Tradeoff values & personal data
In a recent blog post, here, I reflected on a conversation with Thomas Arnold, an ethicist at Tufts University who helps design & engineering teams understand the ethical and long term implications of certain robot/computer-human interactions. We talked about the long term implications of automated vehicle technologies, specifically automated wheelchairs, a technology cohort which inherently uses a certain level of personal & rich data about its users to improve performance. In that post, I explored the broader ethical and market interactions of the proliferation of such a technology, which allowed me to view the situation from outside the system. When viewing systems as an outsider, its easier to spot an unbalanced value tradeoff between user and provider. As the user of a product, however, that value tradeoff is less easy to evaluate without personal bias, and therefore harder to understand if you’re comfortable with terms of the transaction.
"If you’re not paying, you’re the product."
Two services that have a big impact on my (and others) daily lives are Facebook and credit bureaus. Both services use increasingly complex and in depth processes for tracking personal data and turning it into a product that they sell. Both services are technically free, meaning that you & your data are the product. Additionally, both services can seriously impact ones social and societal standing. However, bureaus are a forced necessity the US, while Facebook is an opt-in service. Having worked at Facebook, and specifically with a focus on helping institutional banks drive their brands and adoption by utilizing Facebook data, I was exposed to the processes that both Facebook and the credit bureaus use to monetize their data. During that experience, I’ve decided that while I’m uncomfortable with the in depth use of my data for monetary gain, I’m more comfortable with the way Facebook uses my data, as the transaction & value exchange is rather transparent with a bit of searching.
Facebook
- Collection process: The majority of data collected on Facebook is through ones actions on the platform (likes, shares, searches, videos watched, interactions etc.) Such data is derived from all Facebook Inc. properties (IG, FB, WhatsApp, Oculus etc.) Additional rich data is derived from apps which share data with Facebook via apis, as well as predictive models built around historic use. Up until mid-2018, Facebook also partnered with data provides like Datalogix, Experian & Epsilon to create deeper profiles for their users (ie: income levels, purchase patterns etc.)
- Monetization: Facebook sells ad space, which is powered by the richness of your data. They do not sell user data, and never will. It is against their values (which I believe in) and would also be detrimental to their position in the market. In order to create a powerful and effective ad platform, they utilize user data in certain ways (not a comprehensive list):
- With ~2.5 Billion people on Facebook proper, and 3 other platforms with 1B+ users, Facebook can offer one of the most comprehensive branding platform the world has ever seen. Inventory is strong and the CPMs (cost per 1000 impressions) remains relatively low - around $15 max for most demographics.
- By instructing brands to place Facebook pixels on their site, Facebook can track user journeys and understand the impact of their media…which also allows them to refine targeting and deliver media to people likely to convert on ads.
- Facebook develops a "coefficient" for each pair of nodes (user-user, user-topic, user-brand, group-event, etc), which utilizes the data mentioned to create. They then use this data to estimate potential reactions of a person to certain types of content. A similar methodology is used with ads, where Facebook can deliver the most optimal ad to the most receptive person at the right time.
- Value tradeoff: Most users are aware that Facebook sells ad space in order to make money, and that the process involves using some level of their profile data in order to do that effectively. The value tradeoff for Facebook is use as follows: I’m given free and full access to my Facebook, Instagram, and WhatsApp networks, in exchange for the use of my data - in an aggregated process - to deliver value to brands.
Facebook has given users the ability to download the data they have on file, which provides one of the highest level of internet transparency seen in the industry. See the images below for examples on what this data looks like.
Companies who have uploaded targeting lists including my information.
Ad Target info based on Facebook platform data.
I’ll also concede that the issues around Facebook becoming a place where fake news is prolific, democracy is challenges and elections are swayed gives me grave and serious concern around the efficacy of Facebooks goal. However, I firmly believe that the world has been primed for a large ‘social network’ since the launch of Friendster, and the issues we’re seeing now would have likely taken place on any leading network if Facebook hadn’t come out on top. Our challenge now is not to figure out how to reduce the use of Facebook, but how to make it place that is safe and additive.
Credit Bureaus
- Collection process: A credit bureaus (Equifax, Experian, TransUnion) core data set reflects the credit score of each registered US citizen with a social security number. They use data from any bank or credit union whose product you use, any lender who you owe debt to, and any other fiduciary process in which you are liable for, which involves a loan or a line of credit. They’ll receive tons of bits of information from local credit clearinghouses and map those pieces of information to an individuals file to calculate a credit score.
- Monetization: Unlike Facebook, credit bureaus outright sell your data. Banks and issuers buy the rights to use your data to run directed media, and also pay to do credit checks for more premium products like ‘pre-approvals’ for large financial products (mortgages and auto-loans). Marketing agencies and large brands buy targetable lists of people with certain credit levels, income levels, volume of credit lines, etc.
- Value tradeoff: Every single one of these transactions is hidden from the end user, and no action has been taken to make this process more transparent. Additionally, my ability to edit, change or influence my score if something (like a data breech or identity theft) happens doesn’t exist. The value tradeoff here allows me access to financial products, but does not give me any level of transparency as to how my data is being used and how I can protect it.
Credit Bureaus have not given users the ability to download the data they have on file, and thus they (we) don’t even have the ability to start thinking about how influence or change how that data is being used.
As personal data continues to be the center of the conversation around privacy, large corporations, and ideally countries, should begin utilizing a transparent system to allow users to clearly understand what data is being stored, and how its being monetized or used. Such transparency typically results in a more trusted and trustworthy system, which can then lead to a more valuable product and better value-tradeoff for end users.
