The trans­parency is­sue

With viewa­bil­ity rates as low as 40 per­cent on sup­pos­edly trust­wor­thy lo­cal sites, ram­pant ad fraud cost­ing mar­keters bil­lions and ads being served along­side ob­jec­tion­able con­tent, Damien Venuto looks at dig­i­tal ad­ver­tis­ing’s grow­ing PR prob­lem – and why

New Zealand Marketing - - News -

Damien Venuto looks at dig­i­tal ad­ver­tis­ing’s grow­ing PR prob­lem.

When the vic­tims of famed im­poster Chris­tian Ger­hart­sre­iter were asked how they could fall for the man’s claim that he was a mem­ber of the il­lus­tri­ous Rock­e­feller fam­ily, the re­sponse was gen­er­ally some vari­a­tion of ‘be­cause that’s what he told me’. The con­tact lenses given to us by hind­sight make this ra­tio­nale seem ridicu­lous now, but it’s a re­minder that the de­fault set­ting for hu­mans is one of trust.

Some­thing very sim­i­lar has hap­pened in dig­i­tal ad­ver­tis­ing. Some years back, search and dis­play in­tro­duced them­selves as a pair of so­phis­ti­cated gentle­men, who had ar­rived to change things around town. They came bear­ing brief­cases filled with the finest as­sort­ment of tools, de­signed to cut away the fluff and give mar­keters greater ef­fi­ciency.

Buoyed by well-crafted press re­leases, news of their ar­rival travelled quickly and be­fore long ev­ery ad­ver­tis­ing jour­nal­ist was writ­ing about the ben­e­fits of tar­get­ing cus­tomers through dig­i­tal.

At first, few mis­trusted the claims that were being made. And why would they? Dig­i­tal was, af­ter all, pre­sented as the first truly mea­sur­able media channel. Ev­ery ac­tion could, in the­ory at least, be traced as online users wan­dered from one page to the next, pro­vid­ing valu­able data per­fectly suited for ap­pli­ca­tion to online mar­ket­ing cam­paigns. Be­liev­ing the hype was al­most the ra­tio­nal thing to do.

What no one an­tic­i­pated, how­ever, was the im­pact Campbell’s Law would have on the in­dus­try.

In his 1979 paper ti­tled ‘As­sess­ing the Im­pact of Planned So­cial Change’, so­cial sci­en­tist Donald Campbell ar­gued “the more any quan­ti­ta­tive so­cial in­di­ca­tor is used for so­cial de­ci­sion-mak­ing, the more sub­ject it will be to cor­rup­tion pres­sures and the more apt it will be to dis­tort and cor­rupt the so­cial pro­cesses it is in­tended to mon­i­tor.”

Campbell used the ex­am­ple of a city’s crime rate to show that a re­duc­tion in the over­all num­ber of crimes recorded doesn’t nec­es­sar­ily in­di­cate an im­proved so­cial en­vi­ron­ment but in­stead serves to a re­flect the im­pact of pro­ce­dural changes in­tro­duced to make the num­ber look better. Po­lice may, for in­stance, down­grade the se­ri­ous­ness of cer­tain crimes or change the rules as to what should be re­ported.

The point Campbell makes is that when per­for­mance is mea­sured solely from a quan­ti­ta­tive per­spec­tive, it leads to gam­ing of the sys­tem. Par­tic­i­pants find ways to ac­cen­tu­ate the met­ric being mea­sured, de­liv­er­ing what ap­pears to be a good re­sult on the sur­face.

Busi­nesses dis­rupted by quants—or big data—are par­tic­u­larly prone to such gam­ing of the sys­tem. When met­rics be­come an in­di­ca­tor of suc­cess, players in the game start find­ing ways to de­liver the de­sired re­sults with­out much con­sid­er­a­tion of the over­all per­for­mance.

One game leads to an­other

As big data has dis­rupted ad­ver­tis­ing, we’ve seen var­i­ous forms of gam­ing the sys­tem play out across the in­dus­try. The first and per­haps most rudi­men­tary ex­am­ple would be the ex­am­ple of sell­ing likes online. The mo­ment en­tre­pre­neur­ial types re­alised that so­cial media likes were a com­mod­ity some brands and online in­flu­encers were will­ing to pay for, pack­ages of­fer­ing likes popped up online. It didn’t take long for brands to cot­ton onto the fact that mil­lions of likes were lit­tle more than a vanity met­ric, par­tic­u­larly use­less when fake. At­tempts to game the dig­i­tal sys­tem cer­tainly don’t end there.

