghs o desre he same goods and have access o hem n a shared publc markeplace. 28 Leach and ohers noe ha he democrazaon of desre had some no-so-aracve aspecs; moreover, he deal never fully ook hold. Neverheless, he growh of posed prces (as opposed o hagglng n response o prces ha vary by ndvduals or groups) and he open presenaon of goods (as opposed o dfferen goods offered o dfferen people based on merchans preconceved concepons of hem) are wo feaures of he sandardzed markeplace ha have leveled opporunes for shoppers. One underapprecaed consequence of markeers 21s cenury daa acves s ha hese and relaed feaures of he democrazed markeplace are begnnng o dsappear. Wh he rse of personalzaon as a sraegy, adversers and realers are changng prces for ndvduals (ofen hrough argeed dscouns) when hey shop a home, a work, and even n physcal asles, based on wha he sellers know abou hem. Adversers and realers are also changng he goods hey hghlgh for people based on he profles hey have creaed abou hem from her own and oher onlne and offlne nformaon sources. Alhough some onlne frms have been allowng Amercans o see some of he daa pons hey hold abou hem, he parculars of how and when hey use hese profles and wh wha oher sources of nformaon s a mysery n almos all cases. I s evden ha he amoun of nformaon markeers have already colleced s enormous. For example, he Forreser Research consulancy esmaed n 2014 ha daabase markeng frm Acxom has abou 1,500 daa pons for each of over 500 mllon acve nerne users, mos of hem n he Uned Saes. 29 Anoher 2014 Forreser repor looked o no-oodsan crcumsances where markeers would rounely make decsons based on a cusomer s crcle of socal relaonshps and nfluencers...sensor daa [from n-sore echnologes], sreamng real-me daa, acqured daa [from frms such as Acxom]...anyhng. 30 Amercans surely sense he changes ha are happenng and feel powerless o do anyhng abou hem. Our fndng ha 58% of Amercans are resgned o no havng a say n hese acves ha hey do no wan o lose conrol over her nformaon bu also beleve her loss of conrol has already happened suggess he populaon feels a lack of auonomy even a hs early dae n he shf away from a democrazed markeplace. We also found wdespread suspcon: 72% of Amercans rejec he dea ha Turow, Hennessy, Draper Page 19 of 24
wha companes know abou me from my behavor onlne canno hur me. When we combned he people who are resgned wh hose who beleve wha frms know can hur hem, we found ha 41% of Amercans are no only resgned, hey hold a dark concern ha he basc dynamcs of he emergng markeplace wll cause hem njury and ha hey canno conrol. Markeng and realng execuves have ypcally played down any concerns abou her use of shopper daa. One way has been o depc an empowered publc ha acceps he noon ha s releasng daa wllngly as a radeoff for benefs receves. Our survey challenges markeers ypcal cos-benefanalyss defense by showng que clearly ha mos Amercans do no accep he farness of geng dscouns n rade for her daa. Bu as conroverses around daa mnng and predcve analycs have swrled around he Naonal Secury Agency as well as around rackng by busnesses o ferre ou people wh parcular dseases, fnancal ssues, or sexual orenaons, markeers have aken pans o separae ou rackng people for he purposes of ordnary adversng and shoppng. Consder he commens of Sco Howe, Presden and CEO of Acxom: Some beleve ha all daa s of equal mporance and herefore mus be conrolled n exacly he same way. Ths s jus no he case. Daa regardng personal nformaon ha perans o employmen or nsurably decsons, or ha relaes o sensve healh-relaed ssues or confdenal maers, deserves much dfferen reamen han daa ha would ndcae ha I am a spors fan. Too ofen, legslaon and regulaon seek o pan ssues n broad srokes, ofen wh unforeseen consequences o lberes and nnovaon, despe beng born of he bes nenons. 