Do de­mo­graph­ics mat­ter in me­dia?

Con­tent pref­er­ences and con­sump­tion are no longer de­fined by age, race, gen­der, in­come, lo­ca­tion, etc.

The Insider - - CONTENTS -

De­mo­graph­ics have been used to tar­get the “ideal” con­sumer for over 50 years in re­search stud­ies, ad­ver­tis­ing, politics, and me­dia. Prior to the in­ter­net, it was the pri­mary re­search data that drove many an ad­ver­tis­ing or elec­toral cam­paign.

Us­ing fac­tors such as age, race, sex, in­come, ed­u­ca­tion level, oc­cu­pa­tion, mar­i­tal sta­tus, place of res­i­dence, and na­tion­al­ity, or­ga­ni­za­tions honed their de­mo­graphic-centered mod­els to try and cap­ture the lion’s share of a tar­get mar­ket; and for years, it was enough.

But not any­more.

We are living in a shrink­ing world where the six de­grees of sep­a­ra­tion be­tween peo­ple has been re­duced to a num­ber closer to three. One look at Face­book and its grow­ing pop­u­la­tion of over two bil­lion mem­bers, and it’s not hard to imag­ine three de­grees mov­ing to two in the not-too-dis­tant fu­ture.

That, in and of it­self, may not seem like a big deal, but when the in­ter­net en­abled a many-to-many con­nec­tion be­tween global cit­i­zens, it in­ad­ver­tently also in­flu­enced our at­ti­tudes, val­ues, per­son­al­i­ties, opin­ions, in­ter­ests, and life­styles.

The web has dis­torted the char­ac­ter­is­tics within tra­di­tional de­mo­graphic maps. In short, peo­ple today don’t be­have as their de­mog­ra­phy would dic­tate.

I can still re­mem­ber when the only news I got grow­ing up was in the daily pa­per dropped in the mail­box or on the ra­dio or TV. The dig­i­tal rev­o­lu­tion changed all that. Today we have al­most un­lim­ited choices in terms of con­tent, dis­tri­bu­tion chan­nels, and costs — choices we’re tak­ing full ad­van­tage of. As a re­sult, pub­lish­ers are fac­ing a brand cri­sis as read­ers, who ac­cess mul­ti­ple sources of con­tent ev­ery day on so­cial me­dia and other con­tent ag­gre­ga­tors, of­ten can’t re­mem­ber the source.

To thrive in this world of pro­mis­cu­ous au­di­ences, pub­lish­ers can’t con­tinue to treat read­ers like si­los of de­mo­graphic data be­cause con­tent pref­er­ences and con­sump­tion are no longer de­fined by age, race, gen­der, in­come, lo­ca­tion, etc.

In­stead, they need to put peo­ple first and work at cre­at­ing a re­la­tion­ship with each and ev­ery one of them by:

• Rec­og­niz­ing that de­mo­graph­ics and seg­men­ta­tion can­not help them ad­e­quately cu­rate con­tent for their read­ers

• See­ing con­sumers as in­di­vid­u­als that are con­stantly chang­ing as the world evolves around them

• Dig­ging deeper into the be­hav­iors of each per­son — be­hav­iors that can’t be mea­sured just by what they share, but rather what they ac­tu­ally con­sume

• Adopt­ing be­hav­ioral an­a­lyt­ics that con­tin­u­ally mon­i­tor in­di­vid­ual users and then adapt their con­tent, de­liv­ery mech­a­nisms, and com­mu­ni­ca­tions to meet the ever-chang­ing needs of each and ev­ery per­son

• Ac­cept­ing the fact that they need to be ev­ery­where their au­di­ence is, and their au­di­ence is ev­ery­where The most suc­cess­ful dig­i­tal com­pa­nies in the world al­ready do all of this. Think: Ama­zon, Face­book, Ap­ple, Google, Net­flix, and Spo­tify, to name just a few.

Take for ex­am­ple, Spo­tify’s Dis­cover Weekly playlist — a weekly col­lec­tion of songs se­lected specif­i­cally for you. It is a great ex­am­ple of de­liv­er­ing per­son­al­ized con­tent to an au­di­ence, not based on de­mo­graph­ics, but on an in­di­vid­ual’s mu­sic tastes.

Spo­tify’s cu­ra­tion al­go­rithm is a learn­ing ma­chine that uses one’s lis­ten­ing his­tory, and that of other Spo­tify users with sim­i­lar pref­er­ences in mu­sic, to gen­er­ate for each a unique Dis­cover Weekly playlist. The more a per­son uses the ser­vice, the bet­ter Spo­tify gets to know what that user likes and doesn’t like. What a great way to grow an amaz­ing user ex­pe­ri­ence over time and build loy­alty!

