Facebook sucks for privacy, but how good are advertisers at targeting me?
Amid the massive hue and cry about Facebook’s utter disregard for data privacy, and the illicit use of personally identifiable information, I thought I’d take a step back and examine the company’s data handling from a new perspective. Facebook sells a bunch of targeting methods to advertisers, but are the companies successfully targeting my profile?
To find out, I decided to conduct an informal and, admittedly, subjective data analysis of the companies that are using my personal information to get me to buy their products on Facebook.
In order to do that, I accessed the list of companies who have uploaded my contact information to Facebook to communicate via targeted advertisements (custom audiences). Then I judged whether I’d actually be receptive to their ads (i.e., the accuracy of the advertiser) and highlight the way the contact information got collected.
So read on to see how the advertisers fared in getting my attention — and to get my take on a potential solution to the menace of data privacy in digital advertising space. You can also perform the same experiment on your own data if you’re curious!
Building a lovely data set
Thanks to Facebook, anyone can download a copy of their data directly from the site. I clicked on “Download Your Information” link present in the settings page of my account, and then on “Create file” to grab a ZIP file containing all of my data.
This ZIP file contains a file called ‘advertisers_who_uploaded_a_contact_list_with_your_ information.html’ with an ‘ads’ folder which has the list of companies that uploaded my contact information to Facebook’s ad platform. In the next step, I copied the name of all those companies to a spreadsheet and categorized them based on the industry. Finally, I added identifiers — ‘yes’, ‘no way’, and ‘not likely’ — depending on whether I’d be interested in what they are saying.
For example, Pizza Hut would be put under ‘Food’ category and marked as ‘yes’ (mmm, pizza). Given below is the screenshot of the first three rows of my final data set:
Getting not-so-lovely results
Once the dataset was ready, I fired up ‘RStudio’ to analyze and visualize. I created a diagram called alluvial diagram (a type of flow diagram named after alluvial fans) for visualization.
Hmm… looks like FMCG companies like Hindustan Unilever Limited (product examples: Dove, Taj Mahal Tea House, Pond’s), Pure Derm India, Citra India have absolutely wasted their money while acquiring my information.
Note: You can use the ‘R’ code present in my GitHub repository to create this visualization based on your data.
Coming back to the dataset, irrespective of whether I’ve willingly shared my contact details or not, the data shows I’d be receptive to 65 percent of the companies who have uploaded my information. That’s perhaps not so bad, but I’d expect more precise targeting from the advertisers.
Now, if I look at the companies that are absolutely new to me and there is no way I’d have shared my details, then such advertisers would constitute 40 percent. This is bad, since one would expect Facebook to protect the users from unknown advertisers.
According to Facebook, Custom Audiences has been developed to help businesses reach customers they already know. However, judging by the companies that are targeting me it doesn’t seem the social media giant is doing that.
As mentioned above, around 40 percent of the companies in the list were absolutely new to me and I’m confident there’s no way I’d have shared my details with those advertisers. There’s a bunch of random companies in there — such as a weird company in the agriculture industry — that have no business targeting me or using my data. Barring one or two blockchain projects, I didn’t mark ‘yes’ for any of the unknown companies; the futile attempt of these firms just makes me smize!
Clearly, Facebook hasn’t been able to enforce its terms, and the companies (or their data partners) also didn’t bother to get consent from the targets. But, the question remains: how did they manage to get my data?
And then there is greed
It’s rightly said that greed is the root of all evil and even more so, an advertisers greed to reach more and more people in this case. Essentially, companies want to reach out to as many “relevant” people as possible, even if it means contacting a lot of “irrelevant” people as well. There’s also no dearth of contact information if the sourcing doesn’t bother you.
There are many authentic ways of acquiring contact information — some of which include gifts, coupons, surveys, and loyalty program. However, for majority of the companies these are not enough and data brokers take the onus to satiate the advertising hunger.
A simple example can be a laundry service provider that has recently launched its business. They can start working with a data broker to figure out who can be their customer based on the analysis driven from the humongous amounts of Facebook user data. Once this part is solved, the next step would be to expand the perimeter to reach people with closely matching data profile.
Also, data brokers have already partnered with diverse set of companies (from large airlines to ecommerce companies) to buy customer data. In most of the cases, customers are in the dark about these partnerships and unknowingly provide the access to their data.
Targeting users by acquiring data from data brokers is a common practice on many ad-targeting platforms and Facebook is not an exception — which brings us to the primary issue, the issue of transparency. Many of the players in the advertising domain project ambiguity when it comes to their relationships and operations with data brokers. And this directly leads to data privacy infringements.
The potential fix for the solution ironically stems from Facebook’s own decision. Last year they shut down Partner Categories which allowed third-party data providers offer targeting directly on Facebook. The small statement given by Facebook had massive impact since that option was used by the data partners and brokers.
Hence, one major solution to the privacy issue can be termination of relationships with the data brokers by the internet companies’ ad platforms. Above all, being more transparent with respect to data sourcing and usage will further bolster digital privacy.
Finally, Facebook has pushed the onus on to the advertisers to get consent for targeting after falling short of preventing non-consensual targeting. But, advertisers always have huge incentive in terms of revenue generation to bypass the regulations. So I’ll still see ads from creepy companies that I’ve no interest in, as long as I’m left with the mediocre ‘Manage Your ad Preferences’ option. But at least I get some ads about pizza…