Thread by @WolfieChristl

Facebook plans to require businesses that upload consumer data to the platform pledge “I certify that I have permission to use this data” #CustomAudiences

Facebook's terms already require that businesses have "consent from the data subjects". Looks like a cosmetic change by now.

External Tweet loading...
If nothing shows, it may have been deleted
by @JoshConstine view original on Twitter

Background: with Facebook's "Custom Audiences" tool businesses can upload lists of consumers to the platform, consisting of:

- email address
- phone number
- name
- zip
- city
- country
- birthdate
- gender
- Android or Apple ID
- FB app or page user ID

facebook.com/business/help/…

Facebook has an example file. You can upload even multiple email addresses of a person if you're not sure which one this person uses for her Facebook account:

You cannot just upload lists of consumers to Facebook, but also data about their behaviors, e.g. whether they've seen certain content, searched for a product, added it to a shopping cart, registered, purchased something or any other custom 'event'.

(facebook.com/business/help/…)

You can also add information about how much each person is worth as a customer.

But when rating or ranking consumers you shouldn't use your own rating or ranking system, just upload how much they are worth in $.

But "don't include negative numbers to denote negative value or undesireable customers".

Uploading static lists of consumers is only one way to provide consumer data to Facebook.

With 'Custom Audiences' from websites (facebook.com/business/help/…) and mobile apps (facebook.com/business/help/…) businesses can send dynamic/real-time data on websites visitors & app users to FB.

Facebook's "Custom Audiences" functionalities let companies systematically link data on consumers to FB user data, essentially providing a kind of real-time remote control for Facebook's data universe.

External Tweet loading...
If nothing shows, it may have been deleted
by @WolfieChristl view original on Twitter

Of course, there are more powerful ways to upload consumer data to Facebook than just using a boring Excel/CSV file. You can use any kind of software and Facebook's APIs to automate creating and updating custom audience records: developers.facebook.com/docs/marketing…

Worse, this is not only about Facebook. Google launched similar features in 2015 and extended its functionality end of 2017, they call it 'Customer Match':
adexchanger.com/mobile/google-…
searchengineland.com/google-adwords…
support.google.com/adwords/answer…

Today, most data companies pass different kinds of pseudonymous identifiers (mostly generated from email addresses, phone numbers or device IDs) to each other, resulting in pervasive real-time surveillance.

I guess most people would not agree if they'd understand how this works.

When data and adtech companies use the term 'audiences' they basically talk about lists of IDs referring to people with certain characteristics and behaviors, very similar to how companies have aggregated/traded/used mailing lists with names and addresses for ages.

But today's "audiences" are not simply static lists. They are dynamic real-time feeds about groups of people with certain behaviors, continuously updated with data across contexts, and used to make automated decisions on these people according to complex sets of predefined rules.

Today's platform giants are perhaps the most relevant environments to utilize these real-time behavioral feeds. Facebook introduced basic customer data matching back in in 2012, Google and Twitter in 2015, Amazon in 2017. But *full* adoption by businesses is still in its infancy.

I don't think it is a law of nature that wide-scale personal data matching/linkage at individual level across companies and services must exist in this form.