It’s no won­der Spo­tify con­tin­ues to grow in pop­u­lar­ity world­wide, with over 60 mil­lion pay­ing sub­scribers — more than any other stream­ing me­dia ser­vice, in­clud­ing Ap­ple Mu­sic. Spo­tify’s al­go­rithm uses a com­bi­na­tion of three con­tent rec­om­men­da­tion mod­els.

The first is col­lab­o­ra­tive fil­ter­ing, which con­tin­u­ally mon­i­tors what one lis­tens to so it can open up a whole new world of rel­e­vant songs they may never oth­er­wise dis­cover.

Sec­ond is Nat­u­ral Lan­guage Pro­cess­ing, which scours the web for writ­ten con­tent about mu­sic to learn what oth­ers are say­ing about spe­cific songs and artists, and about the per­form­ers dis­cussed along­side them.

Fi­nally, Raw Au­dio Mod­els an­a­lyze new tracks that may not yet be dis­cussed on­line or have many lis­ten­ers. It’s this in­dis­crim­i­nate anal­y­sis of con­tent that al­lows un­known artists to show up next to pop­u­lar mu­si­cians in Dis­cov­ery Weekly.

Al­go­rithms for cu­rat­ing con­tent based on ma­chine learn­ing are ev­ery­where and ev­ery con­tent ag­gre­ga­tor uses them. Why? Be­cause peo­ple pre­fer al­go­rith­mic cu­ra­tion over hu­man ed­i­tors.

But what about peo­ple’s fear of giv­ing ac­cess to their per­sonal in­for­ma­tion to these sites? Ac­cord­ing to a 2016 study by Ac­cen­ture:

• 70% of peo­ple are gen­er­ally com­fort­able with news sites col­lect­ing per­sonal data if the pub­lisher is trans­par­ent about how it uses it

• 75% are com­fort­able if they can per­son­ally con­trol how it was be­ing used

• 68% are highly sat­is­fied with stream­ing video ser­vices’ (e.g. Net­flix) use of per­sonal data (even though there is lit­tle trans­parency or user con­trol in them) be­cause it helps peo­ple dis­cover shows they like

There is a fine line be­tween us­ing per­sonal data for the good of an au­di­ence and abus­ing it for profit. Hav­ing rel­e­vant and qual­ity con­tent cu­rated for you is mag­i­cal; hav­ing ads fol­low you around the web, for weeks on end, is just plain creepy.

Mis­sions mat­ter

Many pub­lish­ers think that news con­tent is too ephemeral for a Spo­tify-like dis­cov­ery al­go­rithm, but I beg to dif­fer.

Just take a look at Face­book’s cu­rated news­feed. Its al­go­rithm is con­tin­u­ally be­ing en­hanced to pro­vide the most en­gag­ing ex­pe­ri­ence for its users — one that will keep them on the site longer and com­ing back for more. Face­book news is as fleet­ing as a Snapchat photo.

Face­book’s mis­sion is to “give peo­ple the power to build com­mu­nity and bring the world closer to­gether.” It puts its mem­bers at the top of its agenda in ev­ery­thing it does. In Oc­to­ber 2017, the av­er­age time spent on Face­book per per­son per visit was 20 min­utes. Com­pare that to the

top US news­pa­per web­sites had an av­er­age visit time of less than 2.5 min­utes. The numbers speak for them­selves, so why aren’t more pub­lish­ers lis­ten­ing?

Dig­i­tal is just an en­abler to con­nect

I have lost count of the num­ber of times I’ve heard pub­lish­ers re­fer to them­selves as “dig­i­tal-first.” As much as I ap­pre­ci­ate them fi­nally com­mit­ting to 20th cen­tury tech­nol­ogy 10 years af­ter the in­ter­net be­came main­stream, it still amazes me that they don’t re­al­ize that their fo­cus is mis­placed. In the 21st cen­tury con­tent is not king and nei­ther is tech­nol­ogy; read­ers are the rulers on this peo­ple-pow­ered planet and they must come first.

Whether you’re in so­cial, search, ser­vices, man­u­fac­tur­ing, con­sumer pack­aged goods, health­care, ed­u­ca­tion, pe­tro­leum, fi­nance… the list goes on, or pub­lish­ing, un­less you see your com­pany as peo­ple busi­ness whose mis­sion is to make life bet­ter for real peo­ple, you’ll have missed the prover­bial gravy boat.

The in­ter­net con­nected us in ways we never dreamed about — a con­nec­tion that is not just vir­tual, it’s so­cial. If you want to be a suc­cess­ful dig­i­tal com­pany then for­get the tech­nol­ogy and fo­cus on what re­ally mat­ters — peo­ple. Start a so­cial move­ment that be­gins by con­nect­ing them with each other and with you through qual­ity con­tent.

Be­cause de­mo­graph­ics won’t help you reach your ideal cus­tomer; only a ded­i­ca­tion to con­tin­u­ally con­nect­ing with, and de­light­ing, in­di­vid­u­als will.

If you want to learn more about how ma­chine learn­ing can help you con­nect peo­ple through news, let’s talk!

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.