31 Apar from gnorng he ssue of nformaon respec ha Amercans wan o be able o manage her daa n commerce hs vew pushes asde he dscrmnaory poenal ha seemngly bengn peces of daa can have on he opporunes people can have n he publc sphere. In he no-oo-dsan fuure, companes may well merge he caegory spors fan wh dozens of oher characerscs abou an ndvduals eang habs (a he ballpark, for example), ncome, number and age of chldren, house value, vacaon habs, moble-racked locaons, clohes-shoppng habs, and meda-use paerns o creae profles ha dub hem wnners or losers regardng ceran areas of shoppng and he adversemens relaed o hem. For reasons hey don undersand, people may see paerns of dscouns ha sugges hey are beng sloed no ceran lfesyle segmens. Ceran adversers may suppor her meda habs bu no, perhaps, o he exen ha adversers suppor neghbors or co-workers. The ndvduals wll vaguely undersand ha her profles are he cause, and hey may ry o change her behavor o ge beer deals, ofen whou success, all he whle wonderng why he sysem he opaque under-he-hood predcve analycs regmes ha hey know are rackng her lves bu o whch hey have no access s reang hem ha way. Argumens ha play down he profound dscrmnaory poenal of hs sor of daa-drven consumer commerce have he same rheorcal ams as argumens ha nss Amercans wan o exchange daa for benefs: They gve polcymakers false jusfcaons for enablng he collecon and use of all knds of consumer daa ofen n ways ha he publc fnd objeconable. Ye when 3 of every 5 Amercans are resgned o lack of conrol over her daa relaonshps wh markeers, when 2 of every 5 are boh resgned o markeer conrol over her daa and worred ha he conrol can hur hem, and when people wh knowledge are acually more lkely raher han less lkely o be resgned, we have a problem. The fndngs ndcae subsanal ensons n a cenral area of socey s publc sphere ha canno be swep away by execuves asserons of consumer auonomy and raonal choce. Turow, Hennessy, Draper Page 20 of 24
Unforunaely, he socal and nsuonal harms we have denfed here don map easly ono he concep of harms used by he Federal Trade Commsson and oher governmen organzaons. As he FTC self noed n a 2010 repor, approaches o proecng consumer prvacy ha employ a harm-based model end o focus on enhancng physcal secury, lmng economc njury, and mnmzng unwaned nrusons no he daly lves of consumers. 32 The repor goes on o noe ha such an approach ends o leave ou repuaonal harm and he pervasve fear many Amercans face of beng wached ssues ha ceranly nersec wh concerns we have found n hs survey. Ye he concerns we have denfed are boh more specfc and broader: Amercans wan o have conrol over he daa markeers use abou hem. They beleve ha markeers keep ha daa conrol for hemselves. They worry (perhaps because of ha lack of auonomy) ha he commercal relaonshps hey have are dangerous. And hey feel resgned o no beng able o do anyhng abou. A hear here s an asymmercal relaonshp beween wha markeers know abou Amercans and wheher Amercans can learn wha ha nformaon s and how s beng used. The erm ransparency s used a lo n he markeng rade press o sgnfy adversers nssence ha hey mus be able o look no he specfcs of her programmac-adversng acves so ha hey can calculae her reurn on nvesmen and no ge harmed monearly. Ye when comes o ransparency n adversers relaonshps wh he publc, he erm s far less popular and he acves relaed o mplemenng openness far less dlgen. We need naves ha wll gve members of he publc he rgh and ably o learn wha companes know abou hem, how hey profle hem, and wha daa lead o wha personalzed offers. We also need o ge people exced abou usng ha rgh and ably. Here are a few suggesons: One way o encourage more corporae openness abou he commercal use of people s daa may be by namng, prasng, and shamng. Publc neres organzaons as well as governmen agences should develop clear defnons of ransparency ha reflec concerns denfed n hs survey and elsewhere. They should hen sysemacally call ou companes regardng how well or badly hey are dong based on hese values. When acvss, journalss, and governmen offcals name, prase, and shame frms ha don abde by he ransparency norms, hey can aler he publc o say away from bad acors and possbly force hose acors o change her behavors. We also sugges an nave o dssec and repor on he mplcaons of prvacy polces. Acvss, journalss and governmen offcals perhaps aded