The ‘psychographics’ experts who swung the US elections will have a hard time solving Kenya’s tribal arithmetic
Cambridge Analytica gained notoriety for its role in Donald Trump’s presidential campaign last November. After Trump’s surprise victory, analysts began honing in on the previously obscure firm as the hidden factor in the US polls. CA was credited with identifying and responding to important trends that the Trump machine exploited to capture key Electoral College votes in swing states like Florida, Pennsylvania, and Ohio.
Their parent company, the SCL Group, which calls itself a “global election management agency,” cut their teeth by running military disinformation campaigns out of a base in Djakarta that resembles the operations centre set from the James Bond film, Goldeneye. SCL claims to have influenced the outcome of the Brexit vote by “supercharging” the Leave campaign’s voter turnout. More specifically, they identified communities harbouring anti-immigration sentiments and ran inflammatory ads depicting a future Great Britain overrun by foreigners.
Since that time their brilliant computer scientist and ultra-conservative American billionaire owner, Robert Mercer, has merged the companies and moved their operations to the US to avoid the more stringent legal conditions in the UK and the European Union.
For CA, the era of demographic-based marketing is over. Their more specialised approach instead focuses on individual “psychographics” over demographics. Their big data sets include 5,000 data points for each American consumer
Cambridge Analytica is now reportedly in Kenya working for the Jubilee Party campaign, and the implications go beyond whether their methods can tip the balance in a country like Kenya.
WHAT IS CAMBRIDGE ANALYTICA?
According to the controversial company’s spokesman, “We are a data analytics company. Many organisations out there want to serve as the database of record for their clients. We aim to be the layer on top of that, which provides our clients with actionable insights —essentially acting as the brain behind the decision-making process. At our core, we are a data and behavioural science company. Simply put, we help organisations figure out who to talk to and what to say to them.”
For CA, the era of demographic-based marketing is over. Their more specialised approach instead focuses on individual “psychographics” over demographics. Their big data sets include 5,000 data points for each American consumer, and this cache informed the company’s emotionally charged messages and images via advertising, direct mail, robocalls, and in-person canvass scripts delivered during the final three months before the US polls.
They varied dramatically depending on the voter, the issues, and diverse external factors like the condition of the local economy. A voter motivated by issues of immigration, for example, received a different variation on the Trump message than a gun-loving voter from the same demographic.
Their predictive analytics toolbox synthesises three core approaches: data science and analytics, digital marketing based on “persuasion,” and polling research. As a CA spokesman reported, “For the Trump campaign, we served as the data agency of record, but our role quickly evolved as the cycle progressed. Having a large amount of control and input allowed us to be extremely efficient and reactive. Our approach allows our clients to more efficiently spend their resources and better persuade and mobilise their advocates.”
The key word in the statement above is “persuade.” Persuasive technologies, the beating heart of the social media revolution, are why you click the Like button on Facebook and check the number of followers you have on Twitter. They also feed into the big data mantra: “We know what you want before you want it.”
Moi had his bag of tricks, but the fixers behind the 2007 poll employed real magicians who could conjure votes out of thin air and transfer them to computer printouts that were treated as official results. It was done so brazenly it could only be intended to provoke
KENYA’S CONTROVERSIAL 21ST CENTURY ELECTIONS
The 2002 election was relatively conventional by most standards; the peaceful transition to the Kibaki presidency validated this view. Prior to this the conduct of national elections appeared to be on a positive trajectory. But since 2007, the exercise has ended with controversy and acrimony.
Moi had his bag of tricks, but the fixers behind the 2007 poll employed real magicians who could conjure votes out of thin air and transfer them to computer printouts that were treated as official results. It was done so brazenly it could only be intended to provoke, a gambit devised by faceless schemers to legitimise the sitting government’s prerogative to restore law and order, and by doing so remain in power while lawyers argued over the count in court. But the bush fire intensified and spun out of control, requiring the formation of a coalition government to quench the flames.
While the catharsis of violence made Kenyans swear “never again,” it did not purge the sins of the political class. Parliament’s rejection of a local tribunal to sort out the mess and the invitation to turn the problem over to the International Criminal Court also backfired, or so it would appear. Come 2012, the Kenyatta-Ruto coalition employed British spin-doctors to craft a racially emotive narrative that turned the ICC trials and accompanying media circus to their advantage.
The press bought into the amani badala ya haki (peace instead of justice) meme, a blatant oxymoron coined to minimise another outbreak of 2007-style post-electoral violence. The national elections proceeded with the now familiar mix of skulduggery and double entendre aided by the tendering controversies perpetuated by the guardians of the process, the Internal Elections and Borders Commission.
The Ipsos Synovate polls of mid-January showed Raila Odinga with 46 per cent support to Uhuru Kenyatta’s 40 per cent, but the gap closed to within one to two percentage points in most polls by mid-February. A year earlier, the Raila Odinga-led ODM’s lead over the Uhuruto coalition ranged between 12 and 20 percentage points. Jubilee’s Uhuruto claimed a narrow 800,000 vote victory in March.
