Journalism in the age of al­go­rithms, plat­forms and news­feeds

The Insider - - FRONT PAGE -

Re­mem­ber how sim­ple life used to be? It didn’t seem that long ago when we got up in the morn­ing, grabbed a cup of cof­fee and sat down to dis­cover what hap­pened in our com­mu­ni­ties overnight – all of it cu­rated by ed­i­tors of the printed news­pa­pers de­liv­ered to our doorsteps rain or shine.

Sadly, for legacy pub­lish­ers hop­ing for a resur­gence of the gold old ways of print, those days are over. Only 5% of 18- to 29-year-olds of­ten get news from a printed news­pa­per. The 48% that grav­i­tate to­wards print (i.e. those 65 and older) are a dy­ing breed.

Sure, there will be a niche au­di­ence that will con­tinue to go through the nos­tal­gic daily read­ing rit­ual, but they will not be fully en­gaged in to­day’s hy­per-con­nected world – a world where peo­ple spend more time look­ing down at that shiny new thing in their hands than at the peo­ple around them. It’s no won­der that on my re­cent visit to Hong Kong, I noted a new ad­di­tion to all of their travel­la­tors and es­ca­la­tors: signs that prompt the mo­bile-savvy pop­u­la­tion to keep their eyes off of their smart­phones and look up.

But those who think this planet is head­ing to­wards dig­i­tal de­pra­va­tion of our youth, and that ev­ery step for­ward is three steps back, are wrong.

To­day’s dig­i­tal na­tives have the whole world at their fin­ger­tips. They aren’t lim­ited by the perime­ters of the printed page and they’re not con­trolled by the of­ten-bi­ased con­tent main­stream me­dia deems im­por­tant.

The phys­i­cal world we live in is lim­ited by what we can see, feel and hear. The world through the 3D lens of in­ter­net and mo­bile tech­nolo­gies is one we can “ex­pe­ri­ence” - full of dis­cov­ery, amaze­ment and hope.

Tech­nol­ogy’s awak­en­ing of won­der has spawned a gen­er­a­tion of in­quir­ing minds - non-con­form­ists who ques­tion more than ac­qui­esce, ex­plore more than ig­nore, ex­per­i­ment rather than play it safe and par­tic­i­pate more than just ob­serve. And prob­a­bly the most no­table of all char­ac­ter­is­tics these new world cit­i­zens share is that they are less trust­ing of oth­ers than any other gen­er­a­tion be­fore them.

Take a look at the fi­nan­cial sec­tor…

Numer­ous stud­ies have shown that GenYs and Zs are more likely to trust an al­go­rithm (aka Robo-Ad­vi­sor) than a fi­nan­cial ex­pert with their money.

They find it where al­go­rithms, as op­posed to (or, per­haps, in ad­di­tion to) ed­i­tors, serve their need for con­nect­ing with oth­ers through con­tent – con­tent that is cre­ated by both users and the me­dia.

The role of al­go­rithms in Journalism

It is the job of con­tent cu­ra­tion al­go­rithms to show peo­ple what’s

rel­e­vant to them. And with Face­book’s lat­est move to give pri­or­ity to per­son­ally-in­for­ma­tive sto­ries and posts shared by friends and fam­ily over tra­di­tional pub­lish­ing, it’s been made abun­dantly clear that what’s rel­e­vant to users isn’t what me­dia au­thor­i­ties think.

But news­feed al­go­rithms are only as good as their

ar­chi­tects. Here are just a few we en­counter on a reg­u­lar ba­sis, each with their unique cu­ra­tion of me­dia for news con­sumers.

Face­book – LIKEs to fool around

When one thinks of news­feed al­go­rithms, Face­book is prob­a­bly the first com­pany that comes to mind be­cause it has been no­to­ri­ous in its “tweak­ing” of its al­go­rithm since it launched its news­feed in 2006. I would even sug­gest that they are also the most in­fa­mous al­go­rithm de­sign­ers on the web to­day.

For those brands who have been burned by their bait-andswitch tac­tics, you know what I mean. For years, brands in­vested mil­lions of dol­lars cre­at­ing pages on Face­book to grow their fans. Sud­denly all their work was for naught as Face­book de­cided that the fan base was at the point where it needed to be mon­e­tized – not for the brands’ ben­e­fit, but for Face­book’s alone.

With a click of a key­board, Face­book re­duced the reach of page posts to a mere 1-2% of a brand’s fans. To ex­pand the reach, brands needed to start pay­ing the piper.

Fool me once - shame on Face­book.