by crowdsourcng naves should ake on he role of nerpreng hese legally-bndng documens for he publc. The focus should no prmarly be on wheher or no webses abde by her prvacy polces. Insead, prvacy-polcy nerpreers can be mos helpful by uncoverng for her publcs how companes say hey collec and use her daa, and wha he mplcaons mgh be for he ndvdual and socey. When hs nformaon s avalable n a dgesble form, may spur nformed namng, prasng, and shamng. I may well also lead frms o change behavors czens fnd dsagreeable. These consderaons along wh our overall survey fndngs sugges a need o advocae for he rgh o know one s profle and how s used. As long as he algorhms companes mplemen o analyze and predc he fuure behavors of ndvduals are hdden from publc vew, he poenal for unwaned markeer exploaon of ndvduals daa remans hgh. We herefore ough o consder an ndvdual s rgh o access he profles and scores companes use o creae every Turow, Hennessy, Draper Page 21 of 24
personalzed message and dscoun he ndvdual receves. Companes wll push back ha gvng ou hs nformaon wll expose rade secres. We argue here are ways o carry hs ou whle keepng her rade secres nac. Whou access o hs nformaon, ndvduals wll connue o feel dsempowered and dsrespeced by companes. Moreover, he very nsuon of consumer commerce may be hreaened n ways ha damage he broad canvas of Amercan lfe. References 1 Yahoo! Adversng, The Balancng Ac: Geng Personalzaon Rgh (Yahoo!, May 2014), p. 11, hps://adversng.yahoo.com/insghs/balancing-act.hml, accessed May 8, 2015. 2 Joseph Turow, Lauren Feldman, and Kmberly Melzer, Open o Exploaon: Amercan Shoppers Onlne and Offlne (Unversy of Pennsylvana: Annenberg Publc Polcy Cener, 2005). 3 Mary Madden, Few feel ha he governmen or adversers can be rused, Pew Inerne Projec, November 12, 2014, oplne quesonnare: hp://www.pewnerne.org/fles/2014/11/prvacypaneltoplne.pdf 4 How Can Companes Acqure Cusomer Daa Whle Buldng Cusomer Loyaly A The Same Tme? Ask Permsson, Ban & Company, Press Release, May 11, 2015, hp://www.ban.com/abou/press/press-releases/dgal-prvacy-survey- 2015-press-release.aspx 5 See, for example, Sonja Uz and Ncole Kramer, The Prvacy Paradox On Socal Nework Ses Revsed, Cyberpsychology, 3(2) arcle 1, hp://cyberpsychology.eu/vew.php?csloclanku=2009111001%26arcle=1, accessed Aprl 26, 2015; and Susan B. Barnes, A prvacy paradox: Socal neworkng n he Uned Saes. Frs Monday, 11, hp://frsmonday.org/hbn/cgwrap/bn/ojs, accessed May 8, 2015. 6 Brad Sone, Our Paradoxcal Audes oward Prvacy, New York Tmes, July 2, 2008, sec. Bs Blog, hp://bs.blogs.nymes.com/2008/07/02/our-paradoxcal-audes-owards-prvacy/. 7 Jacquelne Renfrow, Increased Spendng on Loyaly Programs No Increasng Engagemen, FerceReal, March 27, 2915, hp://www.fercereal.com/sory/ncreased-spendng-loyalyprograms-no-ncreasng-engagemen/2015-03-27. 8 Gordon Hull, Successful Falure: Wha Foucaul Can Teach Us abou Prvacy Self-Managemen n a World of Facebook and Bg Daa, December 2, 2014, hp://ssrn.com/absrac=2533057. 9 The Truh Abou Shoppng, McCann Truh Cenral, Augus 20, 2014, hp://ruhcenral.mccann.com/wpconen/uploads/2014/09/mccann_truh_abou_shoppng_gude.pdf, accessed May 8, 2015. 10 Chuck Marn, Wha The Shopper Ges Ou Of Beng Tracked, mcommercedaly, May 28, 2014, hp://www.medapos.com/publcaons/arcle/226734/wha-he-shopper-ges-ou-of-beng-raked.hml, accessed May 8, 2015 11 Sco Snyder, Moble Devces: Facng he Prvacy Vs. Benef Trade-Off, Forbes.com, Augus 3, 2012, hp://www.forbes.com/ses/cocenral/2012/08/03/moble-devces-facng-he-prvacy-vs-benef-rade-off/, accessed May 8, 2015. 12 Yahoo! Adversng, The Balancng Ac: Geng Personalzaon Rgh (Yahoo!, May 2014), p. 11, hps://adversng.yahoo.com/insghs/balancing-act.hml, accessed May 8, 2015. Turow, Hennessy, Draper Page 22 of 24