Part of the rea­son why Google con­stantly up­dates its al­go­rithms is to stay a step ahead of SEO con­sul­tants try­ing to take short­cuts to reach the top of Google rank­ings. Glen Maguire, the founder of SEO firm Click­through, says that one of the ma­jor chal­lenges he faces when pick­ing up a new client is un­rav­el­ling all the dodgy back­links bought by “lazy op­ti­mis­ers” look­ing for a quick buck from a dig­i­tally naïve busi­ness owner. While this ap­proach works in the short-term, Google eventually cot­tons on and pe­nalises the client’s web­site. And by that stage, the SEO com­pany has likely moved on to its next mark, leav­ing the busi­ness owner with the mess.

How­ever, buy­ing links and likes re­mains only a mi­nus­cule part of the broader pic­ture. Undoubtedly, the most suc­cess­ful and in­tri­cate form of gam­ing the sys­tem is found in the black market ad fraud in­dus­try.

Bots cre­ated by so­phis­ti­cated crim­i­nal net­works across the world mimic hu­man be­hav­iour and il­lic­itly fun­nel bil­lions of dol­lars of ad spend into their cof­fers ev­ery year. The ‘Bot Base Line’ re­port pub­lished by the As­so­ci­a­tion of Na­tional Ad­ver­tis­ers (ANA) in the United States re­cently found that online ad fraud is likely to cost ad­ver­tis­ers an es­ti­mated US$6.5 bil­lion in 2017. This is down from US$7.2 bil­lion lost in 2016 but make no mis­take, the fraud­sters are only be­com­ing more in­ge­nious and to­day’s bots have be­come much better at mim­ick­ing hu­man traf­fic. Much of the same

The agency pitch cy­cle has dropped from around five years to three years and a big rea­son for this is to max­imise cost ef­fec­tive­ness or cost ef­fi­ciency. - John Mont­gomery

tech­nol­ogy that pow­ers AI in­ter­faces and dig­i­tal al­go­rithms is now also being used by crim­i­nals to make their bots look and smell a lot like hu­man click­ers. With this in mind, it’ll be in­ter­est­ing to see if that US$6.5 bil­lion es­ti­mate isn’t per­haps ad­justed at the end of the year, by which time in­ves­ti­ga­tors would’ve un­rav­elled some of the more elab­o­rate schemes sy­phon­ing away ad dol­lars right now.

The domino ef­fect

Part of the rea­son the ad fraud syn­di­cates have been so suc­cess­ful is due to the CPM (cost per 1,000 im­pres­sions) model upon which online ad­ver­tis­ing is of­ten sold. Fraud­sters have es­sen­tially taken ad­van­tage of the in­dus­try’s al­most sin­gu­lar ob­ses­sion with this met­ric and ex­tracted il­licit value out of it. But the CPM model hasn’t only in­spired ac­tion from fraud­sters; in fact, it has di­rectly in­flu­enced the way busi­ness is done, web­sites are made and how mar­keters choose their part­ners.

Groupm global ex­ec­u­tive vice pres­i­dent of brand safety John Mont­gomery goes even fur­ther, say­ing in an in­ter­view with NZ Mar­ket­ing that at least some of the trans­parency is­sues blight­ing dig­i­tal ad­ver­tis­ing are at­trib­ut­able to “slav­ishly fol­low­ing a low-cost principle and mea­sur­ing on CPMS”.

One place where this is al­ways ap­par­ent is when mar­keters pitch their media busi­ness, he says.

“The agency pitch cy­cle has dropped from around five years to three years and a big rea­son for this is to max­imise cost ef­fec­tive­ness or cost ef­fi­ciency,” he says.

“When­ever we have a new pitch, there’s a re­quire­ment to lower the CPMS and we then have to go to our media part­ners and ne­go­ti­ate very hard to bring the costs down. When those pitches hap­pen in a cat­e­gory, like a big con­sumer group, and one ad­ver­tiser does that, the other ad­ver­tis­ers of­ten fol­low in or­der to make sure their pric­ing is

com­pet­i­tive as pos­si­ble.”

