Archive 7 min read

Splitting Hairs With Facebook Testimony

Facebook was called to the carpet and showed well...because they weren't asked the right questions

Splitting Hairs With Facebook Testimony

Watch this episode on YouTube.

Reasonably Accurate 馃馃 Transcript

Morning. How's everyone doing? I'm thinking almost level. There we go. All right, camera adjusted, all set. So, um, wow, crazy week. Um I just did a radio spot this morning, had a great conversation with uh Robin Bresnahan on CB C's Ottawa morning here in Ottawa.

Um And we talked about uh the Facebook um testimony from Mark Zuckerberg in front of the US Senate and House committees yesterday and the day before, um we were talking about um a bunch of the, the impact of that and the challenges there.

Um And one of the things that really came up for me um in that discussion and read some of the coverage and all that kind of stuff and let me just readjust here perfect. Um was uh asking the right questions and I think that really um kind of surprised some people and Robin seemed a little shocked when I said, you know, they're not asking the right questions.

Um And it's not at the technical depth, it's the way they're being phrased is giving easy out. Um And I think that that really encapsulates the whole problem, um or the whole scenario. Um and the whole challenge we're seeing here uh in general.

Um Is that, so the, the question I drilled in on was really um one of the senators had asked Mr Zuckerberg, um do you sell user information? And he was able to emphatically answer? No, we don't. Um Because guess what?

They don't, they never have, they never um that's not their business model. Understanding. The business model is key here. The business model for Facebook is um selling access to users based on their behavior and a profile created of them. So I post something to Facebook.

Um And this is the same for all the social networks that are ad driven. So this is not just Facebook, it's just Facebook's getting called out for it right now. Um So I post something on Facebook and I post, you know, over time I interact with different things on Facebook like this video, um like somebody else's page like, you know, favorite music.

Um I check in from different locations or I post from different locations and Facebook is not interested in the information. What they're interested in is the signal that that information provides. So the fact that I check in here in Ottawa or that I check in when I'm in Tokyo or London and I'm by check in, I don't mean an actual like four core kind of check in.

I mean, I log into the service and they can geoloc me based on my IP or I post a photo and share that with my friends and my family. Um These are activities that I'm providing user information into Facebook. They don't want to sell.

The fact that I just had Great Shabu Shabu in Tokyo. Um The fact that, you know, hey, here's a cute picture of one of my family doing something or whatever the case may be, that's of no interest to advertisers. That's of no interest to anybody except maybe the audience that I have targeted that post to fingers crossed.

Um What is of interest is the fact that uh here is another indicator, another signal of my behavior and my pattern that profiles me as a potential consumer for advertising. So the fact that I am active in different areas and traveling around says something about me, right?

It's demographic information and that says something very specific about uh uh me. So when I'm liking certain bands, when I'm traveling to certain places and checking in based on certain festivals um or um conferences, um you know, you can start to tell like, OK, so uh middle age, like middle age, um middle age, uh tech savvy tech interested.

Um maybe with a little bit of development, security privacy slant, um probably uh reasonably um well off purchases a lot of tech gadgets, um these kind of things, right? And you build up this profile that gets really, really accurate.

I think that's what's really coming out of. This is a lot of people are shocked at how specific and accurate that profile is, but they don't even know. It's just the tip of the ice um, based on some of this testimony and that's what is user or information about a user based on user information.

And that's a very big difference, user information stuff. I'm posting information about me as a user. Very valuable behavioral model and demographic information that advertisers will pay to get access to. Now, I have um both a personal Facebook account and um that account is the, um, owner admin on a Facebook page that I put sort of my professional face forward.

So my personal account, most people never see um, other than just being posted on my, uh on my page. Um, but, you know, that's where I share with friends and family and I have that locked down my professional page.

So facebook.com/marknca, there we go. Um, is where you'll see stuff I post professionally and I separated those two around Christmas last just to make it easier because I was putting like media appearances and stuff like that on my personal feed just to show family and friends.

But I didn't really have a professional presence. But because I run a page, I also see the business side of Facebook and holy crap. They are pushing all the time for me to spend more money in ad revenue. That's their business model.

I totally get it and I don't begrudge them where this comes back to the current scandal. And the current challenges is the fact that because of, because they are constantly pushing businesses to spend money, those businesses who are the actual customers of Facebook will demand results will demand better tools will demand better action.

That's exactly what Facebook is built. There should be no surprise. I don't mean to sound like a jackass when I say that, but there should be no surprise that this is what we built. The system was set up for Facebook to build this uh outcome.

And I think now the testimony in front of the house in front of the Senate is illuminating sort of the challenges and some of the problems but not nearly to the depth. It needs to either this is good that the awareness is there.

Um But it's not quite to the level we need. Now, I'm going to show you something here on my screen. And despite the fact that you see what I was working at for video streaming in the background, um is I'm going to show you this.

Now, this is one visualization that I pulled out of my Facebook um activity. So I downloaded my personal um data uh from Facebook and I ran through a tool that I put up on github that's freely accessible. So github.com/marknca.

Um So again, marknca and you'll see that um the Facebook data download insight tool and this is a visualization that I created with the output from that tool and Amazon Quick site, which is a great little visualization service.

You can build this kind of thing in Excel as well. But I was just messing around with the cloud and made it easier. What you'll see here is unique IP addresses across the horizontal access. So ABC D blah, blah blah, these are all unique IP addresses and the vertical, these little gray boxes are all different days.

OK. So this covers two months, just February and March and this is all taken from session activity in Facebook. You can see the lighter ones across the top here was when I was traveling and you can see this actually correlates to um different positions around the world I spent.

Um I think this is a marches in the top here or sorry, February is in the top. It goes down to March here. Um I was traveling quite a lot here and you'll see I showed up in different IP S at different times.

But then as this gets darker, it's more logged activity. A number of times that Facebook has seen me logging in not posting anything, not doing anything other than reloading the page. And then it logs this and correlates this and this is user information.

So this shows that hey, I was traveling for a chunk of the month. But then it shows when I'm uh attentive when I'm on the platform most likely. So you can see here this dark bar. Um these couple of days I was on the platform quite a lot and that would be an excellent time to target me for advertisements.

This is just a small sample of the insights um that this uh data that Facebook creates around you. Um That's just the tip of the iceberg and I think that's really what needs to come out of this. So um continuing to write, it's this space of the scandal has been um spinning so fast that it's been really hard to um get on top of this to get this post out.

I'm really trying hard to get it out uh this week in light of the attention around it because I think hopefully showing visualizations like that and the ability to pull that out of your own data tools will really send the message home with how much Facebook just how much Facebook knows about you and why and what we need to do about it.

So, um thanks. I hope you guys are having a great day as always. Hit me up marknca. There we go for the webcam for some reason, marknca on Twitter. Um Facebook, github, everything you can think of uh happy to chat um down in the comments below as well.

Uh I read them all. Um I have seen a lot of requests around getting started in cybersecurity. That is a topic. I'm probably going to be tackling next week at the RS A conference in San Francisco. Um, stay tuned for more on that.

In the meantime, I hope you have a great Thursday and we will talk to you guys tomorrow.

Read next