This is Sunday Supply Chain Stories, where we revisit the foundations that continue to shape how inventory moves, returns and recovers value.
If you enjoy the newsletter, follow The Supply Loop on LinkedIn and ask your friends to subscribe to receive the next story.
Before there were websites, shopping carts, or fulfillment centers, Sears, Roebuck and Company sold everything by catalog. Founded in 1886 by Richard W. Sears and Alvah C. Roebuck as a mail-order watch business in Minneapolis, the company expanded rapidly under the guidance of Julius Rosenwald, who joined in 1895 and transformed a niche retailer into a national institution. The big book — eventually more than 1,200 pages, mailed free to rural households across the country — listed prices on everything from sewing machines to bicycles to ready-made suits to building materials. If you wanted something and you lived outside a city, the Sears catalog was the store. You found the item, filled out the order blank in pencil with the catalog number, size, color, and quantity, folded in your payment, and mailed the envelope. Sears shipped the goods by rail to your nearest depot, or directly to your door once the Rural Free Delivery Act of 1896 opened postal routes into farm country. That was the model: a printed book, a handwritten form, and the postal system as the only connection between buyer and seller.
Cover of the Sears Catalog. Source
By 1907, that model was processing orders at a scale that required its own city. The Sears complex on Chicago's west side covered 40 acres and contained more than 3 million square feet of floor space — the largest commercial building in the world when it opened on January 22, 1906. Twenty thousand people worked there. On a typical morning, government mail wagons (the forerunner of what we now call the U.S. Postal Service) arrived at least four times before noon carrying thousands of envelopes: handwritten order forms tucked in alongside money orders, personal checks, bank drafts, or in many cases, cash folded into the pages. A single day's inbound mail could carry customer remittances totaling as much as $353,000.
Inside the mail-opening department, a workforce of 120 clerks ran those envelopes through machines that stamped each one with the date and hour of arrival at a rate of 800 pieces per minute. The company received up to 90,000 letters and postals on a single day. Each envelope was slit, the order form removed and read for completeness, and the payment separated and verified in the Cash Crediting Division. If the order was regular in every respect — catalog number, size, color, quantity all specified — it moved to the entry department and from there into the picking flow of the merchandise building.
Orders were picked from stock, packed in wooden boxes in the freight packing room, and loaded onto rail cars directly from the dock. The complex sat on the Chicago Terminal Transfer Railway, which connected to every major line passing through the city, so a customer's order could arrive by mail on a Monday and leave by rail that same afternoon, reaching most of the country within three to four days. Letters arrived. Packages departed. Those two flows ran simultaneously and in opposite directions, and together they defined what people in 1907 understood as mail-order retail.
The third flow — packages coming back — was the one that required a different kind of infrastructure.
Richard W. Sears had been making mail-order promises since 1888, when he and Alvah C. Roebuck first sold watches by catalog in Minneapolis. By the time Julius Rosenwald joined the firm in 1895 — buying in when sales were less than $1 million annually — and helped build those sales to $50 million by 1907, those promises had hardened into formal policy. The language in the catalog was unambiguous: "If the goods we send you are not found to be perfectly satisfactory, they can be returned to us at our expense, and the money sent us, together any freight or express charges paid, will be immediately refunded." Sears paid the outbound freight. Sears paid the return freight. The customer bore no transportation cost in either direction.
Freight Classification table inside Sears catalog. Source
That guarantee was not a marketing gesture. It was an operational commitment with a physical consequence. Every item shipped out could come back, at Sears' cost, and the Chicago plant had to be ready for it. The department of returned goods at the complex occupied approximately 50,000 square feet of dedicated space within the merchandise building — a figure that, taken against the scale of the whole operation, makes clear that returns were not a rounding error but a designed flow within the system. They had receiving functions, inspection functions, restocking functions, and disposition logic, all operating under the same roof as the outbound fulfillment that generated them.
The infrastructure that made the returns viable had been assembled deliberately across the years between 1906 and 1929, as Rosenwald built out a national network of regional catalog distribution centers — nine by 1929, eventually more. Dallas was the first outside Chicago, completed in 1906. Seattle followed in 1913, and a North Kansas City warehouse opened the same year, its nine stories of reinforced concrete specifically positioned to reduce freight charges on heavy merchandise shipped to customers across the Great Plains and Southwest. Each center was located to minimize the distance between stock and customer, so that orders could be picked, packed, and dispatched within a few days of receipt — the rail leg to the customer's nearest depot adding more time on top of that. Each center also meant that returned merchandise, traveling back along the same rail network that had carried it out, could be received and processed regionally rather than rerouted to Chicago. The network was built for forward flow, but its geographic logic served the return as well.