Mont­gomery says this then has a knock-on ef­fect in that pub­lish­ers are forced to in­tro­duce new ad zones to make up the loss in ad­ver­tis­ing revenue. At best, this cre­ates an an­noy­ing ad-heavy ex­pe­ri­ence for the users or, at worst, it leads to ads being served on sec­tions of a web­site never seen by hu­man eyes.

To show how far the CPM game can be pushed, Mont­gomery and his team re­cently ran an ex­per­i­ment in which they con­structed a faux cam­paign en­tirely out of in­valid or un­view­able ads.

“You can take your CPMS down to zero if you want to but don’t ex­pect your ad to be seen,” he says.

When you ex­tend this is­sue across com­mer­cial cat­e­gories over a num­ber of years, you end up with what has to­day man­i­fested as a viewa­bil­ity cri­sis se­ri­ous enough for the IAB New Zealand to ded­i­cate an en­tire event to it in Fe­bru­ary.

The gen­eral con­sen­sus among the pan­el­lists was that it was im­por­tant for all the players in dig­i­tal mar­ket­ing to ac­cept the IAB’S oft-ridiculed stan­dard that an ad can only be con­sid­ered viewed if 50 per­cent of the pix­els are view­able for a du­ra­tion of one sec­ond in dis­play or two sec­onds for video.

At the time, Moat part­ner man­ager Dave Good­fel­low con­ceded the stan­dard might be lax but added that mar­keters can al­ways ask more of their providers.

“The im­por­tant thing is to find a base­line that makes it pos­si­ble for us all to work to­gether,” Good­fel­low said. “If there are peo­ple who want to go beyond that base­line and have a better qual­ity au­di­ence, then they should be able to do that.”

The point here is that IAB’S stan­dard is not an ob­jec­tive, but rather a min­i­mum re­quire­ment to en­sure all providers are play­ing by the same rules in the lo­cal market.

What makes the viewa­bil­ity is­sue so im­per­a­tive is that the prob­lem is ram­pant across both lo­cal and in­ter­na­tional sites. So per­va­sive is the prob­lem that ANZ head of dig­i­tal mar­ket­ing David Gas­coigne says he’s seen viewa­bil­ity rates as low as 40 per­cent on some of New Zealand’s big­gest web­sites.

“Imag­ine pay­ing TVNZ to air your TVCS and they come back to you, say­ing that they only man­aged to air 40 per­cent of them but they’re still charg­ing you 100 per­cent – it makes no sense at all,” Gas­coigne says.

Pay­ing for peace of mind

When­ever there’s a mess, you’ll have some­one will­ing to clean it—at a price, of course. And when it comes to the viewa­bil­ity de­ba­cle, ad­ver­tis­ers have a num­ber of clean­ing agents at their dis­posal.

Moat, In­te­gral Ad Sci­ence (IAS) and Coms­core all of­fer tools that gauge viewa­bil­ity, giv­ing ad­ver­tis­ers a sem­blance of se­cu­rity that their ads are at least mak­ing it into the pe­riph­eral vi­sion of a web user.

But this peace of mind comes with a price tag the mar­keter is re­quired to pay. The ser­vices of­fered by these providers are added to the al­ready-com­plex dig­i­tal media sup­ply chain and mar­keters can then choose whether or not they’re will­ing to pay the commission fees at­tached to hav­ing their im­pres­sions guar­an­teed as view­able.

What is some­what un­usual about this is that mar­keters are es­sen­tially foot­ing the bill for a prob­lem on the media owner side. And what fur­ther com­pli­cates the mat­ter is that there’s still lit­tle in the way of third-party mea­sure­ment when it comes to the online du­op­oly of Google and Face­book (and, in­creas­ingly, Snapchat).

In all other media chan­nels (ex­cept per­haps outdoor), media own­ers pay for at least some in­de­pen­dent re­port­ing of their view­er­ship or read­er­ship numbers, whereas the view­er­ship and usage numbers for the big in­ter­na­tional sites gen­er­ally come di­rectly from those big in­ter­na­tional sites.

