Resources: Blog Post

  
August 27, 2015

Metrics Gone Mad at Amazon

Metrics Gone Mad

Any financial adviser will tell you that it is impossible to predict the performance of an individual stock from daily trading results. Look at any stock chart over time and it often reveals massive fluctuations in daily performance even though the long-term trend may be positive.

So why would a company operate a performance management system predicated on an employee’s daily performance? If micro-oversight is misguided when applied to equities, applied humans it borders on the ridiculous. Yet, as reported recently in the New York Times, this is exactly the approach championed at Amazon.com.

According to the report, the tech giant “wants to be in the vanguard of where technology can take the modern office: more nimble and more productive, but harsher and less forgiving.” Amazon apparently prides itself on a perpetual flow of real-time, ultra-detailed metrics.

While, automated human resource systems make continuous individual performance measurement possible, why would an organization want to? Having a manager or peers constantly scrutinize performance is a recipe for inertia because we learn iteratively. Improvement comes from trial and error and that takes time.

Paul Boston, President of Actus Performance Inc., is a Toronto-based human performance expert. He knows from experience as an elite level competitive athlete that “performance improvement comes from direct and specific constructive feedback, where people are continuously reaching achievable goals.”

When and how feedback is delivered is crucial to performance improvement according to Boston. Ironically, Amazon compares its best employees to top performing athletes yet it encourages them to send feedback in secret directly to colleagues’ bosses. Such surreptitious actions fly in the face of the methods athletes use to train taking oversight to Orwellian levels.

Added to this Amazon uses stacked ranking, the much-maligned performance system that ranks employees on a bell-curve. While the statistical model may be easy to understand, research shows that it does not accurately reflect the way people learn and develop. Instead, it has been shown to undermine performance and employee engagement. The system is being abandoned by a growing number of organizations most recently by Accenture.

A survey of more than 5,000 tech workers last week by TINY Pulse asked them about their experience on the job. The results show tech workers are falling behind on several areas of job satisfaction. Only 17% report they feel valued by their employer.

TINY Pulse rightly concludes the drivers of disengagement they identified are “holding IT employees back from doing their best work. This isn’t just bad news for the tech industry it hurts everyone whose work relies on their innovation.”

We are able to track and collect dizzying amounts of data. Without the perspective of a company’s performance over time or the economic context in which it operates, what conclusion could you make about its stock based on a single day’s trading activity?

Before embarking on Amazon’s brave new world of data driven performance we need to explore the notion that more, ever-detailed data are inherently good. Where human performance is concerned, data are only valuable if they support a specific outcome, teach us something and help us to close gaps.

Is the Uber business model an über idea?
Join the conversation on LinkedIn

Posted with permission.

About the author

Tracey WhiteTracey White is a negotiator, mediator and coach who specializes in strategic planning, execution, business operations, and analysis. She combines conceptual business acumen with a focus on metrics and data analysis to support evidence-based decision-making, planning and priority setting. Her strengths include Enterprise Project


Filed under: data, feedback, management, measurement, metrics, performance, tracey white Tagged: data, feedback, management, measurement, metrics, performance, tracey white
LinkedIn Facebook Google+ Twitter