Everything about you, your customer, your goals - and what should be the optimal combination for all. Comfortable handing over that much power to Facebook? You may not have a choice. JL
Burt Helm reports in the New York Times:
The
algorithm weighs what it knows about the company, the ad and the Facebook user. (It) draws inferences from personal interests, online
behavior, the user’s value to each advertiser and the ad’s appeal. Sometimes the winner is the advertiser that offered the most money. Sometimes the algorithm decides you are
likely to click a different ad and awards the space to that advertiser. Millions of auctions take place every minute. The algorithm
is constantly learning, using past results to inform how it weighs bids
in the next auction. No one would mistake Ben Cogan and Jesse Horwitz for “brogrammers,” the jockish male coders swaggering across the tech landscape. Nor are they hustlers, the relentlessly outgoing types who quit their jobs to gamble on audacious ventures. They are two bookish friends, ages 27 and 29, who until recently lived across the street from each other on the Upper West Side of Manhattan. Horwitz worked for Columbia University’s endowment fund; Cogan had a job analyzing consumer behavior. Their hobbies are quiet. Cogan dreams of earning a Ph.D. in philosophy someday — “after all this is said and done,” he says. Horwitz enjoys tracking various aspects of his life in Excel spreadsheets: restaurants visited, books read, jogs taken. Scrolling through those files, he says, fills him with a sort of data-based nostalgia. For years, the two men met for dinner every week or so, where talk often turned to business ideas. Spit-balling plans for start-ups became their equivalent of fantasy football. One night in the summer of 2015, over Sichuan at Han Dynasty on 85th Street, Cogan asked Horwitz for advice about his latest notion: selling contact lenses online. The contacts business was dominated by a handful of companies like Johnson & Johnson and Bausch & Lomb, which seemed to charge whatever they wanted — at least in Cogan’s view, based on the price increases for his own lenses. Surely a low-cost competitor could tempt away customers. Cogan pulled his laptop from his bag and opened it at the table in the middle of dinner, pushing aside plates of dumplings and scallion pancakes. He had two plans to show Horwitz. They could sell a cheap disposable lens to doctors. Or they could mimic Cogan’s employer, a wildly successful start-up called Harry’s. By late 2015, Harry’s, which sold safety razors and shaving cream, was in the vanguard of upstart online retailers known as direct-to-consumer companies. The business model works like this: Firms sell only their own products, only through their own websites. By cutting out retailers and distributors, they can charge less for their specialty products than entrenched competitors. That year, Casper, then a direct-to-consumer mattress and bedding seller, was reported to be on track to exceed $100 million in sales in its second year of business. Dollar Shave Club, another seller of razors, had reached $152 million in revenue. Warby Parker, the eyeglasses purveyor that many credit with pioneering this business model in 2010, had recently closed an investment round that reportedly valued the company at $1.2 billion. Venture capitalists — convinced that consumers would increasingly patronize specialty online retailers as they grew more comfortable shopping online — were pouring money into direct-to-consumer start-ups, more than $2 billion over the past four years, according to CB Insights.Thanks
largely to Facebook, Hubble is on track to finish its first full year
in business having made $20 million in revenue. In August, Hubble raised
$10 million, valuing the company at $210 million. In January, Hubble
will use those funds to expand its business to Continental Europe. Its
advertising strategy? Robo-Dan, with some help from Rosen. As Hubble
advances into new territories, Facebook and the algorithm will be
tagging along with them.Thanks
largely to Facebook, Hubble is on track to finish its first full year
in business having made $20 million in revenue. In August, Hubble raised
$10 million, valuing the company at $210 million. In January, Hubble
will use those funds to expand its business to Continental Europe. Its
advertising strategy? Robo-Dan, with some help from Rosen. As Hubble
advances into new territories, Facebook and the algorithm will be
tagging along with them.Horwitz,
by now bored in his job, pushed Cogan to keep pursuing his idea and
volunteered to help him do research. “It’s hard to intellectualize
whether an idea is good or not,” Horwitz says. “You have to just start
doing it and see.”Thanks
largely to Facebook, Hubble is on track to finish its first full year
in business having made $20 million in revenue. In August, Hubble raised
$10 million, valuing the company at $210 million. In January, Hubble
will use those funds to expand its business to Continental Europe. Its
advertising strategy? Robo-Dan, with some help from Rosen. As Hubble
advances into new territories, Facebook and the algorithm will be
tagging along with them.