To put this into con­text, imag­ine the in­dus­try up­roar if, for in­stance, the mag­a­zine or tele­vi­sion in­dus­try sud­denly told mar­keters that they would now be re­port­ing their own numbers and that mar­keters would have to pay Coms­core to en­sure the ads were ac­tu­ally air­ing in the right place.

So why the dou­ble stan­dard? Why are dig­i­tal providers al­lowed to mark their own home­work while other media own­ers chip in to pay the bill for in­de­pen­dent anal­y­sis? Part of it is down to mo­men­tum. Mar­keters and agencies seem will­ing to forgo cer­tainty to be on these dig­i­tal chan­nels sim­ply be­cause they are so pop­u­lar. And you only have to look at re­cent growth in revenue to see this ‘cri­sis’ isn’t harm­ing Google or Face­book (at least not yet). Also, it hap­pens to be a lit­tle more dif­fi­cult to mea­sure dig­i­tal au­di­ences than an­tic­i­pated.

“It’s tricky be­cause there will al­ways be vari­a­tions, even from third-party sup­pli­ers,” says Gas­coigne. “We see vari­ances in Google An­a­lyt­ics and Adobe An­a­lyt­ics numbers even with the same code being ap­plied to the same pages and ac­tions across the site. [It’s] all down to the way in which these two ven­dors col­lect the data.”

The dif­fi­culty of the task aside, ar­guably the most ob­vi­ous rea­son we don’t have in­de­pen­dently pro­vided numbers is that the likes of Face­book, Google and Snapchat have been re­sis­tant to al­low­ing third-party ver­i­fi­ca­tion of their au­di­ences, pre­fer­ring to keep their cards very close to their chests. And a few promi­nent voices across the in­dus­try, Mont­gomery’s among them, have raised the opinion that this breeds mis­trust.

“When there isn’t third-party mea­sure­ment al­lowed on a plat­form like Face­book, Snapchat and most of Google’s prop­er­ties, there’s al­ways a

Imag­ine pay­ing TVNZ to air your TVCS and they come back to you, say­ing that they only man­aged to air 40 per­cent of them but they’re still charg­ing you 100 per­cent – it makes no sense at all. - David Gas­coigne

con­cern about what’s go­ing on that we don’t know about,” says Mont­gomery.

That said, the dig­i­tal jug­ger­nauts are start­ing to give the other players a few mo­men­tary peeks, with one ex­am­ple being Face­book’s com­mit­ment to an au­dit with the Media Ratings Coun­cil.

While a good first step, Mont­gomery doesn’t be­lieve this is enough to en­sure ac­tual trans­parency in the in­dus­try.

“It’s af­ter the case and it’s also not real au­dit­ing,” he says, ex­plain­ing that a re­port re­leased once a cam­paign has run does lit­tle to give mar­keters any se­cu­rity in the fast-mov­ing dig­i­tal space.

For this rea­son, Mont­gomery is a pro­po­nent of ap­pend­ing what he calls “a brand-safety tag” to ads run­ning across so­cial media net­works.

“If we had a tag on those ads, then at least we would’ve known about it or we would have been com­plicit in it. With­out any con­trol, it’s an in­cred­i­bly un­com­fort­able sit­u­a­tion.”

A com­mon ar­gu­ment against al­low­ing a third­party to tag ads in Google, Face­book or Snapchat is that this could place the data of the sub­scribers in each of those chan­nels at risk. While a leak or pri­vacy breach would be ex­tremely dam­ag­ing to these so­cial media gi­ants, Mont­gomery be­lieves it’s well within the ca­pa­bil­i­ties of the com­pa­nies to keep the in­for­ma­tion se­cure.

“Even if we were al­lowed to ap­pend a tag to an ad, there would be ex­tremely strict agree­ments in terms of what would be al­lowed to be col­lected. Re­mem­ber, this is not per­sonal in­for­ma­tion… They’re not shar­ing names, ad­dresses or any per­sonal de­tails. What they’re shar­ing is where the ad ap­pears.”

