Facebook Will Cut Off Access to Third Party Data for Ad Targeting
techcrunch.com
techcrunch.com
>"Facebook clarified that it will still work with companies like Experian and Acxiom in order to measure ad performance and provide metrics, although it will also be conducting a review of those ongoing relationships."
These FB announcements the last couple of days seem little more than grooming for possible upcoming appearances in front of legislators. They can point to this and the surface-level app redesign and say they've "already" taken efforts to shore up concerns.
I think most people have no idea that FB and Experian - a consumer credit reporting agency have been sharing data at all. Where was that explained in the app redesign that was trotted out yesterday?
It's worth reading how Acxion who facing similar pressures recently handled their own superficial offering of "transparency":
https://www.nytimes.com/2013/09/05/technology/acxiom-lets-co...
If its just helping them target ads then they should explain why Experian is allowed to continue, but I am more concerned with Experian being "creative" with Facebook's tools and incorporating some of that into their credit scoring models.
Pulling data from more sources doesn't change the underlying problem of faceless businesses assigning people scores based on vague unknown requirements.
People will just create fake accounts that "like" things like "saving money" and "paying off debt."
Sure, I think most people who use FB would. Zuckerberg has been talking a big game lately about regaining user's trust and how "we have a lot of work to do." If he were at all serious about those sentiments FB would begin providing meaningful disclosure about what data they were collecting and sharing with which third parties.
They would also update their privacy settings and give user's the ability grant or deny access to the sharing of that data with third parties. Similar to how users now have to grant apps permissions to access contacts, the camera etc.
FB's desire to continue operating in total secrecy and simultaneously court public trust seems to be the very height of cognitive dissonance.
> FB's desire to continue operating in total secrecy and simultaneously court public trust seems to be the very height of cognitive dissonance.
Is it cognitive dissonance if they're self-consciously lying?
My take is nothing's fundamentally changed for them, and "the work they have to do" is distracting people until the spotlight's off, making only marginal changes in the process.
Take this change, for instance, it may have nothing to do with user privacy and everything to do with pushing advertisers to FB's own targeting platform more heavily.
I don't think it would be no, at least not clinically. And that actually makes the rhetorics even more insidious.
From [1]:
"Partner categories are, importantly, only available in specific areas. Acxiom provides data from Australia, France, Germany, the US, the UK, and Japan. CCC Marketing provides data from Japan. Epsilon provides data from the US. Experian provides data from Australia, Brazil, the UK, and the US. Oracle provides data from the UK and the US, and Quantium provides data from Australia."
[1] https://boostlikes.com/blog/2017/08/target-facebook-users-in...
I don't have a specific use case because its not my industry, but Facebook's approach seems to be to give people tools they know can be misused in creative ways but then frown sternly if they get misused, so it wouldn't be out of character.
Facebook knows what pretty much everyone in the Valley knows; 90%+ of the politicians in this country (and likely most other countries) don't have the technical education to understand just how little Facebook is actually doing to fix the problem.
If this is Facebook's business model, Facebook must go.
No one would allow the kind of pervasive intrusion put in practice by Facebook (and others) by the part of governing bodies, the police or the military. So why are we still allowing marketing companies to do it? Why are willing to let credit, insurance and health companies access our data in order to know us better than ourselves? When is this the price of a "better user experience"? It is disingenuous to say that people simply "agreed to this". No one can expect the average user to fully comprehend the inner workings of the modern web, and to what extent their data is being used, or even to keep up with changing terms of use for every single service.
Some, like Tim Cook, are starting to address the issue but not as directly as one would expect.
None of that needs Facebook.
Still want to play this game?
Credit card companies were extending people credit and doctors were making accurate diagnoses long before the internet came along, they are definitely not a necessity for the functioning of society in ways that really matter.
If you wish to argue that there are direct benefits to consumers from either giving credit card companies or health care providers access to your social media data then the the onus is on you to prove this.
That means no credit, insurance, or health related data collection of any kind at all.
What do you think of that?
The world is a lot larger than just the United States and in other places banks, insurance companies and healthcare providers function just fine even if they do not have the data gathering capabilities that are considered 'normal' in the United States.
So there's what I think of 'that', data collection is not a pre-requisite for any of this.
Partitioning the data and more heavily regulating the credit industry may be part of a solution.
Where do healthcare providers function well with less data than those in the US? I'm not experienced enough with the healthcare industry and nationalization of healthcare data to know, so if you have examples I'd really appreciate it.
Your whole argument is based on a misdirected premise anyway. In fact credit reporting and health care are very heavily regulated industries in terms of how that data is stored and what can be done with it. Facebook saw no such regulation, and look where we are.
