Fortnightly Newsletter - Unions are (badly) losing the data arms-race

An arms race is under way that will deeply affect the future for working people, in Australia, and everywhere.

It’s not about who is building next-generation nuclear weapons or who controls South-China Sea.

Instead, it’s a struggle over who controls data, who has the best data, how unions use data to achieve industrial and political objectives, how corporations use data to control workers, consumers and democracy, and how political parties use and regulate data.

The data-wars are here, workers are losing, and unions are being destroyed.

Unions need to understand data, and become masters of data.

Let’s be absolutely clear: the absolute, historic, advantage unions had to win was our monopoly of the “going rate”. This information and data advantage meant that we knew what the wage-rates of workers were, and we used that data to organise workers and build power.

What’s more: we knew more about our members, about workers, their lives, their hopes and ambitions, their working conditions, and their industries. We knew more than the boss, we knew more than the government.

But the movement lost that advantage in the 1990s, and by the 2000s, we more or less forget we ever had it.

Meanwhile, around the same time, corporate control of data started to ramp up. The consumer database companies and the consumer credit ratings agencies combined with the banks, insurers and frequent flyer programs. Then, by 2010, Google and Facebook’s data dominance in the consumer space super-charged this.

Unions have always disregarded the power of data, but political campaigns and the corporate sector haven’t.

The microtargeting tactics (pioneered by the Bush campaign in 2004) took decades-old survey and consumer data techniques from the catalogue-sales and charity donations sector to identify individual voters up for grabs and determine what motivated them.

In 2012, the same time the union movement had put the nails in the coffin of its digital Your Rights at Work campaign infrastructure, the US Democrats were dominating the data battlefield through using massive consumer data-warehouses combined with Facebook targeting and the Blue State Digital system to target 15 million individual swing voters (you can read about my experiencing working on the Obama 2012 campaign here).

Facebook has removed the data-stalking features that Obama (and Cambridge Analytica) used, but by 2016, it was the Republicans who were dominating data.

How did they go from being utterly crushed to kicking the Democrat’s arses?

They created an information warehouse (called Data Trust) that has exclusive Republican party contracts to be a data exchange and provide data to Republican campaigns.

This model is widely used by large charities. They call them data cooperatives. The purpose of the data cooperative is to act as a large warehouse of donor data, which members (the charities) put data into. The bigger the warehouse, the more accurate and useful. The charities then contract the data cooperative to send fundraising appeals to individuals on their behalf. The people who respond (donate) then are added to the charity’s own list.

Data cooperatives are not new initiatives, and their corporate equivalents have been weaponised by surveillance capitalists and advertising sales conglomerates.

The Data Trust was a major reason why Donald Trump won in 2016. His campaign (and the Republicans) had a bigger database, that was more accurate and this resulted in better targeting of social-media messages, video sharing, direct mail and walk-and-phone programs for voter registration, persuasion and turnout.

(Contrast this to the Clinton campaign, who inherited “nothing of value” from the Democratic party with respect to data. This was because the Obama 2012 campaign refused to turn over its massive data-trove and digital tools to the Democratic National Committee after the election.)

Having a resource like the Data Trust has given the Republicans (and their wealthy corporate sponsors) a continued edge, and this not only means they can target better, but it also means they spend money more effectively. Karl Rove (infamous Republican campaign strategist) estimated that the Data Trust saved the Republicans over $100 million by allowing them to stop spending money on people who’d already voted.

The dominance of the corporate sector on the digital and data battlefield is also massive. While this is hardly surprising given that the Big Banks have resources of several tens of billions of dollars, but it is important that this data dominance didn’t happen by accident.

A decade ago, in response to the sudden appearance of “fintech” companies, the big banks invested hundreds of millions in data capability. They realised that becoming masters of data would enable them to keep their oligopoly, and that meant they had to dominate the data war.

(Obviously, their omniscience is not absolute, and their use of data is corrupted by insatiable greed, as the Banking Royal Commission and the recent Westpac scandal have shown).

On the consumer side, surveillance capitalism is not only all-pervasive and deeply intrusive, but is also not very effective.

Thankfully, while surveillance capitalism is a massive and growing threat to the rights of citizens and to democracy, most capitalists don’t know how to use this super-weapon very effectively, as The Correspondent recently highlighted.

On the worker side, workplace surveillance is taking even more extreme and sinister forms than consumer surveillance.

Employers are using a range of technologies to monitor their staff’s web-browsing patterns, keystrokes, social media posts and even private messaging apps. This, combined with “traditional” surveillance (e.g. blood and drug tests) are giving employers unparalleled and unacceptable power over workers.

This kind of technological surveillance is exactly what workers and unions fought against in the late 1800s and 1900s, up to the 80s, in factories and classrooms everywhere.

However, although employers are collecting vast amounts of personal data about their workers and their activities, to-date, they don’t really know how to use this, other than as the basis for sacking people. Only a few very large corporations (like BHP or Amazon) that invest billions in capital equipment and plant are using data about workers to replace them with machines.