By February 2016, after many nights and weekends
of emailing Asian manufacturers and reading up on Food and Drug
Administration compliance, the vision of a viable business was coming
into focus. The pair had found an F.D.A.-approved manufacturer in Asia
and figured out how to meet the necessary regulations. Still, Cogan was
reluctant. He had been accepted to Wharton and had even put down a
deposit. He believed that was the smarter option. At best, the
contact-lens business would become a side project.
Before
shelving their venture, they decided to try one more tack. They
recruited two friends: Paul Rodgers, a buddy of Horwitz’s from Columbia
who knew how to write computer code, and Dan Rosen, a
friend
of Cogan’s from Bronx Science who was handy with Adobe Photoshop and
Illustrator. Together the four built what is known in the world of
online retailing as a demand experiment. The technique, credited to
Harry’s founders (who give away its basic code), amounts to a two-page
website. The first page explained the concept of a monthly subscription
for contacts and asked those who were interested to submit their email
addresses. Visitors who did so were taken to a second page and were made
an offer: Share this referral code with friends, and if enough of them
sign up, you’ll get free contacts.
They
posted a link to their site on the walls of about 40 Facebook friends.
Within a few days, not only had their own friends signed up, but friends
of friends of friends had, too — some 2,000 people in all. Some of
those distant connections were even evangelizing the company on their
own Facebook walls. “It went mini-viral,” Cogan says.
He and Horwitz
applied to tech incubators — organizations that invest in and coach
young companies in exchange for minority stakes — using the demand
experiment as one slide in their 16-page PowerPoint presentation. They
pitched a few venture capitalists based in New York as well. They
decided that if they were admitted to an incubator, they would work on
the project full time. If not, Cogan would go to Wharton. By April, they
had not only been called back for interviews with five incubators;
venture funds were also offering to invest a total of $3.5 million in
their idea.
Cogan
dropped his Wharton plans. He and Horwitz ordered 50,000 contact lenses
and, with Rosen as creative director and Rodgers as chief technology
officer, began working out of their investors’ offices, stacking boxes
and boxes of lenses along the walls by their desks. They eventually
named their enterprise Hubble, after the orbiting telescope that can see
into deep space.
Facebook
helped them succeed with their demand test; now it would generate their
first sales. During the summer of 2016, a friend of one of Hubble’s
prospective investors, a start-up veteran named Joshua Liberson,
recommended that the founders try a new type of Facebook advertising
called Lead Ads. No outside website was needed: Would-be customers
simply clicked a button on the ad to submit their email addresses,
directly from Facebook. Hubble directed its ads to ZIP codes in New York
and Chicago, where they had already signed up optometrists willing to
write prescriptions. After people clicked the ads, Horwitz emailed them
to coordinate appointments and take their orders.
When
Hubble’s online store opened officially on Nov. 1, 2016, Cogan and
Horwitz knew how to run a Facebook advertising campaign, and they were
confident it would continue to generate sales. They planned to spend the
additional $3.7 million they raised almost entirely on Facebook ads.
In 2017, everyone
seems to be wondering: Is Facebook taking over the world? Most of us
now realize that the social network has become far more than a
repository for selfies and political rants of its more than two billion
users. To ad sellers, Facebook is now a gluttonous monster, which, along
with Google, is gobbling up the digital advertising business in the
United States; according to Pivotal Research Group, the two companies
controlled 70 percent of the market and most of the growth in 2016. From
the perspective of American intelligence agencies, Facebook is
practically a weapon, used by a company linked to the Kremlin to foment
extremism and influence the 2016 presidential election with at least
$100,000 worth of targeted ads. For those with privacy concerns,
Facebook plays the role of Big Brother, compiling ever more data on what
we like, what we post and what we buy and even tracking where we are
both online and in the physical world by tapping into the GPS locator on
our phones.
In
considering Facebook’s far-reaching influence, it’s worth keeping in
mind the perspective of the more than five million advertisers whose
money is financing the social network’s rampant growth. For them,
Facebook and Instagram, which the company also owns, are the stuff of
fantasy — grand bazaars on a scale never seen before. By advertising
directly in users’ news feeds, companies can, at any time of day, target
potential customers at moments when they are often bored and open to
novelty. What better time to hear a product pitch?