The only health companies that diagnose are providers. They have been explicitely REJECTING patient-generated data for the longest time because of concerns about poor data. Are your blood-pressure cuff, scale, thermometers, etc. calibrated and regularly tested? Did you take the measures properly? Etc. Now some of that caution is excessive and we're starting to see bridges where patient-generated data is being considered.
But by and large, providers don't want crappy data. Even claims data sent to insurance companies are known to be fudged/imprecise/incorrect and aren't relied on for clinical decision. Even for population health that data is bad (some PH vendors still tried to rely on it but that turned out to be a bad idea).
So Facebook data has zero value to most providers.
Patient-generated data, both objective and subjective, is frequently relied on by providers for diagnosis (and always has been); where available, provider-collected data is preferred and more heavily weighted in the event of direct contradiction, but providers simply have no practical choice but to rely on patient-generated data.
> Even claims data sent to insurance companies are known to be fudged/imprecise/incorrect and aren't relied on for clinical decision.
Well, yeah, since the provider making the clinical decision provides that data after the fact, of course they aren't going to rely on it. Even if it was perfectly accurate, they would use their own source data from which data would later be extracted selectively to support claims.
> Well, yeah, since the provider making the clinical decision provides that data after the fact, of course they aren't going to rely on it.
That hasn't stopped some population health vendors to try to sell products that use claims data to build PH assessment tools. And yes that works about as well as you'd expect. :)
More data allows price discrimination so that some people pay more than others, which we tend to think of as good when its based on things like age in life insurance, but which we tend to think of as bad when its based on things like race in life insurance. These companies have repeatedly shown they are unable to self-limit themselves to only considering things we think of as good, and price discrimination is almost never beneficial for consumers in practice, so its worth having a conversation about whether they should be able to consider data like that at all.
What it means is insurance company or credit company does not explicit state how do they use my data to evaluate my price on certain insurance or credit to mortgage. If I applied for mortgage. By some random parameters in the model, I may be deemed not credible to have a mortgage. And I don't even have real profile data except my name, photos, and posts.
On the other side, if some insider knows what parameter contribute success mortgage, they can game the model by adding fake information to the facebook profile, or even sell this information for profit.
Facebook by no mean represent real identity, although it usually does. I don't think it is fair to use facebook information without consent from customer.
I am interested that if facebook states that using facebook will affect your credit history, how many people will still willing to use it?
The whole of Europe is addressing this with various data protection laws, the latest being GDPR.
And at the same time, Europe is rolling out eCall, which will make it mandatory for all new vehicles starting next April to have a dedicated system for emergency calls that logs GPS data and that has access to speaker and mic.
Car owners can, in alternative to the basic factory install, choose proprietary systems that may or may not share data with 3rd parties.
Addressing public concerns about privacy, the EU told us not to be concerned, because in EU's opinion this is completely safe.
So imho the EU's position on this is schizophrenic at best.
snooped my profile ?
harvested my profile ?
snooped my high res pictures ?
snooped my friends list ?
snooped comments on my pictures ?
snooped comments on my profile ?
snooped my time line ?
snooped on posted/reposted videos ?
snooped on my about page ?
snooped on my sports page ?
snooped on my Videos ?
snooped on my Check-Ins ?
snooped on my Sports ?
snooped on my Music ?
snooped on my Movies ?
snooped on my TV Shows ?
snooped on my Books ?
snooped on my Apps and Games ?
snooped on my Likes ?
snooped on my Events ?
snooped on my Questions ?
snooped on my Reviews ?
snooped on my Groups ?
snooped on my Notes ?
snooped on my friend and clicked his/her name and left my profile to start snooping that person profile ?
snooped my friend (clicking thru my profile) ?
snooped on "X shared" button ?
snooped on "view X more comments" button ?
Maybe this data should be available to download. Users could have the right to know why visited their profiles.
The visitor has his privacy protected while the profile visited does not.
People who care about the extensive personal data harvesting are free to stop using Facebook. You say "No one can expect the average user to fully comprehend the inner workings of the modern web" but there is no shortage of people summarizing the salient details for them, and those who hear the summary and choose not to use Facebook are free to do so.
I don't like Facebook and use it rarely. I'm glad they exist so I can check up on my friends.
What you are proposing amounts to increased government control over what kinds of services can be offered in the marketplace. I think you should be more afraid of government control of private interactions. In the US and Europe it's already far too intrusive.
I say that like I would say "shady credit pushers must go", although I also think people shouldn't resort to such services. Facebook provides a service free of charge but the contract isn't clear: there are consequences that must remain obscure in order to the whole thing work.