Play­ing it safe

Know­ing where ads ap­pear is im­por­tant not only to en­sure viewa­bil­ity but also brand safety, the lat­ter of which is par­tic­u­larly per­ti­nent at a time when count­less brands are being called up for ad­ver­tis­ing along­side ob­jec­tion­able ma­te­rial.

In­ter­est­ingly, the in­tro­duc­tion of a brand safety tag might not only be in the best in­ter­ests of clients but also the media own­ers. As seen in April when myr­iad ad­ver­tis­ers sus­pended ad­ver­tis­ing on Youtube and Google’s dis­play network as it came to light that their ads were ap­pear­ing along­side ex­trem­ist and misog­y­nis­tic con­tent, brand safety is an area ad­ver­tis­ers aren’t will­ing to ne­go­ti­ate on.

Since this scan­dal, Youtube has taken dras­tic steps, flag­ging large swathes of its online cre­ators as ‘non-ad­ver­tiser friendly’ in a bid to coax ad­ver­tis­ers back to the plat­form.

Gas­coigne says he un­der­stands why so many ad­ver­tis­ers are un­will­ing to risk their rep­u­ta­tions on the channel.

“If they don’t have the skillset and the un­der­stand­ing of how to man­age trans­parency and brand safety, you’ll find a lot of busi­nesses are turn­ing and run­ning, and that’s re­ally the feel­ing I’ve had from the in­dus­try of late.”

He says ANZ is still us­ing the channel but has taken a slightly more cau­tious ap­proach.

“At the mo­ment we stick with premium place­ment. We’ve got a 200-strong whitelist of top-per­form­ing and safe Youtube chan­nels that we run a lot of our ac­tiv­ity on.”

This still gives ANZ scale, but it doesn’t nec­es­sar­ily bode well for the channel as a whole.

Part of what makes Youtube so pop­u­lar with younger con­sumers is that it of­fers edgier, racier con­tent that might not be con­sid­ered safe from a tra­di­tional brand­ing per­spec­tive. And if these con­tent cre­ators are pre­cluded from mak­ing money on ac­count of strict en­force­ment of brand safety rules, then they’ll ei­ther have to find a new site that of­fers mon­eti­sa­tion or stop producing their videos. Nei­ther of these sce­nar­ios is ideal for Youtube—or for brands look­ing to en­gage with younger con­sumers.

Chang­ing the con­ver­sa­tion

De­spite trans­parency and brand safety is­sues run­ning ram­pant across the in­dus­try, mar­keters aren’t giv­ing up on dig­i­tal. In­stead, they’re find­ing ways to mea­sure the ef­fec­tive­ness of cam­paigns with­out pay­ing at­ten­tion to the dodgy met­rics float­ing around.

The rea­son why we haven’t seen a mass ex­o­dus of ad­ver­tis­ers from Face­book, de­spite its ad­mis­sion of a string of mea­sure­ment er­rors, is that ad­ver­tis­ers aren’t all that wor­ried about ba­sic view met­rics as long as the cam­paign still de­liv­ers the re­sults ini­tially in­tended.

“With any dig­i­tal media channel, you need to make sure the budget you are spend­ing is de­liv­er­ing to the busi­ness ob­jec­tive,” says Gas­coigne. “You need to take in­sight from met­rics like views and en­gage­ment but judge the suc­cess of the ac­tiv­ity in an ac­tion. For us, the im­por­tance of our cost-per-ac­qui­si­tion in­di­ca­tors are the num­ber one pri­or­ity, and yes we look at cer­tain met­rics with a ‘lens’, be­cause at the end of the day you need to see the re­turns.”

In much the same way that online dis­cus­sions have shifted from im­pres­sions to view­able im­pres­sions and then to safe view­able im­pres­sions, we’re start­ing to see mea­sure­ment lan­guage evolve.

Rather than talk­ing about reach­ing mil­lions of peo­ple, mar­keters and pro­gram­matic providers are in­creas­ingly talk­ing about re­turn on in­vest­ment and at­tri­bu­tion mod­els that claim to show the link be­tween an ad and a con­sumer ac­tion.

The only ques­tion is how will these met­rics be gamed and how many mil­lions (if not bil­lions) of dol­lars will mar­keters lose be­cause of it?

Newspapers in English

Newspapers from New Zealand

© PressReader. All rights reserved.