Case study: Amazon Fulfilment Center

I had the unique opportunity to visit the Amazon fulfillment center in Koblenz, Germany. Amazon, the company that sets the pace in e-commerce and leads logistics innovation with picking robots and experimental drones. Inside, I found a well-oiled machine with some unorthodox elements and learned some interesting lessons.

Nearest airport

I entered through a tourniquet and walked up the yellow tower and along the bridge into the 110 thousand square meters building. In the lobby there was a monitor with the latest performance scores. The employees of CGN1, Amazon names fulfillment centers worldwide after their nearest airport, had earned a performance bonus of 6 percent in February.

Yellow semi-trailers

From the lobby we strolled into the fulfillment center, towards the goods receipt area. There, it was remarkably quiet. Despite the huge scale, only 20 trucks arrive each day, rising to 40 in peak season. Incoming shipments are pre-consolidated into full truckloads by logistics service providers. Consequently, you see many yellow semi-trailers and swap bodies with DHL imprint on the yard.

Peak season

The fulfillment center employs 1,500 people. Add another 1,000 in peak season. Amazon claims to hire people regardless of age or education. Our guide Andreas, a man in his sixties and active in the fulfillment center since the second month of operation, was grateful for this opportunity. The atmosphere in the workplace seemed good. People were friendly and willing to help me out. Although operators worked at a decent pace, it did not come across as the sweatshop that is sometimes portrayed in the media where individuals are pushed to the limit.

No bulk stock

Large products on pallets went from the goods receipt hall directly to one of four picking halls. This, in effect, is the first notable principle in this immense operation: there is no bulk stock! All stock is on pick locations.

Barcode

Small items in boxes are received at work stations. Boxes arrive without paper documents. Suppliers print purchase order numbers in barcode on shipping labels, while they communicate box contents digitally to Amazon in advance. The receipt operator scans the barcode and Amazon's systems know which items to expect. The operator then scans the individual items in the box and places them in yellow totes. Once the tote is full, the operator pushes it onto a conveyor belt. If there are open orders on an item, then the system displays a message and the operator puts the item in an empty tote that leaves directly for the cross-dock zone for expedited shipping. If something is wrong, then a lamp lights up above the work station. The operator may continue with the next item, while a problem-solverlooks into the issue on his behalf. 

Weights and dimensions

Items that arrive for the first time are brought to a special station where weights and dimensions are measured on a Cubiscan device. Results are shared globally among all fulfillment centers. Apparently, Amazon has its measures well under control, since the station was sitting idle during our entire visit.

Pick Tower

The conveyor belt delivers the yellow totes to one of four floors in the pick tower. There, an operator takes the tote and distributes its contents, at his own discretion, across shelving locations. He uses a mobile data terminal for confirmation, but it gives no guidance. Amazon calls it chaotic storage, a principle that they use since 2005. An item may therefore be stored in several locations. To preserve the fill rate of shelves, different items may be placed together in the same location.

Trans-shipments

Amazon has nine fulfillment centers in Germany. Not every item, however, lies in each fulfillment center. For example, printing on demand of books only happens in Leipzig. Amazon's systems evaluate for each customer order which site is best suited to supply it, while taking workload and stock availability into consideration. If products for one customer order come from multiple sites, then the systems make a trade-off between shipping separate parcels to the customer or merging the items. Amazon refers to the resulting shipments between fulfillment centers as trans-shipments. These arrive in black totes in the fulfillment center, which are subsequently stored in a special picking area.

Amnesty

Order pickers walk around with pick carts and pick items in yellow totes for multiple orders simultaneously. Per round they pick a batch of approximately 50 items from the shelving racks. Single-piece and multi-piece orders are picked in separate rounds. With each pick, the mobile data terminal displays how long the picker may do about it. After completing the pick round, he puts the tote on a conveyor belt. Should the picker encounter items along the way that are damaged, then he puts them in a damage location. Items lying around loose, he puts in an amnesty location. The problem-solver then solves the matter, so that pickers can always continue with their work.

Robotics

In the German fulfillment centers of Amazon you will not find any robots yet. The orange robots, rebranded from Kiva into Amazon Robotics after the 2012 takeover by Amazon, were first introduced in Europe in a brand new facility in Poland in 2015.

Year-end Peak

It was clear during the tour that Amazon copes with seasonal peaks. For example, an entire hall was kept empty awaiting the year-end peak. For the same reason, several lanes with pack stations were covered under sheets.

Wrangling

The conveyor belt delivers yellow totes with single-piece orders directly to pack stations. Totes with multi's from different zones come together in so-called wrangling stations. Here operators sort totes by hand into batches. Given the immense scale of the fulfillment center, this is an surprisingly small and simple operation. Typically, warehouses have sizable buffers and sorting installations to consolidate goods from different zones. Amazon, however, manages to accurately synchronize progress between zones, so that batches are ready at approximately the same time in all zones. To achieve this, Amazon's self-made warehouse management system needs intelligent controls. The essence of the control logic are, on the one hand, the standard times which order-pickers see on their devices. On the other hand, the chaotic storage plays an important role. Since an item is in several locations, the system can select pick locations so that the workload is evenly distributed across zones.

Re-bin station

From the wrangling station the yellow bins go per complete batch to the re-bin station. Here, items are sorted to mobile racks with slots per customer order. After the batch is sorted, the mobile rack goes to a pack station. Here, customer orders are packed. The operator closes the parcel and attaches a small barcode label. Same as with receipt, there are no paper documents. Parcels move on the conveyor belt to the shipping area. Along the way, a shipping label is attached automatically. At the end of the conveyor, staff load the parcels into trolleys. The trolleys finally leave in trucks.

Lessons learned

Now what were the most notable issues and what can e-commerce companies learn from Amazon?

1. The process is totally paperless

Definitely a best practice and an obvious tendency in the market. Yet, many companies are not ready yet.

2. There is no bulk stock

Continue reading?

Share this article

Continue reading
asds