Hacker News Reader

Comment by fps | Hacker News Reader

I had an interesting experience last year when I was working in a co-working space. I noticed that LinkedIn started suggesting to me the other people who were working in the same space. I had no verbal or virtual contact with these people, other than basic office small talk, but I never knew their name and they never knew mine. I mentioned it to them, and they all claimed they hadn't looked me up on LinkedIn, and that that was very weird. My assumption was that LinkedIn knew we were sharing the same internet connection and so they linked us together.

View on HN

I'm pretty sure Facebook does this too. I moved cities and got a new job without using my network of friends to get it and I don't keep any [accurate] personal contact data on Facebook. Suddenly my boss and coworkers are suggested people I may know despite the only thing we shared in common w.r.t. Facebook is a set of IP addresses.

I have multiple Facebook accounts (I created them for testing a long time ago). I use only 1 of them. But it's funny to see Facebook claiming that Account #2 sent a "friend request" to Account #3. And offering up the other accounts as "people you may know". The only commonality is the IP.

What do you think all the "data scientists" hired by these companies do? IP addresses (or IPGeo) is one strong signal that goes into the "people who you might know" model. One can argue how effective these models are, but it's quite likely that they are found to be useful, for some metrics that the companies care about, like user growth, revenue or whatever. It's also likely that negative sentiment on a forum like HN is not part of the metrics these models are optimized for.

As a data scientist, I have always debated the efficacy of the IP address as an identifier. It could be useful, but used as-is, I think it is not that great a signal.

Just because a signal gets you some positive results, does not make it useful. You have to also look at the second-order effects. Does it creep out people? Does it make people less likely to use your service, thereby negating the positive effect of added connections?

I had the same experience with my roommates a few years ago, I knew their names offline, but I never had any contact with them online (no email, not friends on Facebook, not followed on Twitter...) , and Linkedin suggested them to me. The only online thing we shared was our IP address...

I had a similar experience with LinkedIn, which also suggests they are using IPs to discover people who may be connected to you. In this case, I had gone into a local coffee shop, used the free WiFi, and then when I went home and accessed LinkedIn from my desktop PC it suggested several of the employees of that coffee shop as possible connections.

I thought that was actually pretty scary - normally, if you're a service employee at a business, the most a customer is going to learn about you is the name printed on your name tag. In this case, after buying my coffee and going home, I was presented with the last names of these employees and whatever other information was publicly visible in their LinkedIn profiles.

Maybe someone had all of your email addresses (like the organizer of the coworking space) and uploaded them to LinkedIn via the contact finder.

I can't say that wasn't the case, but I believe it would have required a lot of sleuthing. I was a recent hire to the company and never added my company email address to LinkedIn. The co-working space organizer, if they even had my email address, would have had my company email and not my personal one.

When I moved into my apartment building (only 4 units) the first day linkedIn started asking me if I knew my neighbor (who happens to work at LinkedIn) even though we had not had any contact. Best guess was the linkedIn app knew we were in close proximity (wifi networks around?) and thought we must know each other. Yeah, creepy.

Or perhaps you had the same workplace info on Linkedin?

we worked for different companies - even technically in different cities - I live in Boston and was working for a company in SF. My LinkedIn entry for my workplace listed the SF address.

Reply on news.ycombinator.com