Al­go­rithms, pri­vacy and the fu­ture of shop­ping.

Harper’s Bazaar (Australia) - - Contents -

“When an e-com­merce brand sends you an email that says ‘What’s new for you’,

the prod­ucts re­ally were se­lected for you.”

THE DAY Mark Zucker­berg an­swered ques­tions be­fore the United States congress, I de­cided to see what he had on me. Like mil­lions of oth­ers, I won­dered if my data had been har­vested for some ne­far­i­ous pur­pose. So I down­loaded it all. what I got back — be­yond party-plan­ning group chats and friends of friends I’d qui­etly un­fol­lowed — looked less like my psy­che laid bare and more like the deep­est, dark­est re­gions of my credit card state­ments.

It was all there. Ev­ery on­line store I’d ever shopped from — and plenty I’d never heard of. Net-a-porter, Match­es­fash­, Moda Operandi and Far­fetch all had my con­tact in­for­ma­tion. They’d all reached out with ad­ver­tis­ing, and it was work­ing — I was click­ing on them. Psy­cho­graphic tar­get­ing, re­mar­ket­ing (aka re­tar­get­ing) and ‘looka­like au­di­ences’ aren’t just the do­main of shad­owy po­lit­i­cal op­er­a­tives. your friendly neigh­bour­hood lux­ury brands are do­ing it too.

“I think the av­er­age con­sumer is highly aware of re­tar­get­ing ... they no­tice when they visit a web­site and then see ads for that site over and over ...” says Re­becca Walker Reczek, a pro­fes­sor of mar­ket­ing at Ohio State Uni­ver­sity. “what they may not be as aware of is when they get an ad for a web­site they haven’t nec­es­sar­ily vis­ited in the past but that is be­ing shown to them based on their pat­tern of be­hav­iour on­line. this is the type of be­havioural tar­get­ing that might slip un­der the con­sumer’s radar; they are be­ing shown an ad based on an in­fer­ence of who an al­go­rithm thinks they are.”

One brand I’d been see­ing a lot of was Tome, a New York la­bel by Aus­tralians Ryan Lobo and Ra­mon Mar­tin that makes vi­brant work­wear and dresses for the kind of so­phis­ti­cated sum­mer party you wish you’d get in­vited to more of­ten. So I asked Nicki Wil­son of CVR Mar­ket­ing, who han­dles their dig­i­tal ad buy­ing strat­egy, why.the an­swer was sim­ple. “if you fol­low Tome,” — which I do — “you’re go­ing to get in­for­ma­tion about Tome in your feed. ”typ­i­cally, I was see­ing ads from Tome, not news­feed posts. Tar­get­ing cus­tomers who al­ready love your prod­uct, whether they’re ex­ist­ing shop­pers, so­cial me­dia fol­low­ers or those who’ve signed up for a news­let­ter is “re­ally your low-hang­ing fruit”, wil­son ex­plained.

If you read the pri­vacy pol­icy of pretty much any web­site you shop on, you’ll find they store data about what you buy and how you browse. Of­ten, this will be used to send you more per­son­alised mar­ket­ing ma­te­rial. web­sites use snip­pets of code called cook­ies to track your move­ments on­line. this code feeds into other web­sites — such as Face­book and Google — where mar­keters might try to reach you. So us­ing Face­book’s pixel, an e-tailer could, for ex­am­ple, cre­ate a ‘Cus­tom Au­di­ence’ of peo­ple who’ve browsed Stella Mccart­ney dresses on its site, to re­tar­get later. (Google has an equiv­a­lent fea­ture called Cus­tomer Match; Linkedin has Matched Au­di­ences; Twit­ter has Tai­lored Au­di­ences; Pin­ter­est has its own tar­get­ing tools). the e-tailer won’t be able to iden­tify in­di­vid­u­als within this Cus­tom Au­di­ence — but Face­book matches them to

its users. that’s why my data showed I was con­nected to brands I hadn’t bought from but which I had browsed. Face­book also al­lows mar­keters to cre­ate ‘Looka­like Au­di­ences’ of peo­ple who ex­hibit sim­i­lar pat­terns of be­hav­iour to their cus­tomers (Google has a fea­ture called Sim­i­lar Au­di­ences; Pin­ter­est has Ac­ta­like; Twit­ter has an al­go­rithm with a sim­i­lar func­tion). So if you’re be­ing served an ad for a la­bel you’ve never heard of, it might be be­cause their cus­tomers also tend to, say,‘ like’ Harper’s BAZAAR on Face­book, live in Aus­tralia and have an In­sta­gram ac­count.