Once Face­book got brands used to the no-free-ride re­al­ity, the next thing the com­pany needed to do was grow con­tent. So Zucker­berg made an of­fer that pub­lish­ers couldn’t refuse, “Put your con­tent on our site and we’ll share it with read­ers and help drive traf­fic to your web­sites. Use In­stant Ar­ti­cles to post full con­tent ar­ti­cles and we’ll share ad­ver­tis­ing rev­enue with you.”

I am still rather shocked by the naïveté of news ex­ec­u­tives and their blind trust in the provento-be un­trust­wor­thy. In June 2016 the bot­tom fell out when Face­book started rank­ing pub­lisher con­tent less than that of users. That didn’t sur­prise me in the least. What did sur­prise me was the me­dia ex­ecs who ac­cused Face­book of rob­bery.

Fool me twice - shame on us!

Brands fed Face­book ad­ver­tis­ing rev­enues while pub­lish­ers do­nated their “paid con­tent” for free and now they are yelling, “Foul!”. Oh and let’s not for­get the mil­lions of par­tic­i­pa­tory read­ers pub­lish­ers drove away from their dig­i­tal prop­er­ties to share com­ments and opin­ions on so­cial me­dia. It’s no won­der the so­cial gi­ant’s net rev­enue last quar­ter topped US$2.3B while pub­lish­ers’ tanked.

Face­book is a bril­liant ar­chi­tect of al­go­rithms so if there was one piece of ad­vice I could give to pub­lish­ers, it would be this, “Don’t be fooled again!”

Google – The keeper of search se­crets

Ev­ery year, Google up­dates its search al­go­rithm hun­dreds of times. It hap­pens so of­ten that most al­ter­ations go un­no­ticed. Fa­mous for chang­ing search laws un­der the cov­ers and an­nounc­ing them later, the ruler of rule changes has cre­ated fear, un­cer­tainty and more fear in the hearts of SEO man­agers.

But in terms of what it means for pub­lish­ers, it’s an on­go­ing bat­tle to keep re­fer­rals com­ing from search. Ear­lier this year, Google started pun­ish­ing what it con­sid­ered to be “old” con­tent. Branded and high-vol­ume key­words (e.g. Net­flix, mil­len­ni­als, etc.) which once drove mas­sive traf­fic to trusted pub­li­ca­tions re­sulted in in­vis­i­bil­ity for brands that used them be­cause they were sud­denly con­sid­ered to be out­dated or in­sti­ga­tors of click­bait.

I’m not against al­go­rithm changes to im­prove user ex­pe­ri­ence, but it would be nice if Google gave us the car­rot first - a heads up on rule changes be­fore they took ef­fect so we could learn and act in ac­cor­dance with what’s now good with Google. Give us the chance to avoid the stick of re­duced vis­i­bil­ity and rev­enues when we break rules we didn’t even know about.

Twit­ter – The news breaker

Al­though Twit­ter has been around al­most as long as Face­book, it’s been much slower in al­ter­ing its news­feed al­go­rithm. It did re­design its UI on many oc­ca­sions with stream op­tions like Ac­tiv­ity, Dis­cover and Men­tions and their “while you were away” fea­ture. And now it al­lows slightly

more con­tent in a tweet, but it wasn’t un­til the spring of 2016 that it fol­lowed in the foot­steps of Face­book by re­plac­ing its re­versed chrono­log­i­cally-or­dered time­line with an al­go­rith­mi­cally-pow­ered “best tweets first” queue.

How­ever, un­like Face­book, the new de­fault al­go­rithm can be turned off (for now). It’s most likely due to the fact that Twit­ter has be­come the de facto stan­dard for break­ing news and emer­gency com­mu­ni­ca­tions where the or­der of the posts is crit­i­cal.

Snap Inc. – The in­ven­tor of im­per­ma­nence

Per­haps the hottest so­cial net­work on the planet to­day, the “now you see it - now you don’t” plat­form is where pub­lish­ers with younger au­di­ences strive to be – so much so that they are will­ing to pay to play in Snap’s Dis­cover through guar­an­teed ad­ver­tis­ing min­i­mums.

What’s hard to wrap one’s head around is the fact that traf­fic is a one-way street from pub­lish­ers to Snap. Snap never links back to a pub­lisher’s web­site. It feels more like old-style broad­cast TV than in­ter­ac­tive so­cial – to­tally con­trary to what one might think is needed to be suc­cess­ful in dig­i­tal dis­tri­bu­tion. But it seems to be work­ing for pub­lish­ers like Cos­mopoli­tan.

