空运和地面运输：亚马逊供应链服务能在履约市场中占据多少份额

Transports: An empirical look at how much of the fulfillmentmarket Supply Chain by Amazon could realistically eat Over the last several weeks, the market has been trying to gauge the potential for AmazonSupply Chain Services and what it could mean for the shipping market (UPS and FDX). Intoday’s note, we lay out the supply chain math behind the structural limitations of a forwarddeployed inventory model and quantify what percent of the market might realistically use“Supply Chain by Amazon.” David Vernon+1 917 344 8333david.vernon@bernsteinsg.com Mark Shmulik+1 917 344 8508mark.shmulik@bernsteinsg.com Background and context. Supply chains are engineered around specific strategicpriorities that vary greatly across industries. Design choices reflect the product’scharacteristics (value density, shelf life, damage sensitivity, handling requirements),customer requirements (order quantity, speed, reliability, customization, kitting), anddemand patterns (volatility, seasonality, geographic concentration). They are also shapedby cost-to-serve economics, as shippers work to minimize the sum of transport costs,inventory handling costs (including storage and real estate), and soft costs like financing,obsolescence, and markdowns. Because these factors vary across verticals, Amazon’ssupply chain is not suitable for many industries. Justine Weiss+1 917 344 8433justine.weiss@bernsteinsg.com Deeksha Pandey+1 917 344 8447deeksha.pandey@bernsteinsg.com Wenhuan Chang+1 917 344 8546wenhuan.chang@bernsteinsg.com Our thesis: Amazon's forward-deployed (local-stocking) model is only suitablefor a subset of supply chains.That model wins on “cheap-to-carry, fast-to-sell, low-value-density” assortments where the shipper lacks the volume to extract a strategiccarrier discount. It is structurally uneconomic for “expensive-to-carry, slow-to-sell, high-value-density, heterogeneous” assortments (unless turns are unrealistically high). Thedecision between the two models comes down to the trade-off between Holding Costs,Slotting Costs, and Transport Savings (Exhibit 1). We reached this conclusion by buildinga comprehensive, transparent toy model to pressure-test the thesis and then abstractingthe key findings to package in a more compact, tunable model (Exhibit 2). Thekey structuraldeterminantson which companies might be able to use a forward inventory modellike Amazon's areproduct valueandhow much of the transport savings can be realized(sensitivities in Exhibit 4). The cost wedge that decides it is not transportation, it isinventorycarrying cost plus fixed slotting footprint. Specialist Sales Steve Song+1 917 344 8401steve.song@bernsteinsg.com We estimate an upper boundary of ~20% of current physical retail sales could besuitable for a forward-deployed model (Exhibit 5). Tolerance to data protection andservice failure limit the realistic market to low-to-mid single digits (Exhibit 6), withsector level considerations laid out in Exhibit 7. Our goal here is to present a frameworkan investor can use to reach a conclusion on how big of a risk this is — as such, while wepresent our view, there is as much value in the analytical framework (and if you think morethan 1 one in 5 suitable shippers might take the plunge, then the risk is higher). BERNSTEIN TICKER TABLE Source: Bloomberg, Bernstein estimates and analysis. INVESTMENT IMPLICATIONS We do not see a significant risk to UPS and FDX parcel volume from Amazon’s Supply Chain Services launch. Amazon has builtan amazing fulfillment model, but supply chain constraints limit the addressable market to ~20% of the market (based on retailsales). Of that amount, data and service tolerance factors would likely further limit the risk to a low-to-mid single digit number.Like all things Amazon, one has to pay attention, but the inventory hit to higher value shippers will likely mean continued demandfor longer distance transportation networks. DETAILS Over the last several weeks, the market has been trying to gauge the potential for Amazon Supply Chain Services and what it couldmean for the shipping market (UPS and FDX). In today’s note, we lay out the supply chain math behind the structural limitations ofa forward deployed inventory model and quantify what percent of the market might realistically use “Supply Chain by Amazon.” BACKGROUND AND CONTEXT Before we dig into the model, we are going to set the stage by saying that supply chains are not one size fits all. There is no onesupply chain to rule them all. Supply chains are engineered around specific strategic priorities that vary greatly across industries.Design choices reflect the product’s characteristics (value density, shelf life, damage sensitivity, handling requirements),customer requirements (order quantity, speed, reliability, customization, kitting), and demand patterns (volatility, seasonality,geographic concentration). They are also shaped by cost-to-serve economics, as shippers work to minimize the sum oftransport costs,