Th­ese pat­terns aren’t spot­ted by hu­mans; it’s all pow­ered by ma­chine learn­ing al­go­rithms, a form of ar­ti­fi­cial in­tel­li­gence. For Amie Stepanovich, US pol­icy man­ager at dig­i­tal rights or­gan­i­sa­tion Ac­cess Now, it’s what com­put­ers, not peo­ple, know about you that can be par­tic­u­larly un­com­fort­able. “We’re see­ing com­pa­nies know more about in­di­vid­u­als than [the in­di­vid­u­als] do about them­selves, be­cause they’re tak­ing in­for­ma­tion that’s given to them, or that’s avail­able in some other data­base, and us­ing al­go­rithms and other tools to make de­ci­sions about peo­ple based on it.

“I think that’s the next piece that we’re fi­nally get­ting a look at ... [is] the ma­nip­u­la­tion,” Stepanovich con­tin­ues. “Peo­ple can be mi­cro­tar­geted through their pref­er­ences and ma­nip­u­lated … Know­ing when some­body might be more likely to pur­chase some­thing. In­fer­ring how much money they might have to spend at a par­tic­u­lar time ... It takes the process of pur­chas­ing and puts a lot of power in the hands of the re­tail mar­ket.”

While it might feel a lit­tle creepy to have a pha­lanx of for­mu­las fol­low­ing your ev­ery move, I have no­ticed lately my dig­i­tal shop­ping ex­pe­ri­ences are im­prov­ing. wil­son says her clients all use pro­gram­matic logic (au­to­mated de­ci­sion-mak­ing based on al­go­rithms) to build bet­ter, more rel­e­vant email mar­ket­ing. when an e-com­merce brand sends you an email that says “What’s new for you”, the prod­ucts re­ally were se­lected for you.

“The more data you have [as a mar­keter], the more in touch you are with your au­di­ence [and] the more so­phis­ti­cated you can be in per­son­al­is­ing that ex­pe­ri­ence,” says Hana Abaza, head of mar­ket­ing at Shopify Plus, which pro­vides the e-com­merce plat­form for thou­sands of startup and es­tab­lished brands. “i think more and more we’re mov­ing in that di­rec­tion.”

I ac­tu­ally en­joy be­ing served ads for K.j ac­ques san­dals when I’m shop­ping for my new sum­mer wardrobe. And I’m not alone. In a re­cent study, Walker Reczek found that “once con­sumers know an ad was tar­geted at them as a re­sult of their click be­hav­iour, they were more in­flu­enced by it … So that means con­sumers might feel more so­phis­ti­cated af­ter re­ceiv­ing an ad for a lux­ury brand ...A lux­ury brand could use be­havioural tar­get­ing to make con­sumers feel more like the kind of per­son who buys lux­ury brands, there­fore in­creas­ing pur­chase in­ten­tions.”

Yet the study found that only works if you know why you’re re­ceiv­ing an ad. If the ad doesn’t align with your in­ter­ests, it won’t change your self-im­age. That’s why trans­parency is key, not just for con­sumers, but also for mar­keters. All this data is be­ing used for the same thing any good sales as­sis­tant does the sec­ond you walk into a shop — pro­fil­ing you based on your tastes. there’s one key dif­fer­ence, how­ever: when you shop in store and a sales as­sis­tant is be­ing too at­ten­tive or overly fa­mil­iar, you can al­ways walk out. On­line, opt­ing out is more com­pli­cated.

There’s a good chance you’ve been asked for your con­sent to be tracked by cook­ies quite a bit lately. we’re see­ing pri­vacy leg­is­la­tion such as Europe’s Gen­eral Data Pro­tec­tion Reg­u­la­tion put power over how data is used back into con­sumers’ hands. Stepanovich thinks this is a great first step, but is wary that there are still op­por­tu­ni­ties for com­pa­nies to mis­use the mas­sive troves of data col­lected.

Know­ing big data can be used to ma­nip­u­late me, I de­cided to do what ev­ery­one does on so­cial me­dia. I cu­rated the com­pa­nies that could speak to me, to cre­ate a more pol­ished pic­ture of my­self. I combed through my Face­book pref­er­ences and deleted brands that re­in­forced habits I’d rather break. Candy Crush and De­liv­eroo went into the bin. But Tome, which serves me my dream sum­mer party dresses? I let them stay con­nected. we’re all be­ing profiled based on our pro­files, so we may as well pick out a flat­ter­ing an­gle.

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.