The disputed results ended up being arbitrated in Kenya’s new Supreme Court. The opposition lost out once again, arguably due to the methodical dismembering of the ODM case by the Jubilee team’s counsel, Fred Ngatia, whose pinpoint rapier thrusts shredded much of the opposition lawyers’ evidence.
Amani badali ya haki prevailed. Perhaps more importantly, it imparted a measure of legitimacy to the outcome. Despite the High Court’s ruling, a number of anomalies fuelled the counter-narrative explaining how the usual suspects rigged the endgame. The IEBC and their tendering antics featured prominently. Unlike their British partners in the Chickengate scandal who went to jail, the Kenyan IEBC commissioners were “retired” with a generous benefits package.
CAMBRIDGE ANALYTICA WAS ALREADY ON THE SCENE IN 2013
A succession of developments had helped the Jubilee campaign close what began as unbridgeable gap during the long run-up to the March 4 elections. As it turns out, Cambridge Analytica was already on the scene in 2013. Their political research project, based on a sample of 47,000 voters, pointed to young voters as an underutilised asset that could be highly influential if mobilised. CA tells us their communications and strategy team devised an online social media campaign that generated an active online following.
Now it’s polling season again, and because institutional uncertainty remains part of the equation, we should expect the unexpected
We have seen similar momentum shifts before and since. But in Kenya, the question, “How did they win?” is now as important as “Did they win?” Did CA’s contribution in 2013 lay down a marker for future electoral strategies, or did the narrative crafted by the BPL consultants make the critical difference?
Now it’s polling season again, and because institutional uncertainty remains part of the equation, we should expect the unexpected. The fact that Jubilee has employed Cambridge Analytica again extends the uncertainty to the role of data mining and political algorithms.
ETHNIC ARITHMETIC VERSUS ADVANCED ANALYTICS
Up to this point, Kenya’s political system has shared more in common with Stalin’s “who counts the votes” principle than the “those who can best influence choices” techno-nerd data-tweaker school.
The import of the CA contribution to the 2013 campaign is debatable. Everyone knew the youth demographic was a critical factor, and the major parties and savvy politicians were already using social media to expand the influence of their campaigns. The New York Times challenged the impact of their input; other analysts also contest the firm’s claims to be the critical influencer behind the Brexit and Trump victories. Others see their “secret sauce” as manipulative, unethical, and a threat to participatory democracy.
But we assume CA will bring some useful intelligence and expertise to the allocation and deployment of resources to the task. This does not mean CA psychographics and other methods of voter targeting will significantly alter the outcome of the August 8 polls.
Even though CA is no doubt accessing all the social media data available, their cache of Kenyan data is not necessarily representative of the 19 million voters, and probably does not reach the 5,000 North American data point threshold. It is doubtful the data in Kenya’s noisier political arena impart the same level of clarity used to fine-tune electoral messages in the US. What we do know is that they are smart, and may well understand this far better than those of us trying to look in from the inside.
If their genius is being invested in social media, the evidence has yet to show on the radar. The Jubilee Facebook page with its 237,000 followers is not setting the world on fire, nor is its 46,000-follower Twitter feed. Uhuru Kenyatta in contrast, has some 3 million followers on Facebook and some 2.2 million on Twitter, while William Ruto has just over 980,000 on Facebook but 1.26 million on Twitter. Raila Odinga, meanwhile, has about 870,000 followers on Facebook and another 1.13 million Twitter fans at this moment in time.
The Jubilee Facebook page with its 237,000 followers is not setting the world on fire, nor is its 46,000-follower Twitter feed. Uhuru Kenyatta in contrast, has some 3 million followers on Facebook and some 2.2 million on Twitter, while William Ruto has just over 980,000 on Facebook but 1.26 million on Twitter.
In the end, the only numbers that will matter are the ones tallied after the smoke and rhetoric clears in August. There are other firms with comparable expertise that could provide the same services to the opposition if CA really confers an advantage in the political marketplace — like Palantir Technologies (named after Saruman’s seeing stone in the Lord of the Rings), a company started by Peter Thiel (another big Trumperian) with a $2 million boost from the CIA’s venture capital arm, Q-Tel.
Kenyan elections are still a different kind of political animal that inhabits a weaker, if evolving, institutional habitat. At the onset of the campaign cycle, I proposed a disaggregated research and metric-based approach similar to CA’s three-pronged methods to a colleague running for a high office; he liked the idea but said, “The problem is 65 per cent of our people are irrational.”