“Facebook
created the world’s greatest infomercial,” says Roger McNamee, a
founder of Elevation Partners, who invested early in Facebook but has
since become critical of the company’s influence. “It’s really
inexpensive to produce ads and unbelievably inexpensive to reach exactly
the market that you’re looking for.” As a result, Facebook has become
especially lucrative for companies trying to sell new products online.
The leaders of more than half a dozen new online retailers all told me
they spent the greatest portion of their ad money on Facebook and
Instagram.
“In
the start-up-industrial complex, it’s like a systematic transfer of
money” from venture-capital firms to start-ups to Facebook, says Charlie
Mulligan, the founder of BrewPublik, which uses a “Beergorithm” to
deliver personalized selections of craft beers to customers every month.
At 500 Startups, the tech incubator based in Silicon Valley that funded
BrewPublik, Facebook advertising is a topic covered in classes. In
fact, social-network advertising is an assumed prerequisite for anyone
studying marketing at a tech incubator these days — or at any business
school across the country. “There is a formula for this stuff,” Mulligan
says. “And the reason why there is a formula is because it works.”
The
process is easy, cheap and effective. With a few hundred dollars and a
morning’s effort, an entrepreneur can place his or her ads before
social-media users that same afternoon. Companies unsure which ads are
best can upload a handful of them and let Facebook’s
artificial-intelligence software test their efficacy. If they don’t know
who should see their ads, they can embed code on their websites that
enables Facebook to monitor the traffic and then show ads to recent
visitors. Or companies can send the email addresses of their existing
customers to Facebook, and it will locate their Facebook accounts and
put ads in front of so-called Lookalikes, users who like and click on
the same things that your proven fan base does. It’s all about as
straightforward as setting up an online dating profile. Steph Korey, a
founder of Away, a luggage company based in New York that opened in
2015, says that when the company was starting, it made $5 for every $1
it spent on Facebook Lookalike ads.
The
ease of opening a business on Facebook has in turn spawned a wild
proliferation of specialty digital sellers that depend on the social
network’s algorithm to find their early customers. Many of them follow
the same playbook and even share a similar aesthetic. They spend money
on traditional public relations, on sponsored links that appear next to
Google search results and on “influencer” marketing, or giving away
their product to people with large social-media followings, in hopes of
creating buzz. Then they buy ads on Facebook and Instagram. Inevitably
you will encounter them there: They feature a sleek photograph or a
video loop of a product — a wood-handled water filter, woolen shoes, an
electric toothbrush. At the top, in bold, the company’s name appears,
often ringing with the same friendly, typically two-syllable whimsy.
Soma. Allbirds. Goby.
“Sometimes
we’ll look at each other and say, ‘God, there are just so many of
them,’ ” says Ellie Wheeler, a partner at the venture fund Greycroft
Partners, which invested in Hubble last year. Her firm has also taken
ownership stakes in Thrive Market, which sells health foods; Plated, a
meal-kit delivery service; Trunk Club, which mails a box of clothes to
its customers; and Eloquii, a fast-fashion retailer specializing in plus
sizes.
By
advertising directly in users’ news feeds, companies can, at any time of
day, target potential customers at moments when they are often bored
and open to novelty.
While
not all of these companies and others like them will survive, plenty
are encroaching on established brands, which are taking the threat
seriously. In July 2016, Unilever, the European consumer-products
conglomerate, acquired Dollar Shave Club for a reported $1 billion. In
June, Walmart agreed to buy Bonobos, an internet-based apparel brand,
for $310 million. Companies that sell products exclusively online
continue to grow faster than any other type of retailer in the United
States — some 17 percent annually since 2011, more than six times the
rate of retail over all, according to Euromonitor International.
And
Facebook has even been taking steps to influence offline sales, in
order to bring traditional retailers into its orbit. In September, the
social network introduced a tool that lets businesses with physical
stores show ads to shoppers and their Lookalikes even if they visit the
store but don’t buy anything. Day by day, Facebook is extending its
reach further and further into the American marketplace.
One afternoon in
March, I watched as Rosen selected three new ads from an extensive
photo shoot the week before, his third in four months. Rosen resembled a
sleep-deprived new parent — mussed hair, dull gaze. He spoke in a
monotone. He attributed his fatigue, I would learn later, to Facebook’s
artificial-intelligence software that placed Hubble’s ads. Rosen and his
colleagues simply referred to it as “the algorithm.”
The
basic building block of Facebook advertising is an ad set. It consists
of the ads themselves and choices in three other categories: audience,
goal and budget. That day, Rosen was designing a set to reach an
audience of people on Instagram who had visited hubblecontacts.com in
the past 30 days. His goal was “conversions,” or persuading users who
had seen the company’s ad to make a purchase. Finally, he set a budget
of $1,000 per day. He uploaded the three images. Now they were ready to
be tested, to see if any of them were winners in the eyes of users and
the algorithm.
What
happened at 8 a.m. the next morning, when the ad set became active, was
complex — and far removed from human sight. Just before Facebook places
an advertisement in a user’s feed, it holds a sort of instantaneous
auction to determine which advertiser gets the space. The amount of each
advertiser’s bid is influenced by its budget size, of course, but the
algorithm also weighs what it knows about the company, the ad and the
individual Facebook user. Seeking to act like an intuitive matchmaker,
the algorithm draws inferences from personal interests, current online
behavior, the user’s potential value to each advertiser and the ad’s
general appeal. Sometimes the winner is the advertiser that offered
Facebook the most money. Sometimes the algorithm decides you are more
likely to click a different ad and awards the space to that advertiser for less money.
This
detailed handicapping process involves thousands of advertisers per
auction. Millions of auctions take place every minute as users across
Facebook load their feeds. The process is never the same twice. The algorithm
is constantly learning, using past results to inform how it weighs bids
in the next auction. The intent, Facebook says, is to maximize value
for everybody: to pair the advertiser with its likeliest customers, and
to show ads that users want to see. And, of course, to make money for
Facebook.
But
from Rosen’s perspective, nothing much had happened before he ambled
into the office a little after 10 a.m. Facebook had spent a grand total
of $1.86 on his ads. It had shown the first ad to 51 people, the second
to 45 and the third to only two. The first ad had been clicked once.
Rosen, unperturbed, poured himself a cup of coffee from the single-serve
machine. The algorithm takes a little while to get warmed up, he said.
“In an hour, it’ll get exciting.”
Twenty
minutes later, Rosen refreshed his browser. The Ads Manager window
displayed the latest numbers: Rosen could see only the results, not the
process that produced them, but it seemed as if the click had inspired
the algorithm to favor the first ad. During those 20 minutes, the first
ad appeared before 76 more people — that is, it won 76 more auctions
than the other two ads. Over the next hour, the algorithm showed the
first ad, which featured a photo of colorful Hubble boxes against a blue
background, to more and more users; the algorithm had begun to favor
it, apparently. As Rosen refreshed his browser, the sensation was like
watching a seed sprout. The ad got more views. Some led to clicks. And
eventually, sometime between 11:28 a.m. and 11:53 a.m., one of those
clicks led to the test’s first sale. Commerce was in bloom.
The
moment felt odd. Obviously there was science behind the scenes; the
algorithm was a set of rules written by Facebook engineers. But from
where Rosen sat, the operation might as well have been run by the Holy
Spirit. Facebook’s artificial-intelligence algorithm had wound its way
through the server farms, reached out among two billion users, found an
individual and showed her a Hubble ad on Instagram — and she used her
credit card to buy a subscription for contact lenses.
In
quick succession, the first ad generated two more sales. The algorithm
started increasing how much it bid on Hubble’s behalf, thus winning even
more auctions for ad space and spending more of Hubble’s money on it —
first $1 a minute, then $2 a minute, then more than $3. By 2 p.m.,
Facebook’s A.I. had charged Hubble $306.50 to put that ad in front of
9,684 users. The second ad, after an outlay of $8.03, had been all but
abandoned. And the third ad was hardly given a chance: Since 8 a.m., it
had appeared before only 30 people.
