Political microtargeting
Inching towards the Dutch elections, increasing amounts of claims are being made about the (micro)targeting of different campaigns. Here are some: New Social Contract does not do any online ads. The Greens-Labour Party spent €323,000 - more than twice as much as the Democrats 66 (D66), who spent €153,000. The Think (Denk) Party is profiling for Turkish ethnicity. The senior 50PLUS party microtargeted on people who listen to Angerfist.
These claims surrounding (micro)targeting are interesting because they can only be made in light of transparency registers for ads. There are various laws that make that possible - like the European Digital Services Act and more. As a result, these claims regarding microtargeting should be easy to verify: Find the relevant transparency register, find the relevant party within a given timeframe, which should give you all the campaigns they have run. This includes the amount of money spent, the amount of impressions, and what the targeting criteria were.
Yet, the news agencies I see reporting (micro)targeting claims hardly ever link out to the transparency registers to support their claims. As journalists, I would suspect that providing the relevant documentation or citation is important to credibly scaffold the article itself. However, this is yet another data point in the limitations of modern day reporting - there is zero linking to supporting documentation or methods utilized.
Journalism is a form of research, and research needs to be verifiable. As a result of this lack of claim support, I started looking into how to verify these claims myself. I hope journalists have this information themselves, but credibility should not only be available upon request.
First off, there is no central register as far as I could find. Each platform has their own transparency register. For your convenience, here's a reference list - if I missed any, please share it in the comments and I'm happy to update.
Even these links would have been helpful to start verifying the claims, despite that they do not link out to any specific information used in the reporting.
Digging into some of these registers more directly indicates that you can link to specific search results. It is possible, even if the information is not provided. For example, find all the VVD campaigns on Meta-based platforms by clicking here. It tells us that since 2019, they've spent €598,598 on ads[1] and €6,701 in the past seven days.[2] By exploring these ads, it becomes clear we can also permalink a specific ad in order to document a claim - here's a large VVD ad that reached between 500,000 and a million people. If all else fails, we can always take screenshots of the information.
By finding a specific page and clicking into the "Audience" tab, it is also easy to identify how they've targeted. I was able to verify the the Think/Denk Party microtargeted based on all kinds of Turkish interests and demographics. Other parties, such as BIJ1, also targeted based on interests. The Greens-Labour Party did the most detailed targeting - including things such as "climbing," specific banks, and their own collected profile IDs - on Meta alone they spent €193,210 in the past 90 days to campaign.
This small guide may be helpful to refer to in case you're looking to verify some of the claims regarding (micro)targeting.
All to say, ad targeting is happening and the reporting is incredibly important. Microtargeting is problematic for various reasons - especially as political parties are not keeping to the relevant laws for placing tracking cookies on their websites. Journalism is a tool to gather and synthesize information - it can also help people identify routes with which they can investigate things and provide the basis for more journalism.
Finally, I was especially tickled by the claim that 50Plus microtargeted to Angerfist - listen to the music below to see why that claim caught me off guard. I was unable to verify this.
Comments ()