Subscribe to DSC Newsletter

What if the Story Doesn't Match the Data? NYTimes & Amazon Case Study

Amazon’s practices & tools make it one of the most important companies in the world of data science. Given this, unless we have been off the grid over the past few days, it is almost impossible to ignore the tech & engineering community discussion about the New York Times article on the brutal culture at Amazon. The article shares anecdotes of employees being expected to work long hours with a brutally frank atmosphere & how unhappy employees are. Defenders of Amazon have called it a biased viewpoint based on anecdotes from dissatisfied ex-employees. Silicon Valley luminaries have come down largely in favor of Amazon saying that this is normal for high growth innovative companies. Instead of opinions, perspectives & finger pointing, why don’t we use data to shed more light on the situation.

One of the richest sources of information on employee satisfactions & ratings is Glassdoor. Employees go on Glassdoor to rate the company, share salary information & view job information. Each person can rate the company overall and also along several factors such as compensation, work-life balance, career opportunities etc. We decided to use the dataset to focus on employee satisfaction & ratings given by engineers for Amazon & its peer group. This group of 20 companies are the top publicly listed tech & internet companies and includes Microsoft, Google, Apple, Facebook & LinkedIn.

Our rankings showed that Amazon was the median ranked company with ratings of about 3.5. Companies with similar ratings include Apple, Twitter, Yahoo, Microsoft, Intel & Cisco. While the ratings are below top ranked companies by engineers such as Facebook & Google, being in the same level as other iconic companies such as Apple & Microsoft is not a terrible thing.


This case study of anecdotes Vs data goes to the heart of discussions on data driven journalism Vs story driven journalism. Traditionally, journalists interviewed “100s of people” and wrote a story from it. However, in today’s world when data is easily accessible, it may disprove the opinions and anecdotes that reporters base their story on. Is it time to change the reporting process in news organizations? Along with fact checkers, should there be a data science team to validate & cross check a reporter's story to see if easily available date contradicts a story line?

A table on the detailed ranking & the number of reviews for each company & notes are available on Zimlon.

Views: 4282

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Max Galka on August 20, 2015 at 10:47pm
It's an excellent point.
A Glassdoor rating is not the final word either, and you can debate its relative importance here. But it is sigificant enough that the journalist should have at least looked into it / explained why it didnt sway him.

Can't see any excuse for condemning a business based on anecdotes without at least making an attempt to see if the data agreed.
Comment by Vincent Granville on August 18, 2015 at 4:09pm

Great analysis, using numbers rather than anecdotes. My experience with Amazon is not good, but it's just anecdotal: as an author, being victim of fake reviews; as a publisher, being victinm of price dumping; and as an interviewee (job interview), being victim of IP (intellectual property) theft. I learned my lesson, moving away from Amazon, launching my self-publishing business, and if I (or someone on my behalf) ever interview again with Amazon, it will be to gain insider secrets to better trade their stock or anything learned from the job interview.

That said, I'm sure they hire great scientists - quite a few of them are my friends - but Amazon is preventing them from designing fantastic algorithms (e.g. to eliminate fake reviews) because that would hurt the bottom line. I am surprised that their margins are so low (part of it because of price dumping) when myself, as a data scientist, running a similar business (on a smaller scale), have much higher margins, largely due to automation, outsourcing, and absence of payroll. 

My guess is that Amazon did something really bad to some powerful guys at the New York Times, and they are now paying the price. 

Follow Us

Videos

  • Add Videos
  • View All

Resources

© 2017   Data Science Central   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service