And on the workers’ rights and industrial information side, an increasing number of corporations (especially in the legal tech sectors) are investing tens of millions of dollars to provide consumers with workplace law information.

For example, a privately-owned company is currently providing insurance companies with an automated wage-theft calculator. This calculator can accurately and automatically review tens of thousands of payslips in a few seconds. The same task would take an industrial officer months.

Unions as a movement are basically absent from this battlefield, and workers (both as workers and consumers) are losing the data war.

So, what should unions do?

Firstly, we need to completely re-think the kind of organisations we are.

Unions need to recapture the idea that we are information organisations, that we are data organisations.

Unions need to understand that our principle asset is data. Both data on members, but also data about workplaces, about wages, about industries, about safety incidents, about industrial compliance.

The data we need to value is not just about members, but about employers.

We need to overcome, as the charity sector has done, the unwillingness to share data between unions. Charities (and the Republicans) have achieved this by creating data-warehouses as separate, independent entities.

Unions need to invest in financial technology and data. The most important of which is the payment platforms unions use to process membership fees.

This is one area where unions are exceptionally vulnerable, and where we hand over vast amounts of member information to hostile corporate entities (banks).

Compare this to the one resource that the Democrat still have an edge over the Republicans: ActBlue.

ActBlue is the online fundraising platform that was created in 2004. It holds the credit card numbers of 8 million people. This gives the Democrats access to a vast and growing universe of small-dollar donors.

In response, the Republicans and their billionaire sponsors, have created a clone.

Unless the Democrats invest a lot more in ActBlue, they’ll soon be overtaken.

A few years ago, I worked with Luke Hilakari at Trades Hall to develop what eventually became Megaphone. This was in response to the business model of Change.org, a data-sales company masquerading as a petition website.

The establishment of Megaphone is good, because it ensures that the union movement stores its own member-data, rather than giving it away for free to Change.org. (By way of comparison, in 2012, Facebook stopped largely handing out access to its social-graph because it realised the enormous value of its social data; instead, it make companies spend a fortune to access it. Compare this to unions who give data away for free to companies like Change.org and other non-union petition platforms.)

But we need to invest a lot more into both tools like Megaphone, and a union version of ActBlue, that allows the movement to merge supporter data with financial payments.

The second very important thing we need to do as a movement is become more digital and data-literate.

At almost every level of our movement from organisers and admin staff to senior leaders, there are only a handful of people with a strong grasp of data and digital tools.

Unions need specialist data staff, and senior data staff. We need our own coders and developers. (and I’m pleased to see that the ACTU has started down this path, an excellent decision.)

We need to invest in technology for the movement that builds on and amplifies our knowledge of both industrial law and employers.

Lawfirms invest vast sums of money into legal databases that give their lawyers an edge. These databases are basically repositories of cases and precedents, and integrate into Word and other tools used by lawyers to draft letters and legal opinions.

Tens of millions of dollars are invested in new legal-tech, including automated tools that automatically generate letters of demand, appeal letters and the like.

Similarly, we’ve lost our data monopoly about whether or not an employer is good or dodgy – the likes of Seek, Glassdoor are now the go-to sources of information, whereas a few decades ago only unions had this economy-wide knowledge. The effort of Fairplate is admirable, but it’s like bringing a knife to a gun fight.

Over in the US, there’s tens of millions starting to flow into political data. The Democrats have a new data “dark money” group called Acronym, aimed at rebuilding the DNC’s digital armaments to challenge the Republicans.

Unions already spend a small fortune on IT and technology, but almost none of this is accessible for the movement, and as I noted, only a handful of union leaders have good data and digital literacy. The consequence is that a handful of tech and CRM vendors have scammed unions into expensive, proprietary and outdated tech solutions.

A few weeks ago, I attended the launch of the Centre for Responsible Technology. The discussion largely centred on the risks and unregulated nature of surveillance capitalism.

There’s no doubt that unregulated oligopoly capitalism is an enormous risk to civil society, demoracy and freedoms.

Unions should definitely be involved in the discussion about the ethical use of data, and the regulation of data companies like Google, Facebook and the various second-party data companies. This is especially worrying as the “internet of things” and smart assistants start listening to conversations in our homes, and as insurance companies weaponise our activities and biometric data from wearables (e.g. the Apple Watch) to even further gouge oligopoly profits.

When journalists and business executives call data the “new oil”, what they’re really saying is that the personal and private information about individuals is being privatised and monetised. What’s more, companies are creating data on our behalf, which the company, not us, own, process and sell without our knowledge or consent. This data becomes the property of the company.

Ultimately, these large companies and the machine learning algorithms they create need to be regulated; the large platforms like Facebook and Google, if they’re not broken up, need to become public assets that operate in the public interest.

But unions absolutely cannot remain in the losing position we are in now. We’re being outgunned, and the consequences for workers is devastating.