Amazon hiring of humans has been relatively flat for 3 quarters. Meanwhile, they have continued exponential hiring of robot employees and increased automation across new generations of fulfillment centers. This trend demonstrates Amazon’s focus on their competitive advantage of headcount productivity, but also indicates the payout of past investments in automation capital.
As this trend continues, Amazon is approaching an inflection point where headcount will plateau, followed by an ongoing increase in headcount productivity, driving additional operating profit for reinvestment.
Amazon is already one of the most productive retailers with retail sales per headcount (HC) of $330K. For context, other large retailers are at $230K for the same metric. However, Amazon still lags their tech peers and Chinese eCommerce competition due to the 500K people involved in the order fulfillment process. But, as Amazon continues to invest $3B of capital per quarter in technology and automation improvements the gap will narrow. The payout from these investments will begin to show in $/HC productivity and reduced shipping cost as a % of sales.
Amazon trends on productivity, hiring, and investment
- Amazon capital investment = $2.7B per quarter. It continues to ramp at 45% YoY growth (8% above revenue growth rate)
- The pace of hiring growth QoQ at Amazon has slowed significantly, while the rate of automation and robot employees grows 2x as fast
- Kiva (automated-shelf) robots represent 8% of Amazon’s workforce
- Every new Generation 8 Fulfillment Center will be staffed with 15K Kiva bots and 2K humans
- If all facilities were currently Gen 8, Amazon could reduce roughly 10% of the warehouse workforce (37K) and free up $1.1B of cash
Why Amazon needs to continue these trends
Amazon still has some pain-points in their business model that can be improved through automation
- High cost of fulfillment = 17% of revenue (and growing) — This can be positive and negative because the high spend enables rapid delivery which is a very expensive moat vs. competition. However, improvements in this metric could generate significant profit given their huge throughput.
- Low profitability — most of Amazon’s profit comes from AWS, and the profit margin of the retail side of the business is lower than competition. Given that their low-price model will never change, Amazon is tied to delivering profit through efficiency.
- Grocery is hard to sell online — while a large white space for Amazon, fresh product and grocery have an expensive supply chain. Investments are needed to deliver even faster than the 2-day prime and minimize cost of this high-touch business.
Impact on Earnings
Holiday is typically a huge hiring season for Amazon, when they will flex up by 50–80K employees. If this trend continued in 2018, Q3 hiring numbers would be 10–20% increase over Q2 (assume 625K+ employees). However, numbers came in at only a 6% increase, demonstrating a continued flattening of the hiring trend. I believe efficiency will allow Amazon to handle future holiday shipment spikes without incremental increase the fulfillment workforce. Theoretically, more automated shipping can deal with variable capacity without the need to significantly ramp up staffing leading to fewer hiring spikes.
- Headcount — expecting 10% or less quarterly increases, as the hiring flattens and eventually declines in the warehouse. (est. sitting around 600K total)
- Fulfillment cost reduction as a % of revenue
- Revenue per HC — increase YoY
Ongoing, I believe that Amazon has reached a tipping point for automation in which HC as a % of sales will continue to decrease as more automated fulfillment centers come online. I expect HC as a % of sales will drop from 22% to 15% by 2020. This will free up $1B of operating profit to reinvest.
Over a longer time-frame, Amazon is moving to full “lights out” facilities. Generation 9 or 10 DCs will have less and less human touch until eventually the supply chain is fully mechanized to maximize efficiency. Examples of this trend include
And outside of the fulfillment center…
- Moving to self-serve vendor sales — as the business shifts toward sellers managing their own business on Amazon-as-a-platform, office HC growth will lag retail sales (20%). This was evident in recent hiring slowdown in Q1 and Q2.
- Shift to more marketplace sales — As third-party “marketplace” sales increase, Amazon gets more efficient. These sales need less HC to manage as proven by the Alibaba model.
- Increasing advertising — On-site advertising is big, automated, and high margin. This will continue to be managed with significantly less HC at a increasing revenue.
Alibaba Aside: Alibaba delivers an estimated $700B of third party sales annually from distributors, small businesses, and large manufacturers selling on their platform. To enable this, Alibaba needs 25K ops employees (66K total employees). Even assuming US cost per headcount, this model has <1% HC as a % of revenue. This has allowed Alibaba to grow rapidly in a Chinese marketplace that is now 20%+ eCommerce.
Into the future
The trends on Amazon hiring will certainly have an impact on the sector. eCommerce continues to grow as a % of US sales (to 10%), and every $1 of online sales requires 33% less people to deliver than traditional retail. If that trend holds, there number of US retail workers needed decreases by 14K for every point of growth in eCommerce sales (4.3M US retail workers). Further, if the US were at China’s level of eCommerce sales today, there would be 140K less need for retail workers.
These jobs won’t disappear, but they will certainly change shape. There will be more need for third party shipping companies, small business owners, marketplace sellers, Amazon sales people, tech talent, and a host of auxiliary businesses. While the Amazon of 2020 will have a lot less need for unskilled and semi-skilled labor in the warehouse, it will remain a large portion of the job ecosystem.