The UK has a parliamentary system with 650 seats, representing 650 distinct geographically bounded constituencies. Each seat elects an MP. An MP wins if they get more votes than anyone else from voters in their constituency. The party with the most MPs forms the government. That party leader becomes Prime Minister.
But you know all this.
In 2018, following pressure from organisations like us, Facebook started up an Ad Library, designed to provide near-real time transparency into what people were doing with political ads.
Unfortunately, right from the start, there were two major gaps.
One was targeting transparency (more on that elsewhere), the other, was detailed geographical information about where ads were being shown.
Facebook provides some, but it’s different in every country. Sometimes they break down national geography in a way that lets you see how just a few thousand people are being targeted. Other times, the geographical units are much larger, comprising tens of millions of people.
The people living and voting in those bigger geographical units get the worst deal from Facebook.
For example, Facebook offers geographical transparency for the four countries that make up the UK: England (which contains 533 parliamentary constituencies and 56 million people), Scotland (59 and 5 million), Wales (40 and 3 million) and Northern Ireland (18 for just under 2 million people).
So, if you want to know who’s running ads in a single constituency or ward in a general election, or get a picture of the totality of ad spending in the same, you’re out of luck.
In our estimation, England is the worst served place in the world when it comes to Facebook’s geographical transparency. In an electoral system that’s totally dependent on the outcomes of a large number of electoral units, Facebook hands the English a paint-roller and asks them to fill in a colouring book.
The trend extends to many highly populous areas. For example, Sao Paulo state in Brazil (45 million people), California in the US (39 million), NordRhein Westphalia in Germany (18 million) and Ontario in Canada (15 million) all also get a raw deal.
Why might that be?
Well, we can’t see a good reason for it.
Facebook knows pretty well where ads are being targeted and delivered. It shouldn’t be a problem for them to provide this information via transparency tools. (By comparison, Google provides quite granular geographical transparency for political ads, though the rest of their offering leaves much to be desired).
There’s no obvious privacy or population size issue (after all, Ireland gets a county level breakdown, meaning that County Leitrim, with a population of just 32,000, smaller than any UK constituency, gets to see exactly which ads are running there).
Nor is data about constituency and ward boundaries hard to get hold of.
Perhaps there’s an issue around advertiser confidentiality? After all, political advertisers might prefer Facebook to hide what they’re up to for them. So what. There’s no obvious democratic or public interest reason for a political campaign to run a secret strategy.
If more transparency is to lead to more accountability, it needs to be worth having. One way platforms (and regulation) can help with this, is by helping to show what’s really happening in elections, where it’s happening and to what extent. This would empower researchers and journalists to tell a fuller, richer story and, where they find it, make it easier to expose bad practices.
In the end, the only conclusion to draw is that Facebook has been lazy on increasing geographical transparency for the last four years.
It’s obvious what they need to do, and they should get on with it.