In June 2016 Snap joined “who has the best army of al­go­rithm ar­chi­tects” with its over­haul of its Dis­cover al­go­rithm, al­low­ing users to sub­scribe to chan­nels and putting more fo­cus on con­tent than on me­dia brands. Some say this move has led to more click­bait con­tent and loss of brand affin­ity, while oth­ers see it as a new fric­tion­less dis­cov­ery mech­a­nism that will drive more ca­sual users to sto­ries they might not oth­er­wise have read. I’d say they’re both right.

The US$16B-val­ued dar­ling of Fidelity also launched a new ad­ver­tis­ing API in July, fol­lowed up with be­hav­ioral tar­get­ing.

All these changes seem to be a sign of even more as the newly re­branded pre­pares to head down the path to an IPO. It’s al­ready sur­passed Twit­ter in terms of ac­tive users and ex­pects to hit US$300M in 2016 in rev­enue, so one can’t help but think it won’t be long in com­ing.

I can see why mag­a­zines high in vi­su­als want to be on Snapchat, but the ques­tion still re­mains in my mind, “What’s re­ally in it for news­pa­pers in the long term?”

In­sta­gram – Face­book’s Mini-Me

Like Twit­ter, this in­stant pho­to­shar­ing app is not so “in­stant” any­more. In March 2016 it started “op­ti­miz­ing” time­lines based on users’ re­la­tion­ships and in­ter­ests, not time. Un­like Twit­ter (and just like Face­book), there is no way to re­vert back to the old chrono­log­i­cal feed.

In July 2016, In­sta­gram launched its Sto­ries fea­ture in what some call a fee­ble at­tempt to steal eye­balls from Snapchat. Early re­sults would in­di­cate that it will not suc­ceed any more than Face­book’s at­tempt to de­throne Snapchat with Poke, Sling­shot and Bolt.

Again, the draw for mag­a­zines is some­what clear, but I don’t get the value propo­si­tion for news­pa­pers. Do you?

PressReader – Cu­ra­tor by the Crowd

In ad­di­tion to full-con­tent repli­cas of pub­li­ca­tions, PressReader au­to­mates its home feed of freemium news to en­sure that read­ers get easy ac­cess to their fa­vorite me­dia brands and the con­tent that feed their pas­sions, whether they love tech­nol­ogy, sports, fash­ion, and, yes, even knit­ting.

But, it also in­cludes in a user’s home feed the lat­est ar­ti­cles that have re­tained the in­ter­est of other read­ers the long­est. This de­liv­ers not only all the con­tent read­ers want, it pro­vides fric­tion­less dis­cov­ery of new qual­ity con­tent they might oth­er­wise never see if the al­go­rithm re­lied purely on their in­di­vid­ual be­hav­iors.

The good, the bad and the ugly of al­go­rithms

Al­go­rithms aren’t go­ing away be­cause they add value in terms ef­fi­cien­cies in the news­room and user ex­pe­ri­ence in news­feeds. But with all the good they pro­vide, there

are some gotchas.

The Good

We live in a world that is in­creas­ingly be­com­ing more au­to­mated. Many of to­day’s news­rooms in­clud­ing The New York Times, Forbes and ProPublica to name a few, are al­ready us­ing al­go­rithms to help jour­nal­ists pro­duce ar­ti­cles on busi­ness, sports, safety and ed­u­ca­tion.

Google Trends uses its anal­y­sis of search terms to show pub­lish­ers what’s trend­ing in near real time across var­i­ous re­gions of the world.

Even the As­so­ci­ated Press is us­ing ro­bot re­porters to cover mi­nor league base­ball.

The Bad

So, yes, al­go­rithms have their plusses, but should we trust them?

Al­go­rithms are only as ac­cu­rate as the data which drives them. And if the data is in­her­ently bi­ased, mis­in­ter­preted or in­com­plete, the re­sults could be any­thing from com­i­cal to cat­a­strophic. Just check out the em­bar­rass­ing re­sults of Face­book’s au­to­ma­tion of trend­ing top­ics. Even with the stock­pile of an­a­lyt­ics ac­cu­mu­lated from its 1.7B users, the largest so­cial net­work in the world still can’t ac­cu­rately

“There needs to be a fine bal­ance be­tween giv­ing peo­ple what they say they want, or ap­pear to want ac­cord­ing to be­hav­ioral an­a­lyt­ics, and giv­ing them what they might want or need if given the op­por­tu­nity to dis­cover.”

re­port what’s trend­ing in me­dia with­out hu­man in­ter­ven­tion.