I believe it's pretty clear that Facebook depends on people inadvertently permitting systemic surveillance.
> Facebook is a private company that people freely choose to interact with.
Like many other companies that are, unlike Facebook, heavily regulated in their primary activity.
>People who care about the extensive personal data harvesting are free to stop using Facebook.
But Facebook has been collecting data of people outside of Facebook as well. In any case, people have elected representatives to protect them from harm in the meanders of technology concepts they cannot grasp.
> there is no shortage of people summarizing the salient details for them
That is hardly a substitute for regulation. Take the example of the agrobusiness: no one expects lay people to ponder the cost-benefit of many drugs and antibiotics present in food they "freely" choose to eat.
> I think you should be more afraid of government control of private interactions.
I'm not sure about that. Currently I am equally afraid of both.
>In the US and Europe it's already far too intrusive.
I agree. Unfortunately, Europe is becoming nothing more but a meta-corporation. Lobbying being legal in Brussels, common people have no way of matching corporate influence.
Also, I don’t really see how this does much to protect anyone’s privacy, this mostly just makes running effective campaigns more complicated and expensive. The data is still being collected and sold, now you can’t just get it directly through Facebook anymore.
I think by shutting down the marketplace for 3rd party data, Facebook is allowing it to continue elsewhere but also washing their hands a bit, so next time they will have better optics when explaining themselves to regulators if data misuse hits the headlines again.
It will make some marketers feel less in control...but i actually doubt that data actually warranted the additional costs facebook charged for using it.
Excellent.
> Also, I don’t really see how this does much to protect anyone’s privacy, this mostly just makes running effective campaigns more complicated and expensive.
If you don't have access to data you shouldn't have access to that improves privacy. Exchanges leaking data which then gets cached is a huge privacy issue, allowing third parties to link up this data with data they already have is absolutely terrible. Data files in isolation are bad enough, allowing the joining of disjoint datasets is about as bad as it gets when it comes do de-anonymizing people.
> The data is still being collected and sold, now you can’t just get it directly through Facebook anymore.
Good. Not all data will still be available, and hopefully some more of these holes will be closed soon to make it even harder to make running highly targeted campaigns more difficult.
From where I'm sitting the sooner marketeers lose the capability to run targeted campaigns the better, and every little bit helps.
On the plus side: your budgets will go up to reach the same effect so why complain?
I've worked on every single vertical and 3rd party data / Partner Categories is rarely used by big or small advertisers a like.
The cost of the advertising increases i.e. higher CPMs for using this data. I've spent north of $40ml across all digital ads with a large chunk being on $FB and it isn't the main kind of targeting that is used and its the most expensive kind as well.
I am not hugely convinced this will dramatically affect $FB given there are so many other targeting options out there it isn't much of a concern in my eyes just a shame this is one avenue that will now be closed but its one of so many marketers can use.
Poor thing.
I hope you weren't expecting any sympathy. A camera in our bathroom would also make it easier for you to suggest us a "better" shampoo, but no thanks.
>>this mostly just makes running effective campaigns more complicated and expensive.
Again, that's your problem.
>>The data is still being collected and sold, now you can’t just get it directly through Facebook anymore.
Until the outrage hits a tipping point.
Are you suggesting that I shouldn't be allowed to give companies this data if I choose? Because that's what it sounds like when I read this but I'm not sure, and I'd like to be clear about what you are saying and not make any assumption.
Am I risking jail time or just a fine? :)
You should read somewhere else regarding permissions, default settings for 99%of users and "I agree" when it's a leave or take.
Like surface-level changes that look good in a congressional hearing but actually just raise prices and increase ad-spending to maintain the same result.
> Contact targeting is a LinkedIn Marketing Solutions feature that allows advertisers to upload lists of contacts to include as part of their target audience for ad campaigns. If you have already interacted with a company and provided them with your contact information (e.g. to sign-up for a newsletter or webinar), they may include you in a target audience for a LinkedIn ad campaign using contact targeting.
It's part of "Matched Audiences": https://www.linkedin.com/help/linkedin/answer/81195
Twitter's rough equivalent is called "Tailored Audiences," though as far as I know, it relies only on hashed email addresses and device IDs uploaded directly by the advertiser, not data appending companies: https://developer.twitter.com/en/docs/ads/audiences/overview
TechCrunch doesn't say exactly what part of Facebook's product will be rolled back - uploading any hashed user lists? hashed user targeting only when done by third-party data appenders? only when based on offline events? - so it's hard to tell how far Facebook is going.
