Knowing When It's Time To Go
I was closing in on my 6th year at Amazon. Since joining, I'd done quite well: I was promoted, took on newer and larger scope, and had generally gotten better at the game of work. I was sharper, built stronger mental models, became a better collaborator and stronger executor. Going into my 6th year, I was closing in on another promotion after getting two strong performance reviews in a row.
Yet despite earning top of the performance band back-to-back years, my salary was decreasing by $30k.
Amazon's compensation strategy includes a few levers: base salary, sign on bonus, and restricted stock units (RSUs). When new hires join, they get their base salary and shifting values in sign-on cash and RSUs. Based on performance and the salary bands for job families at particular levels of seniority, these values are adjusted each year, usually in a positive direction for the first four years. This provides team members with a clear path and incentivizes them to stay through the first four years, as stock compensation ramps up in years 3 and 4, which are often the most lucrative long-term and make up for the loss in cash from the sign-on bonus ending in the second year.
After year four, continued increases usually come from promotion, moving to new (and more lucrative) job families, and high performance. Employees often refer to the "vesting cliff", where after their fourth year, they see a meaningful decline in total compensation because their initial grant payouts end. Unless employees get promoted or receive refreshers for high performance, they will see a drop.
I'm mixed on this strategy. On one hand, I appreciate the performance incentive to employees to make earning high performance ratings a meaningful differentiator. In 2025, Amazon revised its policies to provide escalating bonuses to employees who consistently scored into the highest performance tier, being able to earn 110% of their band, meaning they can exceed the standard band if strong performance is sustained. Lower performers get smaller increases, if any. On the surface, this feels fair and draws a clear throughline between performance and compensation.
On the other hand, employees seeing wage decreases despite achieving significant tenure, feels like a losing strategy. Demotivating tenured employees who possess institutional knowledge and have survived in what is already a tough, high-achieving environment, feels counterintuitive. Encouraging even average performers to leave and replacing them with new hires -- who will inevitably cost more -- creates churn, loss of institutional knowledge, and a worsening reputation of the firm by pushing people out. By definition, the average performer is competent at their job. The cost of replacing them is higher than the cost of retention.
So when someone labeled a "high performer" also sees a significant compensation drop, how should this be determined? While a promotion would help, promotions at higher levels take time to put together, require the right level of scope and visibility with leadership, and also don't immediately solve the problem of the drop in compensation, especially when the market will be able to offer an increase on a much shorter time frame. Somehow, despite proving that one can survive and even thrive in an environment as notorious as Amazon's, employees may see their compensation drop.
It is rare to see any field where more experience leads to less pay. The only examples that come to mind are past their prime athletes who take shorter contracts at lower pay in their twilight years, or executives that are intentionally choosing a less stressful role in the C-suite and choosing instead to consult. In both cases, there is a shift in expectations that makes the greater level of experience "worth" less in terms of salary.
For me, I could stick it out, investing most of time schmoozing to get the support needed for promotion and maybe earn more in a year's time, or I could explore my options outside the company. I decided to go with the latter.
Is this really what you want for your so-called "top performer"?
Unlike the typical adage of "people don't leave jobs, they leave managers", I did not want to leave my manager. I had solid relationships with my leadership and they were supportive of me and valued my work (hence the high performance rating). This wasn't a situation where management wanted to push me out unfairly. They wanted to help. So I wasn't leaving a manager, I was leaving a system.
At the scale of Amazon, losing one employee over something like this would hardly register more than as a few basis points in a "regretted attrition" stat in a quarterly doc somewhere, but this is a situation that could be avoided altogether with stronger system design. While the company would certainly continue to thrive without me, the cost to find and ramp my replacement would be greater than if they hadn't dropped my salary significantly.
This is one of the reasons why performance management is so important to get right. I was deemed a strong performer, yet the system wasn't set up in a way to properly retain me. To continue earning at the same level I enjoyed previously or to even increase my salary, I'd need to invest time away from my core responsibilities to go on the campaign trail and lobby leaders for support and spend time writing up my accomplishments to put in front of a promotion panel.
Alternatively, I could make a case that I belonged to another job family that would increase my total compensation further. Almost everyone who was at Amazon for a while wanted to be in a product management job family because it offered one of the best pay bands for non-technical and non-sales work.
I could've spent my time trying to weave a narrative about how my work was actually a product I was managing to justify increasing my salary further. Both cases generally felt wrong to me -- wasting company time for my self-enrichment, when the performance management system should be capturing my accomplishments and rewarding me accordingly. While everyone needs to do selling to highlight the value of their work, the extreme load put on the employee, who is evaluated against their core deliverables, feels again counterintuitive. If I was a leader, I would not want my employee focused on navigating internal complexities to earn more money -- the system should be set up to identify and reward accordingly based on their efforts and outputs, allowing employees to focus on the work that drives value in the first place. Towards the end of my tenure, I was encouraged to spend more time focused on improving my optics to get promoted than actually doing the work that supported customers and got me in a position where I would be considered for promotion in the first place. Many conversations became about how to get "visibility" with leaders so they could support the promotion.
