How "Bruising" is Amazon's Use of Workplace Algorithms?
How "Bruising" is Amazon's Use of Workplace Algorithms?
The New York Times' article on Amazon's intense corporate work environment revealed mountains of employee metrics and data. But how insidious is this e-commerce giant?
Join the DZone community and get the full member experience.Join For Free
Hortonworks Sandbox for HDP and HDF is your chance to get started on learning, developing, testing and trying out new features. Each download comes preconfigured with interactive tutorials, sample data and developments from the Apache community.
Amazon.com is a billion dollar retailer. Bloomberg Business reports that their revenue from 2014 to this year increased from $74.5 billion to $89 billion with an overall quarterly gain of $241 million. In their annual Top 100 Retailers list, the National Retail Federation put Amazon.com in fourth place in its list of mass merchants behind Wal-Mart, Costco, and Target for retail sales in the United States last year -- a significant achievement considering Amazon.com has no physical storefront for its goods. Amazon.com is growing so quickly that they are reportedly finishing up their brand new office space (or "tower" as The New York Times called it since it spans 37 floors) with plans to add additional towers on its South Lake Union campus in Seattle.
But statistics on employment retainment, satisfaction, and departures? That's the kind of data Amazon doesn't want to disclose.
If there was one thing that you could take away from The New York Times' scathing analysis of the online retail giant's work enviroment, it was that Amazon doesn't actually need any managers in its corporate headquarters. From peer reviews to how many boxes are packed in a warehouse per hour, Amazon has distilled almost everything in the workplace down to numbers. Managers are there, more or less, just to tell someone to pack up their belongings -- although Amazon declined to release their own numbers concerning employee retention.
The article, written by Jodi Kantor and David Streitfeld, describes how such use of data analysis of employee performance, while effective in culling the weak from the herd ("purposeful Darwinism") and staving off any onset of the IBM-esque "bureaucracy, profligate, spending, [or] lack of rigor" that Jeff Bezos abhors, actually resulted in a callous work environment. Good employees, according to Amazon, are robots, i.e. "Amabots," who are highly competitive and willing to throw their colleagues under the bus with little to no emotion.
For full disclosure, I am a devotee of Amazon.com. I have an Amazon Prime Membership which I use to take advantage of free two-day shipping and to stream TV shows, and I also use an Amazon.com Cashback Rewards credit card. The site provides me with a lot of savings on items, from running shoes to replacement cables. When I was in school, I always got textbooks from the retailer at a fraction of the cost my on-campus bookstore offered.
So I wondered: Was I contributing to a paradigm of corporate evil?
Where Amazon Gets Its Money
One of Amazon's massive online fulfillment warehouses.
One thing that the Times article forgot to mention is how Amazon.com makes its money.
Remember the figures from Business Insider? Their $241 million gain was actually a loss in comparison to the quarter before, which was $274 million. So how does Amazon stay on top? According to former Amazon employee Eugene Wei, it stays alive thanks to its fixed cost business model.
While you save money on things like household items and electronics, Amazon makes a profit based on the sheer vastness of what it sells in what is a mutually beneficial relationship. Sellers reach the masses under the flag of Amazon's brand recognition and the trust users place on the site. Amazon takes a small cut of each transaction, but at the end of the day, the bulk of the money goes to the seller -- as does the processing and shipping cost. Therefore, the cost of that transaction is very small for Amazon.
So why the quarterly losses? Because Amazon has seen that shipping things to people quickly and cheaply makes them happy and, in turn, want to buy from their site more. Therefore, they've been investing in fulfillment warehouses all over the country (not to mention internationally) in order to keep buyers and sellers coming. This takes out a significant chunk of Amazon's profits, but thanks to investors and its AWS services, its fixed cost base gets higher and higher. Wei describes it like this:
"[Y]ou sell a glass of lemonade for $2 and it only costs you $1 to make it, and you decide business is so great you're going to build a lemonade stand on every street corner in the world so you can eventually afford to move humanity into outer space or buy a newspaper in your spare time, and that requires you to invest all your profits in buying up some lemon fields and timber to set up lemonade franchises on every street corner."