The Ugly

The Pro­ceed­ings of the Na­tional Academy of Sciences pub­lished a study in 2015 that con­cluded that “Users [on Face­book] tend to ag­gre­gate in com­mu­ni­ties of in­ter­est, which causes re­in­force­ment and fos­ters con­fir­ma­tion bias, seg­re­ga­tion, and po­lar­iza­tion. This comes at the ex­pense of the qual­ity of the in­for­ma­tion and leads to pro­lif­er­a­tion of bi­ased nar­ra­tives fo­mented by un­sub­stan­ti­ated ru­mors, mis­trust, and para­noia.”

No minc­ing of words there!

There needs to be a fine bal­ance be­tween giv­ing peo­ple what they say they want, or ap­pear to want ac­cord­ing to be­hav­ioral an­a­lyt­ics, and giv­ing them what they might want or need if given the op­por­tu­nity to dis­cover. This has been one of the ar­gu­ments jour­nal­ists have been say­ing for years and I think most ra­tio­nal peo­ple would agree with them. I cer­tainly do.

Journalism in 2020

As data be­comes more re­li­able, al­go­rithms will be­come

more ac­cu­rate and use­ful to read­ers, busi­nesses and pub­lish­ers. But, can they make us nar­row-minded? Yes, if we let them con­trol us. Can they open our eyes to new pos­si­bil­i­ties and ed­u­cate us? Yes, if we’re open to re­ceiv­ing them.

Al­go­rithms are still in their in­fancy in terms of con­tent cre­ation and dis­tri­bu­tion, but they are pop­ping up ev­ery­where. Com­mod­ity news in fi­nance and sports are nat­u­ral choices for ro­bot journalism given their em­pha­sis on num­bers and sta­tis­tics (e.g. earn­ings per share or runs bat­ted in).

But by 2020 al­go­rithms will be­come a much more fun­da­men­tal part of the en­tire pub­lish­ing value chain from con­cept through to re­search, news gath­er­ing, writ­ing, record­ing, edit­ing, pub­lish­ing, dis­tri­bu­tion and mar­ket­ing.

Through so­phis­ti­cated data-min­ing and ma­chine-learn­ing tech­nolo­gies and the abil­ity to in­ter­pret nat­u­ral lan­guage, al­go­rithms will evolve from just be­ing data-driven to be­ing lin­guis­ti­cally-driven. At that point, it might be very hard to tell if the au­thor of an ar­ti­cle is a hu­man be­ing or a com­puter.

Kris­tian Ham­mond, co­founder of Nar­ra­tive Sciences, a com­pany that trans­forms data into nar­ra­tives peo­ple read for a num­ber of pub­li­ca­tions, be­lieves that 90% of news con­tent could be writ­ten with lit­tle or no hu­man in­ter­ven­tion by the mid-2020s. He also shocked more than a few peo­ple when he said that a com­puter would win a Pulitzer Prize within the next five years.

Now I’m not sure I’m quite as op­ti­mistic about the qual- ity of a robo-re­porter’s work, but I have no doubt that Ar­ti­fi­cial In­tel­li­gence (AI) will in­fil­trate journalism and pub­lish­ing in a ma­jor way within the next few years.

Where does that leave those who live and breathe by their phys­i­cal and men­tal labors – those over­worked and un­der­paid mem­bers of the press that bleed ink, sweat and tears to de­liver the in­ves­tiga­tive journalism we not only crave, we need?

Many hope that they will be lib­er­ated by tech­nol­ogy that re­moves the mun­dane from their pro­fes­sion. It re­minds me of what Leonard Brody, au­thor of The Great Rewrite said to me ear­lier this year about AI and the fu­ture of me­dia, “There’s go­ing to be a lot of AI in jour­nal­is­tic ac­tiv­ity. But we’re also go­ing to see a re­nais­sance of tra­di­tional journalism be­cause it, and the di­a­logue it cre­ates, are in­cred­i­bly im­por­tant.”

I strongly be­lieve in the need for high cal­iber news con­tent and journalism’s role in a safe and pros­per­ous democ­racy. But that doesn’t mean tech­nol­ogy and so­ci­ety can’t work in tan­dem to bring forth a new age where AI, drones, cit­i­zens and pro­fes­sion­als col­lab­o­rate to cre­ate and de­liver ac­cu­rate, un­bi­ased, qual­ity con­tent that would be missed by all of us if it were gone.

“There’s go­ing to be a lot of AI in jour­nal­is­tic ac­tiv­ity. But we’re also go­ing to see a re­nais­sance of tra­di­tional journalism be­cause it, and the di­a­logue it cre­ates, are in­cred­i­bly im­por­tant.” Leonard Brody, au­thor of The Great Rewrite

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.