“No
idea why,” Rosen said, shaking his head. Rosen could see all sorts of
data arranged in neat rows on Facebook’s Ads Manager program: the number
of views, clicks, sales and the average cost, in advertising, of
acquiring each new customer. But none of the metrics at Rosen’s
fingertips could resolve a fundamental mystery: why the algorithm
behaved as it did, why it preferred some ads over others and why the
third ad got little attention whatsoever.
The
morning’s ads were incredibly similar: “hubblecontacts,” the company’s
Instagram handle, appeared at the top, above pictures of boxes in peach,
blue, yellow and green. The only differences were that the first ad
showed the boxes of contact lenses lined up against a blue background;
in the second and third ads, they were set against a split pink-and-blue
background and were arranged diagonally in the second and scattershot
in the third. But they were all just boxes! Did Instagram users really
prefer contact-lens ads with strict rows of boxes or blue backgrounds?
Had rules been written into the algorithm favoring orderly arrangements?
(The Hubble team knew Facebook favored certain aesthetics.) To what
extent was the day’s outcome, apparently set in motion when the first ad
happened to get that first click in the morning, actually random? Rosen
could only guess.
Advertising has always
been an uncertain business. No one has ever known why, exactly, some
people respond to an ad in a newspaper or a spot on TV, much less why
specific individuals decide to buy products when they do. (The oldest
cliché in the ad world, usually attributed to the department-store
magnate John Wanamaker: “Half my advertising is wasted. The trouble is, I
don’t know which half.”) But to make money in advertising, you don’t
have to be all-knowing; your ads simply need to work better than those
of a competitor. To this end, advertisers inevitably pursue some
combination of two major approaches. They test and refine their
messages, trying to craft one as efficient and targeted as possible
(junk-mailers of preapproved credit-card offers, for example). Or they
showboat, putting on a huge spectacle that’s sure to attract someone (Super Bowl advertisers).
Anyone with a
credit card can go online and test ads on Facebook’s platform, one of
the most sophisticated direct-marketing operations ever.
In
the early 2010s, direct-to-consumer companies showboated. But lacking
the money for big TV ad campaigns, they relied instead on old-fashioned
public relations, panache and luck. Warby Parker hired a
public-relations firm to pitch its concept to Vogue and GQ and debuted
its website on the same day issues reached subscribers. It also held an
event featuring bespectacled models at the New York Public Library
during Fashion Week. Dollar Shave Club first succeeded on account of the
exquisite timing, both commercial and comedic, of its founder, Michael
Dubin. He made a funny, low-budget video introducing his company, then
uploaded it to YouTube on the same day TechCrunch reported Dollar Shave
Club’s first round of venture funding. Within days, after some immediate
attention at the South by Southwest festival in Austin, Tex., Dubin had
three million views online.
Facebook’s sales pitch — putting the
right ad in front of the right person, thanks to the wonders of data
technology — isn’t exactly new. As far back as 1964, William Allan, a
business editor for The Pittsburgh Press, reported that in the near
future, “computers will tell businessmen
which
half of their advertising budgets are being wasted.” Thirty years
later, The Economist described an effort to take advantage of American
Express’s transactional records: “Powerful data-crunching computers
known as massive parallel processors, equipped with neural-network
software (which searches, like the human brain, for patterns in a mass
of data), hold out a vision of marketing nirvana.” Companies like
Acxiom, Experian and Datalogix have been offering similar data-mining
services to direct marketers for years. What sets Facebook (and Google)
apart are scale and sophistication.
A
recent study by a Princeton professor, Arvind Narayanan, and a doctoral
candidate, Steven Englehardt, provides a sense of how thoroughly the
two online giants monitor user behavior. In early 2016, they examined
the top one million websites in the world, using special bots they
developed to scour them for tracking mechanisms. Google had trackers on
76 percent of these sites, Facebook on 23 percent of them. (Twitter, in
third place, had trackers on just over 12 percent of the sites.) The
tech giants can examine all this data looking for patterns and then
match them back to prospective customers.
What
also sets Facebook and Google apart from their direct-marketing
forebears is that they give access to everyday advertisers. Anyone with a
credit card can go online and test ads on Facebook’s platform, one of
the most sophisticated direct-marketing operations ever. But while
average people can use the machine, there’s still a lot of mystery about
how it works. The methods and calculations of the algorithm — why it
ends up pushing some ads and not others — are all hidden.