At the end of the day, Amazon cares about its customers and wants to make you happy -- in fact, that's a heavily emphasized aspect of Bezos' work culture. On Amazon's Jobs site, their list of "Leadership Principles" starts with a tenant on Customer Obsession: Earning and keeping customer trust.
Everything that happens in an Amazon.com transaction, including what people put into their shopping carts and whether or not they leave it there or come back to purchase it hours later, is all part of the continuous flow of live data streaming. The retailer's workers then scramble to analyze it all and make sense of what that information entails for the retailer as a whole.
Such data has a direct impact on employee performance and their own metrics; meetings called "business reviews" heavily feature printouts summarizing the previous week or month's data, at times 50 to 60 pages long, which employees are expected to read. Failure to do so can result in an embarrassing moment if a pop quiz is given or if an executive cold-calls on you about a specific number on a specific page -- and while such knowledge can help make a customer's experience better, it can come at a cost to employees.
The performance algorithms for employees' tasks are a little less 1984 and a little more Mean Girls, particularly for the white collar employees. Kantor and Streitfeld report that Amazon utilizes a "self-reinforcing set of management, data and psychological tools." One such tool is Anytime Feedback, an anonymous way to submit peer reviews to management if someone notices a colleague slacking off.
It's a great way to encourage honest feedback and teamwork -- and it's also an excellent way to undermine your competition in what some ex-employees called "a river of intrigue and scheming." Ex-Amazonians described how some groups would make secret pacts to bury a colleague they didn't like while showering one another with praise because managers would often take their reviews directly from the Anytime Feedback stream.
If it sounds like those piles of data could deter employees from other work, that's because it's on purpose. Reading and understanding business metrics and submitting to Anytime Feedback is supposed to give employees a better handle on what or who is performing well and what or who isn't.
The tales of woe covered in the Times stretch from sobbing at desks to one employee receiving a poor performance review following the stillbirth of their child, all shrowded under the company's extensive confidentiality agreement. A human resources executive recounted having to place an employee with cancer on a "performance improvement plan," i.e., not-so-subtly hint that she might be fired if her performance didn't improve.
Corporate isn't the only place in which expectations and competition is intense and you leave your notions of a work-life balance at the door. Factories are outfitted with electronic sensors that monitor how many boxes are completed per hour. In 2011, a local newspaper in Lehigh Valley, Pennsylvania ran a story detailing how, at one local Amazon warehouse, workers were forced to work until they literally collapsed and were only then allowed to leave by way of the paramedics waiting outside the factory doors.
Big Brother or Transparency?
To have so many eyes watching your every move at work, weighing all pros and cons with an algorithm, sounds like a dystopian future. However, such transparency could also increase productivity and weed out people who truly don't have the work ethic to make the cut. How many times has a developer given a false estimate for a task? Doesn't it make sense to have an algorithm that can cut out that sort of behavior? Sean Boyle, who runs the financial division of Amazon Web Services, describes such use of data as "liberating" thanks to the clarity it provides.
In another of Streitfeld's articles, published this past Monday, more and more companies are harnessing data to get a better feel for employee productivity. When companies begin to adopt technology that tracks activity, they found that it was a much more accurate assessment of how an employee's time was used. Mitesh Bohra of InfoBeans told Streitfeld that after adopting such monitoring programs, they uncovered new ways to cut out distractions, like the black holes of productivity that are meetings, and to focus smaller teams on bigger tasks to maximize time.
Certainly, we shouldn't discount the experiences of those who lost their jobs due to poor job performance following the loss of a loved one or while battling cancer. There is no doubt that in all levels of corporate life in all companies, there may be a certain level of corruption or mismanagement. But Amazon is a top online retailer for a reason: Its use of "purposeful" Darwinism has left only those most dedicated to going above and beyond for their employer. It's clear that Bezos only wants a certain type of person working there, and he has mastered data analysis in order to do so.
“This is a company that strives to do really big, innovative, groundbreaking things, and those things aren’t easy,” said Susan Harker, Amazon’s top recruiter. “When you’re shooting for the moon, the nature of the work is really challenging. For some people it doesn’t work.”
Glassdoor reviews reflect the same sentiment, as 82% of reviewers approve of Bezos' leadership but only 62% would recommend working at Amazon to a friend. You may hate how he does it, but love what Bezos does.
Opinions expressed by DZone contributors are their own.