And yes, there's a specific Facebook targeting category for "bought beauty last 30 days", or was before this change.
If you're correct, the part they're rolling back is relatively small.
For the amount of data Facebook collects, the quality of advertising they provide is pretty much garbage. Late night tv infomercial quality.
I'd hazard a guess that not many ordinary users are even aware that retailers can upload their email addresses into Facebook and create targeting segments from them.
As to what these changes do to the AdTech industry is anyone's guess. 'Onboarding' email addresses is basically LiveRamp's business model.
This was a smart biz decision and CA was just a convenient opportunity to roll out without too much backlash IMHO.
I would not be surprised if this thing is undone later.
This is the information from Facebook (via Dennis Yu from Blitz Metrics):
// from Facebook:
Although I have spoken with many of you just today and yesterday, I wanted to make you immediately aware of an update we received to third-party targeting, late this afternoon.
We have a responsibility for the use of data on our platform - and we want to ensure that people have transparency and control over how their information is used. Over the past week, we announced important changes to reduce the amount of data that apps can request from users and ensure that people have more control over their information on Facebook. Now, to build on these efforts, we are going to be more restrictive in the way that we use data for advertising on our platform, particularly as it relates to information from third-party data providers.
Specifically, over the next six months, we will remove the ability to use Partner Categories, a targeting solution that enables third-party data providers to offer their targeting directly on Facebook. While leveraging third-party data is a common industry practice and we've put good protections in place, we believe this step will help improve people's privacy on Facebook.
We understand this may impact your advertising efforts on our platform, and we will work with you through this transition. In an attempt to minimize disruption, we will allow time for you to update your targeting. In light of the General Data Protection Regulation (GDPR) in the European Union, we have created a timeline to comply with the regulation:
· May 10: After this date, you will no longer be able to create or edit campaign using Partner Categories built on audiences from the UK, Germany, and France; however, they will be allowed to continue running until May 24.
· May 25: We will no longer deliver to Partner Categories built on audiences from the UK, Germany, and France, and these targeting options will no longer be available for use on our platform. You will notified to update any targeting containing impacted Partner Categories before this date.
· June 30: Last day for creating new or editing existing campaigns using non-EU Partner Categories; they will be allowed to run until September 30.
· October 1: All other Partner Categories will no longer be available as targeting options on our platform and we will stop delivering against these audiences. You will be notified to update your targeting by this date.
Protecting people’s information is the most important thing we do. You can expect to hear more from us in the coming weeks as we continue to work to make our platform safer.
All of those things are being asked by adtech lawyers right now in planning for the future EU data regimes.
Great tactical game by Facebook.
I realize we're likely too far gone now, but it's fundamentally an issue that we aren't paying for these ad-supported services, so the companies gravitate toward serving the interests of the entities that are paying them.
That's not a justification or reflection of my opinion about privacy and ethics around the use of data, just a thought around removing the problem from the equation all together.
I guess my real question is, if you remove the profit seeking component of the data discussion does the bad behavior completely go away? Definitely interested in other opinions.
[0] https://en.wikiquote.org/wiki/Mark_Zuckerberg
And of course you can down vote, but we will meet again in the next Facebook scandal :)
I left Apple (previously only a Mac user) several years ago over this. It's not merely the issue of ads (which is not to be understated, ads are ruinous, at the heart of tons of problems). The way power structures work in these non-FLO walled-gardens amount to an effective sabotage of the entire potential for FLO software overall.
Quote from the article: "In order to leverage the deep pool of data Facebook collects on users, the company mixes information that it obtains from users themselves (Pages a user liked, for instance) with information from advertisers (membership status in a loyalty program, for example) and with data obtained from third party providers."
It is super interesting that they have no terminology at all for data collected via shadow profiles. I propose "zeroth party" data - they are shadow profiles after all. :-)
The only reason they're even taking this step is because they know there are many other companies that could get just as much if not more data about its users as Cambridge Analytica did from the data brokers that partnered with Facebook.
This would be true especially now that the "cat is out of the bag" so potentially many other malicious groups/rival states could be trying to do the same thing as CA did.
Facebook simply saw the inevitable: Cambridge Analytica multiplied by 100 in its future.
But one question remains: has Facebook made sure that the data brokers deleted all of the FB data, and did they just ask them to "certify" this, as they did with CA in 2015, or did FB audit these companies? (as Zuckerberg committed to do recently with all the companies they suspect of abusing its policies?) Did FB audit Palantir, too, and its use of Cambridge Analytica's data?
https://www.cnbc.com/2018/03/27/palantir-worked-with-cambrid...