Almost
as soon as they began, Rosen, Horwitz and the others at Hubble became
determined to fathom the algorithm’s secrets — to figure out why some
ads succeeded and others didn’t. Soon they were trading hypotheses with
other entrepreneurs, cribbing ideas from other companies’ ads and taking
a formal approach to testing, rooted in the scientific method. They
uploaded ads with identical images but different wordings, for example.
The Hubble team wound up concluding all sorts of things. Ads with
third-party endorsements — from GQ, say, or BuzzFeed — beat those with
their own slogans. Ads featuring close-ups of the Hubble boxes
outperformed those with human models. Ads that included a button that
said “Shop Now” or “Learn More” fared worse than an ad with no button at
all; viewers simply preferred to click anywhere on the picture to go to
the website.
But
even as the Hubble team gleaned more about what yielded successful
Facebook ads, the algorithm could be unpredictable, almost moody. If you
kept loading the same ads into the same ad set every day, they stopped
performing as well. The founders figured at first that users were tiring
of the same ads. But actual viewer numbers revealed that practically no
individual user had seen any ad more than once. The algorithm itself
seemed to grow bored. At night, meanwhile, the algorithm spent lots of
money and rarely found customers. The Hubble executives started
shrinking the budgets at 11 p.m., which they called “putting the
algorithm to bed.” The algorithm could also be impulsive and streaky —
some days it might go on a sudden jag, blowing a thousand dollars in a
few hours with nothing to show for it. At any time, any one of the 15
different ad sets might go haywire. Rosen found himself checking the Ads
Manager compulsively on his laptop and his iPhone. (Facebook offers an
iOS app for advertisers.) “It occupies my brain constantly,” he says.
“It’s that feeling of ‘Did you leave the oven on?’ ”
One
night we went to a standup-comedy night Rosen hosted at a bar called
Muchmore’s in Brooklyn. (For the past four years, he has moonlighted as a
comedian.) But while the other comics were onstage, Rosen was on the
Ads Manager the whole time. “Who cares about jokes?” he quipped
afterward.
Eager
for help, Rosen sought guidance from a former Facebook employee named
Faheem Siddiqi, who now runs his own marketing agency. Hubble’s sales
representative at Facebook told him that Siddiqi had figured out the
best ways to optimize Facebook advertising campaigns. But it turned out
that Siddiqi and his employees checked the Ads Manager even more
compulsively than Rosen — every half-hour, for up to 16 hours a day.
When I asked Siddiqi to share his tips for managing Facebook ads, he
replied, “Step 1 is meditation.”
“It’s
like a baby,” Jesse Horwitz told me. “If you go more than half an hour
without checking in on it, it’s probably dead.” (Horwitz, who is
married, does not yet have children.)
Middlemen — creative
agencies, media planners, publishers — have long ruled the advertising
business. Yet until recently they have not been as omnipresent, opaque
and inhuman as Facebook. The social giant now dictates, more fully and
precisely than ever before, which ads we see and who sees which ads.
Some of the implications of this are amusing, others troubling. In my
house, the strange new world of advertising announced itself in the form
of a water pitcher. The Soma 6-Cup Pitcher is a paragon of Brooklynite
beauty: folksy oak handle, sleek minimalist reservoir, filter cones made
out of coconut shells (or something). I had never heard of it before my
wife ordered one online. Plenty of my friends hadn’t, either. When our
visitors opened the fridge half of them were like me: Soma ignorant. The
other half knew the brand immediately: Hey! You got a Soma? They had
seen the pitcher on Facebook, on Instagram, all over the place. What was
a familiar brand to some was totally unknown to me and others. We had
been divvied up. It’s something I’ve noticed again and again: I see an
ad for Aaptiv, a running app; my wife sees ads for a furniture website
called Article that I’ve still never visited. Just as Facebook steers
conservative and liberal talking points to users who already share those
perspectives, we’re being sorted into commercial bubbles as well.
Recently
ProPublica, the investigative-journalism nonprofit, showed how bad
actors can abuse this process: Facebook’s software gave advertisers the
option to target “Jew Haters,” for instance. In a separate
investigation, ProPublica found that Facebook made it possible to
exclude specific “ethnic affinities” from seeing ads, noting that ads
excluding people based on race are prohibited by federal housing and
employment laws.
