A dark advert is distinguished from a normal advert by the fact that it is never seen by anyone except the intended recipient.
Imagine during an election campaign, a particular party advertises the message “We will ban all migration” on a billboard in one part of the country, and that same party puts the message “We will allow unrestricted migration” on a billboard in another part of the country in an attempt to attract voters from both sides of a debate. Now imagine someone from one end of the country is visiting friends in the other, they notice the mismatch and naturally the party is ridiculed by the public.
Now imagine instead of billboards, those were online ‘dark’ adverts. How would you go about determining if your neighbours or those of different demographics were seeing different adverts to you? How can you be sure you truly understand what a party stands for when the vast majority of campaign material is hidden from view.
The true solution is for Facebook and other advertising platforms to publish data on how campaigns use advertising. Unfortunately this is unlikely to happen anytime soon, and in its absence Who Targets Me is crowdsourcing a repository of political advertising.
As long as there have been elections, campaigns have segmented their audiences in an attempt to tailor popular messages to the right group.
Microtargeting refers to the extreme form of this segmentation, often driven by mining audience data and combining multiple datasets for the use of predictive analysis. This data is used to drive direct mail, phone calls, home visits, television, radio, web and social media advertising, dark advertising, email, text messaging, etc… with content tailored to resonate with that particular group’s unique information or demographic.
Asking our volunteers to anonymously provide Who Targets Me with their basic demographics helps us understand the groups campaigns use to target.
As computing resources become cheaper, datasets previously unusable because of their size become feasible as potential sources for datamining for the use of microtargeting.
Big Data refers to the ability to extract meaning from complex datasets, generally for use in predictive analysis and within politics for informing election campaigns.
The boundaries of which data can be meaningfully processed and used in elections are constantly being pushed. The advent of artificial intelligence and natural language processing among other technologies, combined with the sheer volume of personal data available on social media is inspiring a host of creative analytics firms to look at whether advertising can be reliably tailored to a person’s psychometric profile. There’s no limit to the personal attributes we could see advertising tailored for in the future.