I would argue that we had durability a few decades ago, as according to The Wall Street Journal "American shoppers snap up about five times more clothing now than they did in 1980." With the advent of fast fashion the cost of clothing decreased, but so did quality and durability -- but that was the whole point! You buy more pieces of clothing that are essentially "disposable." I don't think we can have our cake and eat it too. They might be able to make fast fashion more sustainable than it is today (like making an internal combustion car more fuel efficient), but the very concept of fast fashion is at its core, unsustainable.
It all depends on the circumstances of the backorder. If the item is completely out of stock and not available anywhere, then it's either an issue with the plan or, if the plan was OK, a production or logistics issue (e.g. the factory came up short or the item is stuck in a container at the port). These situations are difficult to mitigate as they are out of the control of the retailer. On the other hand, if the item is available somewhere (another warehouse or another store within a reasonable distance) but the retailer can't access it because they don't have an enterprise view of inventory and dynamic order routing, then it's on the retailer to implement or upgrade their order management system.
The real issue here is not how or when to substitute an item and with what. The root of the problem is inventory accuracy, both in-store and online. Retailers that need to substitute don't have true real-time inventory visibility on their websites and (if they're fulfilling from the store) don't have inventory accuracy at the store level. This is what causes disappointments. Fixing inventory accuracy in the store means performing inventory receiving and counting when new stock arrives, and tracking it with Store Inventory Management systems and RFID. At the e-commerce level, it means integrating directly with WMS/POS/ERP systems that feed inventory data to the website with APIs (by way of the OMS) as inventory is consumed. Substitutes are a Band-Aid solution.
With labor shortages, a climate emergency and so much waste in transportation (e.g. half empty trucks, multiple shipments per order), hyper-local fulfillment and the technologies that enable it (order management, warehouse management and automation) will take even greater precedence and enable retailers to meet customer expectations while responding to market and environmental conditions.
If the choice is between over-buying and having to mark down (and let's be honest, even at markdowns, Nike still makes a healthy profit margin off these items), vs. under-buying and having no stock and therefore no sales, the choice is obvious. This is the dilemma manufacturers faced when placing orders to their overseas factories last year, in the face of massive uncertainty and ongoing factory and port shutdowns. Nike erred on the side of caution because they have the deep pockets to do so. As for forecasting, not sure anyone could have forecasted massive gas price hikes (caused by a number of factors) impacting consumer goods, demand and spending in the last couple of quarters.
In an increasingly omnichannel world, fulfillment has grown and continues to grow in importance. Historically, when DCs did most of the heavy lifting for order fulfillment it was a supply chain responsibility. With omnichannel retail, fulfillment is now part of a network that includes stores. The responsibility of fulfillment should no longer rest solely on supply chain; e-commerce, store ops and even merchandising/allocation now play a role in ensuring a seamless shopping experience across all channels, physical and digital.
If the retailer has an alternative option for accepting returns like BORIS (buy online, return in-store) which is free, I can see the charge for shipping returns back to the retailer becoming more commonplace. It would deter shoppers from over-buying (knowing they will return anyway) and reduce margin erosion by making the returners foot part of the bill. It also makes good use of the store network.
Using stores to fulfill orders is a more cost-effective and sustainable method from a delivery standpoint for executing last-mile fulfillment because it shortens the distance the order has to travel (in a BOPIS case, it doesn't travel at all). It also enables a retailer to leverage all of their inventory across their store network, resulting in reduced stock-outs and eventual mark downs. The negatives described in the article such as picking inefficiencies come down to process and technology. A "store-as-warehouse" solution -- which maps the store and/or stock room and directs associates to the item location with a handheld scanner -- streamlines the entire picking process with picking path optimization. This reduces the time spent picking and also accelerates order fulfillment. In addition, a store with the space should use their stock room as the primary inventory location to fulfill the majority of orders and use the selling floor as a secondary location. Regardless of whether a retailer uses the store room or sales floor, store-as-warehouse (sometimes called micro-fulfillment) technology is the key to streamlining the order fulfillment process and further enhances the store fulfillment model.
If we look at a basic transactional view of retail -- buy inventory, market it, sell it, ship it, replenish, repeat -- it seems that Amazon retail is a loss leader for its other businesses such as advertising and marketplaces. The retail side with low prices and free shipping is designed to drive traffic, and the traffic is what they sell to third-parties and make profit from. So, in the traditional sense of transactional retail, Amazon might not be profitable. But that's by design. Is there any point in trying to measure their performance? From a retailer's point of view, they have to come to the realization that they will not beat Amazon on price or shipping cost. They will need to find their unique value proposition to win and keep customers that goes beyond free shipping.
The assortment and availability of brands is what will get people to come to the store--that or low prices/value. Without brands, it'll be private-label city for Amazon, competing with the likes of George and Time and Tru, essentially an omnichannel-optimized Walmart, which would still viable for Amazon, just not brand-enhancing. The real question is, should leading brands play ball with Amazon stores?
What could be interesting is if retailers start taking ownership of the last mile. That is, use their own fleet and drivers to deliver online orders. From a technology perspective, a Transportation Management System (TMS) with fleet management, GPS tracking, notifications, etc. would get the job done with low implementation risk. The harder part would be carving out this outbound logistics division within the existing retail organizational structure.
Omnichannel retail is about the seamless integration of physical and digital retail experiences. It doesn't really matter who owns what, as long as the environments are harmonized. A website, store, pick-up location, return location, delivery company drop-off location, etc. are all pieces of an omnichannel puzzle that need to fit together perfectly, but do not necessarily need to be part of the same organization. The glue that will hold all of these pieces in place is technology, like open APIs.