This
stereotyping isn’t a glitch of Facebook’s machine-learning process —
it’s how the software works. To formulate audiences, the algorithm
scours profiles and analyzes them for shared traits and correlations and
self-identified interests and, it assumes, our preferences, grouping us
into tribes that can be targeted. It’s up to Facebook and advertisers
to constrain this amoral process in ethical and lawful ways. Yet the
ethics of targeting are not clear-cut. In May, The Australian reported
that Facebook employees had prepared a document showing how they could
gather details on teenagers during vulnerable moments — when Facebook
users feel “stressed,” “insecure,” “defeated” or “worthless.” Is that
immoral, or simply crass?
‘It occupies my brain constantly. It’s that feeling of “Did you leave the oven on?” ’
Such
challenges are opening a new front for companies and
corporate-responsibility watchdogs. Bad human actors don’t pose the only
problem; a machine-learning algorithm, left unchecked, can misbehave
and compound inequality on its own, no help from humans needed. The same
mechanism that decides that 30-something women who like yoga
disproportionately buy Lululemon tights — and shows them ads for more
yoga wear — would also show more junk-food ads to impoverished
populations rife with diabetes and obesity.
“Sometimes
data behaves unethically,” Antonio Garcia-Martinez, a former Facebook
employee who worked on the advertising team, wrote in an essay in The
Guardian. He provided an example from his time at the company: “Someone
on the data-science team had cooked up a new tool that recommended
Facebook pages users should like. And what did this tool start spitting
out? Every ethnic stereotype you can imagine.”
As
algorithms sort users in increasingly complex ways — already the
multivaried criteria for determining a Lookalike group defies human
comprehension — regulators and companies will have to confront how to
determine who is being nudged, and why, and whether that’s benefiting
the public or exacerbating societal ills. An algorithm that draws its
lessons from the present reality can’t be counted on to improve the
course of the future on its own.
Facebook’s
A.I. isn’t operating unattended, certainly: Garcia-Martinez wrote that
Facebook decided not to release the recommendation tool. Facebook points
out that it makes efforts to prevent harmful advertising. It does not,
for instance, allow ads for payday loans, which prey on the poor. It
says it has removed advertisers’ ability to target users by ethnicity
when promoting housing, employment or credit; it removed “Jew Haters”
and other objectionable categories and said it would increase human
review of its ad-targeting options. In response to the report in The
Australian, Facebook said the analysis “was intended to help marketers
understand how people express themselves on Facebook. It was never used
to target ads.”
Yet managing a platform this way — seeing what
mischief the algorithm and its users gets up to, then responding with
countermeasures — can be difficult to sustain. “This is a whack-a-mole
problem, one among many Facebook has,” Garcia-Martinez told me. It makes
Facebook, a company still largely controlled by a single man, Mark
Zuckerberg, the ultimate arbiter of morality and taste for all two
billion of its users. It also means the company has unilateral power to
make or break companies when it tweaks its system.
This
is not a hypothetical possibility. In 2013, media sites like those
measured by the BuzzFeed Partner network, which includes BuzzFeed,
Thought Catalog and The New York Times, noticed a huge surge in
referrals from Facebook — a jump of more than 50 million page views from
August to October. A year later, Facebook announced that it had
adjusted its news-feed algorithm to eliminate so-called click bait.
Upworthy, a peddler of stories with headlines like “9 Out of 10
Americans Are Completely Wrong About This Mind-Blowing Fact,” had its
total page views decline by half in the span of three months, from 90
million to 48 million visitors. (At the time of these huge shifts, 30
percent of Americans got news from Facebook. In 2017, 45 percent of
Americans do, according to Pew Research Center.)
“We
always knew that Facebook is sort of like the weather,” says Eli
Pariser, Upworthy’s co-founder and president. “There’s going to be sunny
days and stormy days.” In response to the algorithm adjustment, Pariser
instructed his staff to stop posting as many videos to YouTube, which
is owned by Google, and start publishing more videos directly to
Facebook instead.
“That
certainly served Facebook well,” Pariser admits. “But you know, I also
wouldn’t be able to reach 200 million people on any other medium,” he
says, citing the reach of Upworthy’s videos on Facebook. The platform
may be mercurial, but Upworthy still relies on it.