13 US consumers wan more personalzed real experence and conrol over personal nformaon, Accenure survey shows, Accenure, March 9, 2015, hp://newsroom.accenure.com/news/us-consumers-wan-more-personalzed-realexperence-and-conol-over-personal-nformaon-accenure-survey-shows.hm, accessed Aprl 26, 2015 14 The Loyaly Repor 2014 From Bond Brand Loyaly, Bond Brand Loyaly, 2014, hp://cdn2.hubspo.ne/hub/352767/fle-942578460- pdf/whepapers/bond_brand_loyaly_2014_loyaly_repor_us.pdf?=1430494118043, p. 21, accessed May 1, 2015. 15 The Loyaly Repor 2014 From Bond Brand Loyaly, Bond Brand Loyaly, 2014, hp://cdn2.hubspo.ne/hub/352767/fle-942578460- pdf/whepapers/bond_brand_loyaly_2014_loyaly_repor_us.pdf?=1430494118043, p. 19, accessed May 1, 2015. 16 Yahoo! Adversng, The Balancng Ac: Geng Personalzaon Rgh (Yahoo!, May 2014), 7, hps://adversng.yahoo.com/insghs/balancing-act.hml, accessed May 8, 2015. 17 Yahoo! Adversng, The Balancng Ac: Geng Personalzaon Rgh (Yahoo!, May 2014), p. 2, hps://adversng.yahoo.com/insghs/balancing-act.hml, accessed May 8, 2015. 18 Yahoo! Adversng, The Balancng Ac: Geng Personalzaon Rgh (Yahoo!, May 2014), p. 2, hps://adversng.yahoo.com/insghs/balancing-act.hml, accessed May 8, 2015. 19 See, for example, Chrs Hoofnagle s argumen agans he raonal choce heory of he lae prvacy law exper Alan Wesn: Chrs Jay Hoofnagle and Jennfer M. Urban, Alan Wesn's Prvacy Homo Economcus, 49 Wake Fores L. Rev. 261 (2014), avalable a: hp://scholarshp.law.berkeley.edu/facpubs/2395. 20 Joseph Turow, e. al., Amercans, Markeers, And The Inerne: 1999-2012, Repor from he Annenberg School For Communcaon, Unversy of Pennsylvana, March 2014, SSRN, 21 Alca M. McDonald and Lor F. Cranor, The Cos of Readng Prvacy Polces, I/S: A Journal of Law and Polcy for he Informaon Socey 4, no. 3 (2008). The Federal Trade Commsson has suggesed recognzes he dffculy of readng prvacy polces. For example, n clarfyng he collecon lmaon prncple of s prvacy framework, he Commsson saff wroe any daa collecon ha s nconssen wh hese conexs, companes should make approprae dsclosures o consumers a a relevan me and n a promnen manner ousde of a prvacy polcy or oher legal documen. See Federal Trade Commsson, Proecng Consumer Prvacy n an Era of Rapd Change: Recommendaons for Busnesses and Polcymakers, Washngon DC, Federal Trade Commsson, March 2012, p.27. 22 Joseph Turow Jennfer Kng, Chrs Jay Hoofnagle, Amy Bleakley, and Mchael Hennessy, Conrary o Wha Markeers Say Amercans Rejec Talored Adversng and Three Acves Tha Enable I, Onlne, (Sepember 29, 2009), 21, SSRN. 23 Alessandro Acqus, Curs Taylor, and Lad Wagman, The Economcs of Prvacy, Manuscrp, May 2015, hp://lwagman.org/prvacy_survey.pdf, accessed May 10, 2015 24 Hull, Successful Falure: Wha Foucaul Can Teach Us abou Prvacy Self-Managemen n a World of Facebook and Bg Daa, 9. 25 Yahoo! Adversng, The Balancng Ac: Geng Personalzaon Rgh (Yahoo!, May 2014), p. 2, hps://adversng.yahoo.com/insghs/balancing-act.hml, accessed May 8, 2015. 26 Mark Andrejavc, I Spy: Survellance and Power n he Ineracve Era (Lawrence, Kansas: Unversy Press of Kansas, 2007). A relaed concep s wha Shoshana Zuboff erms he Bg Oher : corporae mechansms of nformaon conrol ha effecvely exle persons from her own behavor whle producng new markes of behavoral predcon and modfcaon. See Shoshana Zuboff, Bg Oher: Survellance Capalsm And The Prospec Of An Informaon Cvlzaon,: Journal Of Informaon Technology 2015 (30), pp. 75-89. 27 Google defnon of resgnaon, hps://www.google.com/?gws_rd=ssl#q=resgnaon, accessed May 18, 2015. 28 Wllam Leach, Land of Desre: Merchans, Power, and he Rse of a New Amercan Culure (New York: Vnage books, 1993), p. 6. 29 Rchard Joyce, The Fuure of Dgal Meda Buyng, Forreser Research, December 2, 2014, p. 10. Turow, Hennessy, Draper Page 23 of 24
30 Randy Heffner, A Radcal Rehnk of Daa Archecure for Cusomer Engagemen, Forreser Research, July 17, 2014, p. 3. 31 Sco Howe, A Prce Call-o-Acon for he Daa Indusry, Adversng Age, Aprl 8, 2014, hp://adage.com/arcle/prvacy-and-regulaon/a-prvacy-call-acon-daa-ndusry/292464/, accessed May 18, 2015. 32 Federal Trade Commsson. Proecng Consumer Prvacy n an Era of Rapd Change. (FTC, December 2010). hps://www.fc.gov/ses/defaul/fles/documens/repors/federal-rade-commsson-bureau-consumer-proeconprelmnary-fc-saff-repor-proecng-consumer/101201prvacyrepor.pdf, accessed, May 16, 2015. Turow, Hennessy, Draper Page 24 of 24