Sign up to myFT Daily Digest to be the first to know about Facebook Inc news.
Ahead of the US presidential election in November, do Facebook’s powerful “microtargeting” advertising services threaten democracy?
Some marketing and digital experts say the company cannot be trusted with selling the services to politicians because of its response to Russian interference in the 2016 election.
“Advertising technology designed for the promotion of products and services has been weaponised for political messaging,” says Nigel Gwilliam, director of media affairs at the UK’s Institute of Practitioners in Advertising, calling for a “ban on microtargeted political advertising online” after the system was “demonstrably abused in the recent past” by operators using Facebook and other platforms.
Google has responded by limiting how campaigns can target ads, while Twitter and TikTok have banned political ads completely. Facebook however continues to allow its clients to serve political ads, with no fact-checking, to selected groups of its nearly 3bn users, based on characteristics ranging from their postcode to their political leanings.
The company’s founder and chief executive, Mark Zuckerberg, initially dismissed as “crazy” the idea that fake news had any bearing on the 2016 election. But then evidence came to light of co-ordinated interference, spearheaded by a Kremlin-backed outfit called the Internet Research Agency, which created fake accounts and pages on social media, and targeted disinformation or provocative material at people using keywords like “patriotism’’. While it spent a mere $100,000 on Facebook advertising, its ads reached 11.4m Americans, according to data shared by Facebook with Congress.
Facebook’s security chief at the time, Alex Stamos, acknowledged that the ads, about a quarter of which were geographically targeted, “appeared to focus on amplifying divisive social and political messages across the ideological spectrum”.
Some marketing experts question whether Facebook has disclosed all the information it has on these Russian campaigns. They note that the click through rates for some of the Russian adverts — the proportion of people who go on to click on advertisements — were far higher than the industry average: sometimes as much as 10 per cent versus an average click-through rate of 0.9 per cent for Facebook adverts across all industries, according to search marketing group Wordstream.
Some say the explanation could lie in microtargeting.
“The reported click-through rate is inconsistent with the general broad advertising targeting information which Facebook shared with Congress,” argued an online marketing executive who declined to be identified. “This apparent disconnect between the ads’ high impact and their low cost suggests a much higher degree of targeting than previously disclosed.”
Such critics ask whether Russian trolls had access to electorally sensitive US voter data, and targeted Americans using Facebook’s custom audiences tool which targets groups of individuals specified by the advertiser.
Facebook admits that the now notorious British analytics group Cambridge Analytica, wielding the data of 87m Facebook users it bought from researchers, did use the custom audiences tool in its work for the Trump campaign in 2016. Separately, information was stolen from Democratic party and Clinton campaign servers by Russian military intelligence.
Damian Collins, until December 2019 head of the UK parliament’s Digital, Culture, Media and Sport Committee, has questioned whether the Cambridge Analytica data made its way into the hands of the Russia disinformation operations, which might explain the high click through rate.
“It is possible that the Russians had the [Cambridge Analytica] data set,” says Roger McNamee, an early investor in Facebook who is now a vocal critic of the company and called the data the “best custom audience ever created for politics”. Cambridge Analytica whistleblower Christopher Wylie said he believed Russian intelligence did have access to the data.
“In order to swing an election you don’t need to blanket the entire country with Facebook ads the way, for example, Mike Bloomberg [the billionaire candidate for the Democratic nomination] is currently doing,” says the online marketing executive. “Combining granular data …with Facebook’s microtargeting capabilities, you could target a relatively small number of households in the key counties that make all of the difference.”
But Facebook denies that the Russian campaigns brought their own data set for targeting.
Asked by the Financial Times whether Facebook would allow a third party to look at the targeting instructions for the Internet Research Agency’s adverts, a spokesperson refused, saying the company had “turned over what we have” to Congress.
Targeting instructions are kept on Facebook servers in perpetuity for all advertising clients, with third party inspection rights granted to the likes of ad agencies and others involved in an advertising campaign.
Either way, critics are still sceptical. “I’m not convinced that Facebook shared all the information they had around data and political advertising in that period of time,” said Jason Kint, chief executive officer of Digital Content Next, a US trade association for online publishers, referring to the Internet Research Agency and related activities beyond paid advertising.
The key, argued the marketing executive, is for regulators to insist on verification and disclosure rather than try to limit rapidly evolving technologies, a task for which they are “ill-equipped”: “If Facebook is unwilling to take a basic step like providing inspection access to the Russian accounts, what would they be doing? Protecting the privacy rights of the Russian operatives?”
On the other side of the Atlantic, the European Commission, seen by many in the US as a more potent regulator of Big Tech, said Facebook’s “black box” of algorithms used in political advertising should be opened up to researchers and other third parties.
“There should be the possibility to check how the algorithms are designed and how they work and the people should have control over that,” said the Czech commissioner, adding she “had nothing against” microtargeting as “a fantastic instrument to sell more goods to consumers”.
The EU’s executive body would even consider asking to inspect targeting instructions kept on Facebook servers for all campaigns if foul play were suspected, an EU official said, adding that regulators are keen to discover information wherever it might be stored.
“The algorithmic amplification of outrage and fear, and microtargeting…the combination of those two things allows for manipulation of attention, but also choices,” said Mr McNamee.
Facebook says its decision is not financial; political advertising will account for just 0.5 per cent of its sales this year, or $400m, according to Mr Zuckerberg, who came under pressure from the campaign to re-elect Donald Trump not to drop a service the US president credits with aiding his 2016 victory.
Proponents argue microtargeting is a vital tool for allowing campaigns, big or small, to reach their audiences. “It’s a key tactic and most political campaigns . . . spend a lot of time and money building up their in-house email list of supporters and donors and volunteers, and that’s a hugely valuable asset,” said Grace Briscoe, vice-president of ad-tech platform Centro’s candidates and causes group. “Without that ability to target the sort of ‘potential’ and ‘somewhat likely’ support, if you have to blanket the entire market, you might have to spend twice as much to reach the people who want to.”
Others remain unconvinced. “It’s amazing to me how microtargeting is largely being justified on grounds that it’s “cost-effective”, retorts Ellen Weintraub, chair of the US Federal Election Commission. “You know what else is cost-effective? Child labor. Dumping sewage into rivers. Fraud.” In a recent Twitter post, she added: “I strongly urge Facebook to go back to the drawing boards and come back with something much more robust. This will not do.”
Get alerts on Facebook Inc when a new story is published