I share this personal story not to garner sympathy, but to highlight the importance of building systems that encourage growth and retention. Do you want your employees spending time trying to justify that they should be earning more, or do you want to build a system that enables them to do their best work and rewards them accordingly? In short, performance management systems should serve to identify top talent and enable them to do what they do best: their work.
Performance management as employee success.
The software-as-a-service industry popularized "customer success" as a function for retaining clients and growing revenue over time. Instead of treating "customer service" as a cost, it is treated as a revenue function, consulting clients on their use of software to get the most out of their investment and encourage future growth and investment in the software.
In my view, performance management should is better positioned as "employee success".
While there are tangible elements of this that are more administrative in nature, such as compensation, benefits, and general "HR" interactions being as seamless as possible (like getting a pay stub for an apartment application), there is also the element of doing your job better. Beyond table stakes items like healthcare and salary, performance management is the next level up from these basic necessities. No one is going to work for you if you aren't paying them. From there, we can focus on behavior change and skill. Ask: what behaviors do we value in this organization and how do we create the infrastructure and build the skill internally to build and help others build towards this ideal?
Performance management, at its core, is the systematic process of setting expectations, providing ongoing feedback, and developing your team members to achieve both individual and organizational goals. This goes beyond the dreaded annual review, but is a system with clear goals and measurable outputs.
The goals set out to drive employee success help inform everything we discussed earlier in recruitment and training, serving as a feedback mechanism for those functions and ensuring the clear throughline during these critical periods. By understanding what makes for a strong performer, we can figure out who we need to hire, what we need to train them on, and how to provide continuous feedback and support. We will focus next on the latter point.
But what about jobs without good outputs?
This is a fair point. Sales jobs are easy: you can measure all relevant activities and more importantly, point to the revenue earned. Customer Success is similar, with retention, renewal rates, and upsells as key performance metrics.
In other functions that are more traditionally viewed as "cost centers", like IT, Finance, HR, Operations, and so on, how do you evaluate their outputs? This is nuanced and really depends on the organization, which should establish clear goals and activities for these roles and how they tie to the overall performance of the business. For example, a component of an HR person's review should be based on reducing the amount of time employees spend trying to get basic paperwork for apartment leases. There is a gamification element to it, but if thoughtfully designed, you can drive people towards what you want them to do and measure the outcome -- allowing you to reward accordingly.
This of course, can be gamed. I am hesitant of goals around activities or inputs that can be misrepresented. For example, is driving thousands of engagements on training courses actually valuable if you're not training the right people? Does it matter if engineers ship a hundred features in a quarter if all of the features are basically broken down to lines of code? There's a lot that can go wrong here, so a basket of metrics: inputs, guardrails, and outputs, can help provide a more holistic view that drives the right incentives.
What does good performance management look like?
Performance management starts with clear expectations. Assuming we have our core values identified and an understanding of how they are lived broadly, we then need to go a level deeper: how are these values best demonstrated in different roles and at different levels of seniority? This becomes our benchmark for good performance that we can use to fairly evaluate employee performance and provide coaching and feedback to help improve outcomes.
Like in recruiting and training, clarity upfront is essential. While we may have different favorite films because we value different aspects of them, at work, we need to align on what we are evaluating so we can get to the most objective version of "strong performance" as possible to ensure the performance management process is fair and consistent.
Managers hold a great deal of sway in the corporate world on employee career trajectories. Mitigating bias, both positive and negative, is important for creating an equitable, scalable, and consistent process for identifying and supporting talent in the organization. This only grows more important at scale when team members are more fully removed from the process.
When designing performance management functions at the company, you need to consider the below:
1) Evaluating what good looks like; what contributes to strong performance 2) The mechanisms for evaluation; what is the process for formally evaluating team performance, determining compensation, and identifying high potential talent 3) The mechanisms for internal transfers and promotions 4) Compensation 5) Coaching philosophy and mechanisms 6) Feedback mechanism to inform future hiring and training needs
Performance management is inherently personal to the employee: it's their performance being evaluated. How their past actions are perceived effects their future prospects at the company, influencing their decision to stay or leave the organization.
However, performance management is also a critical organizational system that ensures the company manages its talent properly and fairly. Aligning the incentives of employees and organizations is difficult. The system designed needs to consider the agency of the employee in the process and determine how to best enable them to do their job while facilitating their growth: not only for their own development, but to support the company's goals. This is a continuous dance that must be navigated carefully.
We'll discuss more on this, but the core takeaway is that performance management is a system for evaluating and rewarding employees to get the best measure of their contribution. If designed well, systems should align personal and business goals, mitigate bias, and help reinforce company values and develop the attributes needed for employees to be successful.