The mechanics of a Sears return in the early 20th century required a customer in rural Iowa or Tennessee to repack the item, attach the prepaid return label from the catalog, carry the package to the nearest post office or rail depot, and wait. The return freight traveled on the same Rural Free Delivery infrastructure that had made the original catalog possible — the 1896 Rural Free Delivery Act had opened mail routes to rural areas, and the Parcel Post system established in 1913 gave the catalog business, including its returns, a low-cost national delivery mechanism that no private carrier could have built independently. Sears had leaned on government infrastructure to build its forward network. Its reverse network ran on the same rails.
By 1916, Sears was mailing more than 50 million catalogs annually. By 1927, it served 11 million customers — roughly one in three American families. That scale meant the returns volume, even at a modest rate, was a significant operational throughput problem. The guarantee kept generating it. The system kept absorbing it. The two were not in tension because Sears had built them together.
That guarantee was not a marketing promise. It was a capacity commitment, and what distinguished Sears from most modern operators is that the company understood the difference. When Richard Sears wrote the return terms into the first catalog, he was defining the full scope of what the fulfillment system had to do — not as an afterthought, but as a design constraint that shaped every infrastructure decision that followed.
Most return policies today are set by marketing or leadership teams who are not accountable for the operational cost of processing what comes back. The Sears model did not separate those accountabilities. The guarantee drove volume, and the volume drove investment in physical infrastructure to handle it. The department of returned goods existed because the guarantee required it. The regional network existed because a national customer base generates a geographically distributed return flow, and routing all of it back to Chicago would have compressed margin on every item processed.
The condition logic inside returned goods was straightforward by necessity. An item that came back in original condition could be restocked. An item that was damaged could be assessed for repair, discounted, or written down. An item that arrived unsalvageable was a loss — but the guarantee made that loss known and finite, rather than a source of customer attrition that compounded over time. Richard Sears held to the motto "We Can't Afford to Lose a Customer," and the returns infrastructure was the operational form that conviction took. A returned item was a data point about whether the forward system had worked. A refunded customer was a retained relationship rather than an abandoned one.
What the system lacked, by modern standards, was any kind of systematic data capture at the item level during the returns process. The condition decision was made by a worker at the receiving dock, and the outcome — restock, discount, write-down — was recorded in whatever accounting system the era supported. There was no item-level triage code, no grading standard documented in writing, no yield rate tracked per category. The volume required trained judgment applied consistently, which is the same problem modern processing centers face at fifty times the scale, with the added complication that the judgment is now expected to be encoded in a WMS and applied by workers who may never handle the same SKU twice.
The Sears system was not sophisticated by the standards available today. It was sophisticated by the standards of its era, which is the relevant comparison. The question it answered, before anyone had a name for it, was: what does it cost to honor this commitment, and how do we build the infrastructure to absorb that cost without destroying the margin on the forward business?
The Sears catalog ended on January 25, 1993. The announcement was made alongside the closure of 113 stores and the elimination of 50,000 jobs — the largest single-day layoff in the history of American retail at that point. The stated reason was cost: the big book had been unprofitable for decades, its distribution overhead growing faster than its revenue could cover. By 1992, the catalog operation that had served 11 million households in 1927 had become a burden on a company that had already moved most of its volume into physical stores.
The infrastructure that supported catalog returns went with it. The regional processing centers were retooled or sold as the retail store network became the operational center of gravity. The expertise built up across 80 years of handling mail-order returns — the condition assessment practices, the geographic routing logic, the restocking decisions — dispersed into the industry as the people who held it moved on. The catalog as a commercial format died, and the returns infrastructure built around it was treated as a specific-to-catalog problem rather than a general capability.
What replaced it, in the years that followed, was a retail model that assumed the customer would bring the item back to a store, where a return could be processed face-to-face, the item could be assessed by staff, and the disposition decision could be made on the spot. That model worked as long as most purchases were made in stores, because the return flowed back through the same channel that had originated the sale. The store was the forward leg and the reverse leg simultaneously.
Ecommerce broke that assumption the same way the catalog had, by separating the point of sale from the point of return. A customer ordering online in 2006 had no store to return to unless the brand happened to have a physical location nearby. The return had to travel — by mail, by carrier, by drop-off point — back to a processing facility that in many cases had not been designed to receive it. The problem was not new. Sears had been solving it since 1888. The institutional memory of how to solve it at scale had been sitting in the inventory of retired operations managers and decommissioned distribution centers for more than a decade before Amazon's return rates started making the business press.