Imagine,
now, that Facebook tweaks its algorithm in a way that — rather than
cause wild swings in web traffic to a purveyor of viral videos — leads
to a steep decline in advertising and sales for a consumer-products
company, one that happens to be the largest employer in a small town. Or
imagine multiple companies shaken up by such an adjustment, or an
entire industry overhauling its practices to suit Facebook. Even the
threat and uncertainty of those possibilities could hurt
businesses, which depend on predictable returns to invest in future
projects.
As
we delegate more control to artificial intelligence, both businesses as
well as users are venturing into uncertain territory. In the meantime,
more and more companies — start-ups, mom-and-pop stores, major
corporations — are handing their dollars and their data to the
social-networking giant. Facebook’s Ads Manager is user-friendly. Sales
are plentiful. And if you don’t take advantage of it, your competitors
will. How could you not go there?
By mid-March, a
few weeks after I first followed Rosen, the Hubble team no longer had
15 Facebook and Instagram ad sets. It had 40 — all pointed at discrete
audiences, each with its own handful of ads. But Rosen looked more
rested, less frazzled. He explained that he and Paul Rodgers had
developed something they called “Robo-Dan,” a few lines of code that
checked the Ads Manager every hour, then adjusted the budget as Rosen
would. He could wake up and let the ads run (although he had to fight
the compulsion to check on Robo-Dan). Soon, he told me, they would
upgrade to Robo-Dan 2, which would switch in new ads, taking over the
nightly bedtime routine. (With 40 audiences, he told me, going through
the process lasted almost as long as an entire episode of “The Late Show
With Stephen Colbert.”) Finally, he said, he was getting some distance
from Facebook’s everyday machinations. Someday soon, he would be able to
go to bed early, he told me. Or have an evening to himself.
But
by the end of June, Rosen’s stress-free life was still a dream. A new
problem arose: No matter what new ads they put in an ad set, the growth
rate of sales declined and the cost per acquisition went up. They began
to think it was an audience problem: Had they found all the customers in
those groups? With their ad sets going fallow, the Hubble team
scrambled to find fresh and fertile ground. Their ideas for new
audiences got quirkier, more outlandish. One week they zeroed in on
people who like Sweetgreen, the salad chain. Next they went after people
who had indicated that they were fans of bottled water, whoever they
are. Each group fizzled after a few days — the cost per each new
customer climbed higher and higher; sales dwindled. As they searched for
more and more audience descriptors, they landed upon a novel idea: They
began trading their Lookalike groups with other online retailers,
figuring that the kind of people who buy one product from social media
will probably buy others. This sort of audience sharing is becoming more
common on Facebook: There is even a company, TapFwd, that pools
together Lookalike groups for various brands, helping them show ads to
other groups.
Cogan
and Horwitz have decided that they need to reduce their dependence on
Facebook advertising, for the sake of their business and their own
sanity. In May, they tested their first 15-second cable-television
commercials. With TV, the data is vaguer, and it takes longer to get
results back. Yet even though the old medium provides them with less
information than Facebook, in some ways the ignorance is bliss. “There’s
fewer levers; there’s less to stress out about.” Rodgers says. “You can
push the button and get on with your life.”
In
August, the Hubble team finally handed over their domestic Facebook
advertising work to an outside agency, Ampush, that charges them based
on how many new customers sign up. Ten people at Ampush now do the job
of Rosen and Robo-Dan. Still, the handoff was bittersweet. “We ran their
numbers — it’s something we could beat,” Rosen says, meaning Hubble
could get more customers for less money if it did the ad buying
in-house. “But it would destroy our lives.”
Thanks
largely to Facebook, Hubble is on track to finish its first full year
in business having made $20 million in revenue. In August, Hubble raised
$10 million, valuing the company at $210 million. In January, Hubble
will use those funds to expand its business to Continental Europe. Its
advertising strategy? Robo-Dan, with some help from Rosen. As Hubble
advances into new territories, Facebook and the algorithm will be
tagging along with them.
As a Partner and Co-Founder of Predictiv and PredictivAsia, Jon specializes in management performance and organizational effectiveness for both domestic and international clients. He is an editor and author whose works include Invisible Advantage: How Intangilbles are Driving Business Performance. Learn more...
0 comments:
Post a Comment