IF INTERESTED #166 : Amazon's Performance Management
IF INTERESTED #166 : Amazon's Performance Management
Some notes on the leaked Amazon performance management documents
hi,
I am working on a few IF INTERESTED topics, but they seem to progress slower than usual. I will blame the summer and the fact that I started a new job.
Recently I came across the leaked Amazon performance management documents. Since it’s an easier topic, and I felt like having a lighter IF INTERESTED issue, I decided to get this one out.
Here is the link to the article, and it starts with a familiar visual:
Yes, it is an extended 9-box grid. I don’t particularly like 9-box grids as they box human beings into boxes of rigid descriptions. Especially the “potential” axis has been too subjective and detrimental for organisations. However, it doesn’t hurt to get some learnings out of this 9-box grid, as it is one of the better ones.
It is a 7-Scale performance rating grouped in three categories. Initial reaction, sounds good. The scale of 7 will allow nuances within the category (see above visual). They will have difficulties to explain the difference between a “2” and a “3”, but the three main categories will help get the main rating across. The leaders usually look for a scale that allows them subjectivity and nuance, for that part a 3 main and 7 sub category scale as Amazon uses, works good. Simpler scales of only 3 like “limited, successful and extraordinary” can be too generic and uncomfortable for some leaders.I have seen that with engineering departments.
Nuances. There is a certain insight knowing the difference between a “3” (out of 7) and a “1” (out of 7). Both of those ratings are “Needs Improvement”, but the leader can express a difference. A “3” rating means that with coaching the person will do much better, but a “1” is a clear misfit and should change the job or department. The scale of 7 grouped in 3 main categories sound good, but leaders need to be equipped for effective use and communication of those ratings.
I love the “High Bar” statement. The fastest way to demotivate a solid talent is to give them “Meets Expectations”. However, if it is known that as an organisation you set high goals, and reaching them is a big deal, and it is in the wording of your performance management, then it probably is a different context. “Meets High Bar” or “Exceeds High Bar” have a better meaning. Words are important, but an aggressive statement like “High Bar” should be followed up by practice. I bet Amazon has a real “high bar” culture.
Real life use. Would be interesting to see the actual distribution of their ratings. Are all of the 7 categories really used? Or everyone is centred around 3-6, hence in practice it is only a 4 scale.
Potential. I really am not a fan of the word potential because it has a static feeling. Once you are high potential (or HIPO they used to be called), then you are forever a star, everyone else is just “average”. Once a HIPO you get all the investment and attention of your company, and non-HIPO people might fall out of the radar. In reality, our capacity to have more complex assignments changes depending on where we are in life. Our “potential” is dynamic, and depending on the context and field we are in. Also we humans are bad at predicting distant future, potential assessments for the next 5 years are unrealistic and full of biases. It is best to keep the estimations of someone’s capacity to a shorter term, like 2 years and continuously evaluate it. There is a much better approach to potential by J Garvey-Berger in her book Changing on the Job.
How vs What. Amazon apparently measures what is delivered and also how it is delivered, from a more behavioural perspective. This is a classical approach that many companies use.The HOW part should be evaluated, after all you don’t want people delivering results in expense of your values. The question is do you want to complicate your performance system having this as a separate dimension.
Some more insights about the descriptions:
Performance Scale Descriptions:
The WHAT. The descriptions for three categories are overall good. They are simple, leave space to the leader, and give guidance with real world examples. What I found odd is lack of emphasis on delivery. Only one row out of five mentions the results and delivery. Companies are moving towards more impact and delivery based evaluations, critical problem solving and decision making are important tools, but not part of the WHAT, it is more the HOW part. Some decisions and solutions are bad ones, you can still pull of a significant positive impact for the organisation.
The HOW. Usually companies connect this part to their values or behaviour models. Amazon does the same without explicitly naming those. Good references to customer centricity, collaboration and synergy. Authentic and simple language use deserves an applause. The avoidance of cliche behaviour mottos like “customer first” is always refreshing. A clear link with the statements in HOW part and the values of the company should be easily perceived by employees.
Potential Scale Descriptions:
As I said above, I am a bit disappointed here. There are far better short term capacity philosophies around potential. A much humane approach around “everyone has great potential” with a more scientific framework would be more suitable to the caliber of Amazon. Learning agility, motivation, complexity and network building are the four pillars of assessing growth, I learned from Gavin Weeks.
I can’t help myself but put one paper as a reference here, to express my stance about performance management:
Performance evaluation will not die, but it should by Kevin R. Murphy : ).
If interested.