About that COVID-related “productivity drag” at Amazon

Amazon’s CFO recently explained that about $4 billion quarterly in additional costs during the third quarter were driven by unspecified “productivity drags” related to COVID-19 had cost the company $4 billion in the third quarter of 2020. What does that mean?

As usual, Amazon  conceals more than it reveals. Maybe it’s all about clusters of sick people in the warehouses desperately moving packages around until they collapse and die. Fortunately (for them and us), the reality is different.

Amazon drives productivity mainly through sticks rather than carrots. Its algorithms flag workers for firing if they miss their quota, and make no allowances for individual circumstances (absent a doctor’s note). Older, frailer, injured – makes no difference; the same quotas apply to everyone in the same warehouse. And those quotas are consistently growing: Amazon itself has admitted that when 75% of workers are making quota, the quota gets raised. Instead of the time-and-motion men who used to help run factories, workers supervise their own work, with help from technology – a big electronic timer on the workstation shows each worker exactly how long in seconds they have been working on each item. They know that if they are expected to shift 350 items per hour, that clock on average has to restart every 10.2 seconds for every minute of every hour of a ten-hour shift.

In the face of COVID-19, Amazon turned off individual productivity benchmarks. The clock stopped, and workers were no longer held accountable for meeting benchmarks. The “productivity drags” that Amazon talks about can be explained by the decision to turn off benchmarks.

In 2019, Amazon’s shipping costs were $37.9 billion and fulfillment was $40.2 billion (from Amazon’s 2019 Annual Report), of which an estimated $26 billion was spent on the warehouses and the movement of goods. It’s not possible to exactly break out the warehouse portion of that $64 billion in shipping and goods movement costs, but it’s reasonable to assume that it’s somewhere between half and two thirds.

Assuming that the warehouses actually cost Amazon $30 billion to operate – which is less than 10% of the value of the goods flowing through them – a 15% decline in productivity would more than account for that $4 billion in COVID-19 productivity drag. And it seems very likely that the sudden removal of productivity benchmarks and of the associated harsh enforcement that’s the norm in Amazon warehouses would cause productivity to slip significantly. Additional slippage probably came from people getting sick and missing shifts, from the massive disruption caused by the change in goods flows related to COVID-19, and for workers being reluctant to take shifts during the pandemic.

The implications are also pretty clear. Amazon turned the productivity benchmarks and harsh discipline back on in October. That will “solve” most of the productivity problem. Amazon has also pretty much digested the enormous boost to ecommerce caused by the pandemic, so efficiency is improving. Amazon also stopped paying pandemic pay, which would have generated more “drag.” And workers have fewer alternatives – staying home with unemployment is less of an option as benefits run out. So that “productivity drag” will likely vanish for the last quarter of 2020, assuming that Amazon doesn’t completely implode from the huge volumes running through its warehouses.