Many of these voters are not so much irrational as prone to be incentivised by material inducements, fake news, and peer pressure than by appeals to personal values. But how many? Low-tech methods like helicopters, the gigantic billboard figures now straddling the country’s roadsides like the ancient kings of Gondor, and even confetti cannons, trump hi-tech wizardry all the time (no pun intended). Against this background, the contracting of Cambridge Analytica may even appear to be an extravagant if not foolish waste of funds, although we should also be cautious in light of Dator’s second law stating “Any useful idea about the future should appear to be ridiculous.”
Ethnic identity and other prime drivers of voting behaviour tend to reduce the number of swing voters on the national level. But this may be changing, and 15 per cent of them currently fall into the latter category according to polls — a relatively large figure compared with the thin margins separating presidential candidates in the last election. All of these reasons suggest we should not dismiss the power of persuasion built into data-based political algorithms.
We may see a surge of propaganda and emotionally provocative messaging during the last three weeks of the campaigns. But the more serious implications of Cambridge Analytica’s local presence transcend questions about their impact on Kenya’s national elections.
WHAT DO KENYANS REALLY WANT?
CA claims they are “politically agnostic” but they are actually quite the opposite. This and other claims by their CEO, Alexander Nix, demonstrate that they are slick dissemblers but not very good liars. Their website boasts of the campaigns they have assisted and prominently showcase the names of Ted Cruz, Ben Carson, and John Bolton alongside Trump. Robert Mercer, a key member of the ultra-conservative Koch brothers network and the single-largest Trump campaign donor, is their major investor.
The accumulation of mistrust and cynicism over the past two electoral cycles imbues the current iteration of electioneering with a large measure of angst and trepidation. If a clean and transparent outcome dissolves this negative legacy, everyone wins
Steve Bannon, who used to sit on the CA board, benefited from the enabling contribution of Mercer to set up Breitbart News, the reigning avatar of alt-right causes. Mercer and Bannon are at the forefront of the movement waging an asymmetric info war on the mainstream media and most other progressive causes they consider to be too liberal. Jane Mayer and Carole Cadwalladr and others have documented how these individuals and their colleagues are using money and power to reconfigure the online news ecosystem and leverage their toxic influence to undermine participatory politics and critical issues like climate change.
Nix also said the company participated in the 2016 US election to use the experience as a pivot for applying their decision-influencing brainware to the commercial sector and product branding. This rings true in contrast because the ultimate goal of Bob Mercer and the other elites populating the Koch Brothers network is to manipulate ideology and political culture to serve their material interests — both in Western political arenas and beyond.
It follows that CA does not need to deliver votes as much as create the perception they can produce results. The murky quality of Kenyan elections makes this a win-win proposition, while Kenya provides an ideal entry point into the larger region. Embedding themselves with ruling elites presents a pivot for exploiting emergent commercial opportunities. It will also provide the cover for rolling-out civilian psyops operations and misinformation in the service of private sector investors and foreign powers with an eye on the region’s resources and its growing numbers of persuadable youth.
In its May post on CA’s presence in Kenya, Africa Confidential reported, “The spectre of a highly secretive data mining company using such techniques is raising serious concern among Kenyan activists,” and added that this is occurring in a climate where the wounds of the 2007 post-electoral violence have not fully healed.
Most Kenyan voters want a fair and peaceful election that will validate the legitimacy of the winning party and candidates, even if they are not the ones of their preference. The accumulation of mistrust and cynicism over the past two electoral cycles imbues the current iteration of electioneering with a large measure of angst and trepidation. If a clean and transparent outcome dissolves this negative legacy, everyone wins.
Be warned, these people are not the Artur brothers; their machinations should be a concern for everyone, not just activists
Jubilee’s CA baggage is problematic from this perspective. If they win, the latter’s presence casts a local shadow, however small, on the legitimacy of their victory. Regardless of the outcome, it links them with the white supremacist Breitbart faction of the Trump political machine and like-minded operatives like Nigel Farage. This could backfire in any number of ways based on what we know about this network’s modus operandi and other potentially more perilous unknowns lurking in the future.
THE ROLE OF BIG DATA APPLICATIONS IN THE DEVELOPING WORLD
Be warned, these people are not the Artur brothers; their machinations should be a concern for everyone, not just activists.
These perhaps speculative considerations bring us to the larger question of the role of fast-moving technological and big data applications in the developing world, including the examples highlighted here. Advocates of persuasive technologies working out of Stanford University argue that applications used to influence consumer behaviour and voter choices can also be applied to promote peace and harmony. They even claim persuasive technologies can be deployed to establish an enduring state of world peace over the next several decades.
Kenya is not a technological backwater when it comes to innovations in this domain. Kenya launched the planet’s first mobile money service, and boasts other contributions to the IT domain like the globally deployed Ushahidi crisis mapping tool. One does not need a real palantir to see that the planet needs more Ory Okollohs and fewer Bob Mercers if we are going to transit beyond the polarising influences and social conflict engendered by the current phase of electoral cyber wars.
By PAUL GOLDSMITH
Dr Goldsmith is an American researcher based in Meru, Kenya. He has lived in the country for over 40 years