The scale has changed in ways that matter. Consumers returned products worth $890 billion in 2024, with ecommerce-specific returns running at approximately 20 percent of total units shipped. The average cost to process a single return now runs $20 to $30, exclusive of the freight that was already paid to send the item out. The global reverse logistics market reached an estimated $768 billion in 2023 and is projected to pass $1 trillion by the end of the decade. These are not the return rates of a mail-order business serving one in three American families. They are the return rates of a consumer economy in which the absence of friction in the purchase decision has been industrialized, and in which the back-end cost of that friction-removal was not modeled into the original business case.
The specific operational problems that modern ecommerce returns operations describe — how to triage condition at scale, how to route efficiently across a distributed return flow, how to make restocking decisions faster than the inventory loses value, how to build regional processing capacity without over-capitalizing in the wrong geographies — are the same problems that Sears was solving in its regional distribution centers between 1910 and 1928. The tools available are vastly more capable. The volume is vastly higher. The fundamental questions are the same, which means the fundamental design choices are also the same: you need a guarantee that creates the return, a prepaid mechanism that enables it, a geographic network that absorbs it, and a condition-based disposition process that recovers the value.
Sears built all four, without digital tools, at a scale that served a third of the American population. The operators building returns infrastructure today are not solving a new problem. They are recovering a capability that the industry spent thirty years dismantling, because it believed the catalog was the problem rather than the prerequisite.
For practitioners: How much of your current returns volume is forecast in the same model that plans your forward capacity, and where does the gap between those two forecasts show up in your cost-to-serve? If you were building your returns processing network from a geographic first principles — not inheriting the footprint you have — where would you put the nodes, and does the answer match your current infrastructure? When a returned item arrives at your facility today, how many distinct condition outcomes does your team actually distinguish, and is that number sufficient to capture the difference in recovery value between them? And if the guarantee you offer customers creates a return, have you quantified the full round-trip cost of that return — outbound freight, return freight, receiving, processing, restocking or disposition — as a single line against the margin on the original sale?
Sources
Vicki Howard. "The Rise and Fall of Sears." Smithsonian Magazine. July 25, 2017. https://www.smithsonianmag.com/history/rise-and-fall-sears-180964181/
Melissa A. Winn. "The 20th-Century Retail Mecca That Was Sears, Roebuck & Co." HistoryNet / American History Magazine. January 16, 2024. https://historynet.com/the-20th-century-retail-mecca-that-was-sears-roebuck-co/
National Park Service. "Sears, Roebuck and Company Complex." National Historic Landmark nomination and summary. https://npgallery.nps.gov/NRHP/GetAsset/NHLS/78001129_text
City of Chicago, Department of Planning and Development, Landmarks Division. "Exhibit A: Final Landmark Designation Report, Sears, Roebuck and Co. District." https://www.chicago.gov/content/dam/city/depts/zlup/Historic_Preservation/Publications/Sears_Roebuck_and_Co_District.pdf
Library of Congress, Business Reference Services. "Sears, Roebuck, & Co. Goes Public." This Month in Business History. https://guides.loc.gov/this-month-in-business-history/november/sears
Encyclopedia.com. "Sears, Roebuck Catalog." https://www.encyclopedia.com/history/culture-magazines/sears-roebuck-catalog
MyTotalRetail. "A Customer-Centric Approach and Other Enduring Lessons Learned From the Old Sears Catalogs." https://www.mytotalretail.com/article/a-customer-centric-approach-other-enduring-lessons-learned-from-old-sears-catalogs-20917/all/
Smithsonian Magazine. "Before Folding 30 Years Ago, the Sears Catalog Sold Some Surprising Products." January 26, 2023. https://www.smithsonianmag.com/innovation/before-folding-30-years-ago-the-sears-catalog-sold-some-surprising-products-180981504/
Capacity LLC. "Returns in 2025: What Ecommerce Brands Need to Know Now." https://www.capacityllc.com/returns-in-2025-what-ecommerce-brands-need-to-know-now/
Pacific Coast Architecture Database (PCAD), University of Washington. "Sears, Roebuck and Company, Catalog Distribution Center and Department Store, SODO, Seattle, WA." https://pcad.lib.washington.edu/building/8900/
Thank you for reading Sunday Supply Chain Stories!