Quantifiable impact on polarization, populism & disinformation

Existential challenges for democracy

EU1.3m

Projects stemming from this solution won funding from the Danish MoFA

42%

Up to 42% increase in mainstream traction (ie, outside the echo chamber)

103%

Mainstream follower growth (98k - 199k) for a pro-democracy media NGO

There’s a lot of talk and a lot of worry about polarization, disinformation and climate skepticism, but still little consensus on effective solutions.

Over the last four years I've led 20+ public and third sector media innovation projects to develop and test a quantifiable solution.

I like the Einstein quote, “If I had an hour to solve a problem, I'd spend 55 minutes thinking about the problem and five minutes thinking about solutions.”

So let’s dig into the problem.

First, I’ll outline the logic of algorithmic distribution and the mechanisms underlying democratic decline.

The solution stems from there.

Examples of disinformation

Disinformation is just digital marketing

In almost exactly the same way that Nike and Adidas compete for brand awareness and affinity in the minds of their shared target audience, hostile states compete with democracies for brand awareness and affinity for their narratives. The tools and techniques they use are almost exactly the same as the tools private sector companies use to persuade us to buy stuff we don’t need.

To put it another way, advertising is storytelling and so is disinformation.

The overall objective of disinformation is always to fragment affinity with the shared narratives that bind a healthy society together.

If we agree on that problem formulation, then our objective in solving it becomes clear. Impact on disinfo and polarization comes down to outperforming hostile or harmful narratives in terms of awareness and affinity; persuading more people to believe our stories than theirs, in other words.

That comes down to traction, and traction is easy to quantify.

Susceptible audiences

Both digital marketing strategists and the strategists behind disinformation aim for the audiences that are most susceptible to the stories they tell about their products (the echo chamber, in other words).

For Nike and Adidas, of course it’s just good business sense to target their product at susceptible echo chambers of sports and fashion enthusiasts. Targeting anyone else would be a waste of money.

Disinformation strategists use the same logic, and target the audience segments that are most susceptible to their stories; those people who feel disaffected or disenfranchised, or who bear a grievance against “the system.”

The emergence of algorithmic distribution (Meta, Google, Twitter et al) in the mid 2000s gave both sets of strategists the perfect mechanism to segment and target their most susceptible audiences.

How algorithmic distribution changed everything

Polarization and the echo chamber effect

The problem underlying polarization, populism and disinformation is algorithmic distribution. Let’s be clear, algorithmic distribution is here to stay – it brings enormous benefits for society (as well as enormous challenges) and it drives the economy. So how is it that algorithmic distribution has such profound effect on politics, economics and culture?

Algorithms are designed to connect people with information they want to access. That drives the economy because, as the algorithms learn in increasingly granular detail what individual users want, they can connect them increasingly effectively with the brands they’re most likely to buy from.

It works very well for selling products, but when it distributes political ideas, it concentrates all the energy at the extremes. Consumer decisions are driven as much by emotion as by logic (this is the fundamental insight that underpins advertising). Algorithms are tuned to serve people content that engages them - that keeps them online longer - and engagement is driven by emotion too.

In the short term, algorithmic distribution seems to work for the public and cause driven sectors too. It gets them engagement and donations from energized echo chambers,

But as we’ve seen over the last 15 years, their macro social impact has shrunk, and society has become polarized and vulnerable to disinformation as the vast mainstream middle segment has lost touch with fact-based information and pro-democratic content. (If you’re not motivated to seek that kind of content out, it won’t get served to you by the algorithm, and there’s more than enough content on whatever it is you are interested in to keep you online and engaged 24/7.)

Using the targeting models prescribed by the platforms to distribute social and political ideas, inadvertently contributes to polarization. (That includes almost all organic reach too)

The complication: User motivation

Most people just aren’t motivated to engage with complex political issues in any depth. NGO and public sector content tends to be dry and intellectual. It takes mental energy to understand. People arrive on social media motivated for engagement and entertainment; they use social platforms to relax in the breaks between work. They don’t arrive motivated for cognitive challenges. Political content tends to be ineffective, in other words, because it misunderstands user motivation.

Disinfo and populism, on the other hand, motivate engagement because they’re entertaining, solution oriented and packed with emotion. They connect with disenfranchised segments’ pain, and they appear to offer real value. It’s effective media, in other words.

The challenge (and the solution) therefore, is to engage the mainstream (the segments who are increasingly vulnerable to populism and disinformation) with ideas that they need to engage with to enable democratic participation, but that they just aren’t really interested in engaging with.

Indeed, the mainstream are the prime targets of the populists and the strategists spreading disinformation. That’s why their influence is growing – they actively strategize to expand and convert segments in the middle – while democratic influence continues to shrink as NGOs and the public sector continue to engage only their echo chambers.

Example: Vote share in the September 2024 German local election

The solution: Expanding reach & engagement

Rethinking targeting is a big part of the solution for rebuilding mainstream consensus on sustainability, rights, facts and democracy. But it doesn’t work to just take messages designed for an engaged echo chamber and target them at everyone. It would be hugely expensive, get terrible engagement, & only succeed in harming an organization’s algorithmic ranking.

The solution therefore is a mix of targeting innovation and content strategy:

Creating messages designed to work for as much of the mainstream middle as possible (audience size is determined by the issue)
Taking the targeting out of the platforms’ hands, and instead targeting for social impact
Testing to learn what drives engagement (and what doesn’t)
Integrating data analysis into strategy, and continually iterating to optimize impact / engagement

Solutions are only sustainable when they mesh with economic realities. The aim of my experimental projects has therefore been to accumulate data that proves it’s possible to reach wider (reach the key middle segment) and improve return-on-ad-spend for engagement & donations.

Methodology: Multivariate AB testing & strategy iteration

Messaging

To expand traction with democratic ideas, first, I map the maximum optimal audience for the specific campaign issue; the segments who could support an organisation’s ideas, but who aren’t algorithmically defined by that specific issue, and therefore just don’t get served that kind of content.

And to package the issue in a way that the maximum optimal audience wants to engage with organically, I use whatever data I can get my hands on to understand them – their culture, language, values, standpoints, ideas, concerns, pains and problems.

Then I synthesize all of that into messaging that aims to resonate with them, and connect the issue with their lives and their concerns right now.

Testing

Until it’s tested, a new strategy is just a best guess. Our own judgement is (by definition) subjective. Our objective, on the other hand, is to make sure our messages land as well as possible with as many people as possible from this key mainstream segment.

To get objective data on which messages audiences respond to best, I run randomized experiments (AB tests) in which I serve new messages and old messages to both the existing echo chamber audience and the new maximum optimal audience.

Through this I get concrete, measurable data that shows what is effective and what’s not.

Real World Applications

We’re going backwards on so many of the values that make our societies cohesive, and our children’s futures secure. New media technology is a big part of the reason why.

Algorithmic distribution put an end to the media touchpoints that used that bind us together around shared narratives and truths, and anti-democratic actors have exacerbated and weaponized the polarization that resulted.

If we’re serious about turning things around, we need to focus more on winning broader support than getting likes from people that like us already. That’s a complex process, but it comes down to understanding what that majority is actually interested in, and concerned by, and connecting with them over that (instead of insisting on talking about what we’re interested in because we know better and everyone should listen to us).

A hypothetical example: Boosting affinity with the EU project

The EU is a net win for anyone lucky enough to live in it, but somehow pro-EU political parties struggle to connect the EU with voters’ concerns. Brexit is a case in point.

The EU is complex proposition. Most people just don’t have the time or the motivation to dig and understand it any depth.

But what people are interested in is themselves and their families’ wellbeing. If we use that simple insight to create media that connects the EU tangibly with their lives – showing them how the EU makes them healthier, wealthier and more secure, in other words – we will have immeasurably more success than the UK’s Remain Campaign did.