Hacker Newsnew | past | comments | ask | show | jobs | submit | twokul's commentslogin

you are making a lot of assumptions, read up on accuracy here https://engineering.linkedin.com/blog/2016/10/bringing-salar...


There's no details in there about valuation of skills and how that plays into the results.

For other industries where titles are a little more controlled, it may be a fine measuring stick, but people hire in software based on what skills you have and what you've done. Not by title.


I'd suggest you to read https://engineering.linkedin.com/blog/2016/10/bringing-salar..., the section about privacy.


I have. I'm aware I have a strong bias against LinkedIn when it comes to their data-handling. The entire company is built around selling me to recruiters, so I'm very wary about their collection of my employment information.

Edit. For example, I'm not sure if this current policy will continue into the future. It's easy for me to get used to handing them my salary information and for them to change the privacy policy under me.


The entire company is built around selling me to recruiters

I think you may have fixed this data accuracy debate/issue. If they ask recruiters only they may get nearly-accurate data.


Here is what the _current_ content says:

  One of the interesting dichotomies with compensation data 
  is that many users want this information, but don’t want to 
  have their individual data exposed or connected back to them.  
  Salary information is personal to each of our members. With 
  this in mind, and consistent with our Members First organizational 
  philosophy, one of the first goals we established when we set out 
  to embark on this project was to provide powerful salary insights 
  in aggregate without risking an individual’s private information. 
  In the end, we built a salary collection system to provide the 
  strongest protections for the anonymity of all of our members—no 
  easy task. Parts of our approach are detailed later in this article.

  Because we proactively separate a member’s submission data from their 
  Member ID when compensation data is submitted, it enables us to 
  secure the system in such a way that we cannot even support the 
  ability for a member to update their previously submitted salary 
  information - they have to resubmit. Furthermore, our system provides 
  protection not just from hackers out in the wild, but also provides 
  access control against unauthorized use by internal users.
  
If I am reading this correctly, what they seem to say is that as implemented currently, they don't have the ability to track a person's salary information to their profile and hence won't be able to sell that information.


True, but that is not the issue. The clickwrap means nothing to me when they have this rider:

"7. Right to Terminate or Modify Software

    LinkedIn may modify the Software and this EULA with notice to you either in email or by publishing notice on the Website, including but not limited to charging fees for the Software, or changing the functionality or appearance of the Software. In the event LinkedIn modifies the Software or the EULA, you may terminate this EULA and cease use of the Software. LinkedIn may terminate your use of the Software, the EULA or the LinkedIn Service at any time, *with or without notice.*" (emphasis mine)[0] 
Fundamentally, I cannot ever trust them.

[0]https://www.linkedin.com/legal/mobile/eula


  Because we proactively separate a member’s submission data from their 
  Member ID when compensation data is submitted, it enables us to 
  secure the system in such a way that we cannot even support the 
  ability for a member to update their previously submitted salary 
  information - they have to resubmit.

How is this possible? If the salary data is not associated to the user in any way can't a user submit their salary over and over again drastically swaying the aggregate salaries?


Presumably they're tracking when the user last submitted a salary, but not what they submitted.


Until 7 months later they quietly update the ToS.


7 months later they would still not have access to the submitted data that was never stored, only to future submissions.


Then when someone resubmits new data, how do they know that a previously submitted data point is now stale and shouldn't be used any more?


What I think happens:

1. You submit a salary info struct, which includes time of submission. The struct does not contain a userId.

2. Your account's lastUpdatedSalaryDate gets updated, with only day/week/month level precision.

Someone with access to the entire database may be able to connect the two if the salary set for your job is small enough.

Then they weigh salary information based on how fresh the entry is. Stale entries don't get a good weight, and you updating your salary is just another info point, and it might go into another bucket if you got a title change (ex software eng -> sr. software eng.). If you got promoted, then the old salary info is probably still relevant for people in lower level positions.


When you submit data, your current state and history will form some basis for the entry into the salery database. When you update your information, you have a different state and history, and thus will generate a new entry.

Both should be in the data.


that's a mighty big hoping that nothing in this regard will change. ever. in for-profit company. that eagerly steals all my contacts from my cell phone (to which I probably unknowingly agreed, but it's amoral).

NO.


LinkedIn waited several years without letting me know that my account information had been hacked. For me at least they lost all credibility so reading their "privacy policy" would be a waste of time.


Yeah, because I trust LinkedIn with privacy.


Sure, because the right combination of words somehow overrides LinkedIn's history of violating users' trust, of security breaches, and of sleazy practices.


we are going to be rolling out search by city soon


one of the engineers from Salary team. we talk about privacy and accuracy at length in this blog post https://engineering.linkedin.com/blog/2016/10/bringing-salar.... TL;DR; we do outlier detection.


What's an outlier? I bet a lot of people feel they're shortchanged by about $10k, and will report that much higher to set a baseline. Outlier sounds more like "AngularJS developer making $250,000 in Cleveland"


it means more accurate baseline calculation as well as median salary which will rule out things like "AngularJS developer making $250,000 in Cleveland"


This is the first salary listing I've seen that actually looks somewhat accurate (at least for some titles) I am confused why you guys aren't using the title standardization. You're surfacing and getting different results for "Senior Software Development Engineer" and "Senior Software Engineer" (whatever those mean). You should talk to someone in SNA about this. I think Qi is the guy now.

Also you're having sampling bias that's effecting levels. There's no way the median salary for a staff engineer is higher than a senior staff's. Again talk to Qi.

Stay cool.


Why didn't you buy industry standard employer wage reported surveys like any employer of size does? They are readily available and the cost would be a drop in a bucket for LNKD.


Because they'd rather make and sell their own, obviously


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: