Networks of Architecture: Keedoozle and Walmart

March 7, 2011

Architect Jesse LeCavalier explores two accounts of architecture understood and deployed as a network of connected buildings in pursuit of a territorial agenda. Looking at these two examples together helps to better understand the possibilities of architecture to operate in unison, beyond the envelope of any single building.


What follows are two accounts of architecture understood and deployed as a network of connected buildings in pursuit of a territorial agenda. Looking at these two examples together helps to better understand the possibilities of architecture to operate in unison, beyond the envelope of any single building. The first concerns an experiment in automated retail by Clarence Saunders, the founder of Piggly Wiggly and inventor of self-service (he holds the patent). Though the project never achieved the ubiquity Saunders envisioned, it is an early example of architecture acting territorially and prefigures the contemporary geopolitics of discount retail. These are most evident in the operations of Walmart Stores, Inc., a discount retailer based in Bentonville, Arkansas. Tracking the process by which the company came to have stores in Vermont shows how Walmart carefully deploys its architecture to override certain restrictions and effectively redraw political boundaries.

Mas issue networks of architecture 01

Exterior of the new Keedoozle in Memphis, Tennessee. © Paul Robert Alvey Family.

Clarence Saunders was convinced that his third attempt at getting rich was fail-safe. On and off for the last twenty years of his life, he worked on developing a project for an automatic shopping environment that he called Keedoozle. The name was always a big thing for Saunders, having founded the Piggly Wiggly grocery chain as well as the “Clarence Saunders: Sole Owner of My Name” stores. Keedoozle is derived from a contraction and modification of the phrase “Key Does All,” in reference to the device customers used to select their merchandise. Saunders was convinced that having a sample of each item for sale behind glass would increase efficiency, increase hygiene, and reduce theft. Patrons (or patronesses, as was more likely the case in the 1930s) would survey the available goods and then register their selection by attaching the key—often described as a cross between a camera and a gun—and pulling a lever that would in turn perforate a strip of paper. That strip, after accumulating a list of all the desired items, would then be brought to a cashier where it would be fed into a “translator” that would in turn trigger a series of belts and trip switches to automatically direct items into a tote that would momentarily and miraculously appear through a small door next to the cashier. The registration process of the items also tabulated their cost, thereby eliminating the time needed to checkout. Customers would pay the total while their items were being retrieved and would then be on their way. This format is different than an “automat” because in the latter, items are purchased one at a time and the item on display is the one the customer will buy. In the early Piggly Wiggly stores, customers acquired the labor-role of the clerk through the process of serving themselves. In the Keedoozle stores, customers took on the role of the inventory manager by effectively notifying the store of what merchandise was being depleted and what needed to be restocked.1 The automatic stores themselves functioned like miniature distribution centers with automated conveyors and “trip shelves” that would allow workers in the storeroom to “pick” the orders placed by the customers and their keys. An examination of the plan shows the division of display room and stock room, but is effectively two buildings in one – a showroom open to the public and a small-scale automated distribution center.

Mas issue networks of architecture 02

Illustration of the interior of a Keedoozle store. From “Saunders Ready to Go Again with a New Store: Keedoozle System Open in March at Union Extended and Poplar”, Clark Porteus, Memphis Press Scimitar, January 23, 1948. © Mississippi Valley Collection, University of Memphis Library.

Mas issue networks of architecture 03

Plan illustration from Sunder’s patent application for an automatic store. © United States Patent and Trademark Office.

Accounts of Saunders’s new venture in popular media did their best to promote it, even if they acknowledged the complications of its baroque mechanism. Saunders was a bit of a local hero and, though eccentric and stubborn, it seems the community was pulling for him. In spite of such popular support, the technology of Keedoozle that was meant to vault shopping into the future proved erratic and costly. While Saunders managed to open three of the stores and had ambitious plans for expanding his empire of automatic stores, the system never developed as he hoped.2 However, in Keedoozle we can locate several tendencies and aspirations that remain legible in the contemporary geopolitics of discount retail, including an early form of a retail prototype and a system that prefigures contemporary automated distribution systems.3

The Keedoozle model was presented as a deployable prototype, infinitely extendable along one axis, 20 feet at a time. Saunders and his contractors evidently saw the capacity for these new projects to occupy in-fill sites. Keedoozle was designed as a format: deployable, extendable, and refined through wartime use. According to the contractor hired to build the first store in Memphis:

The all-steel buildings come in 20-foot sections, and we have eight different widths…The Keedoozle will be 40 x 160 feet, and there won’t be a piece of wood in it. It’s been improved since the war. It’s fireproof. We recommend concrete floors. All we need is to get vacant lots… These buildings standardize the stores and they will be exactly alike wherever they are put up. This means all the electric connections and plumbing will be alike.4

Significant in this description of the consistent formatting is not that the buildings are of the same design, but that all the connections are standardized. The standard Keedoozle format is really a matter of infrastructural connection and signage and the stores themselves were a combination of engineering and marketing. The focus on concrete floors further reinforces this attitude. Rather than foregrounding the shape of the buildings or their vertical surfaces, the prototypes are described in terms of the required formatting processes necessary to enable their repeated implementation.

The deployability of the stores was part of Saunders’s specific expansion agenda. Having already gone through two multi-million-dollar cycles of boom and bust—the first with Piggly Wiggly and the second with Clarence Saunders: Sole Owner of My Name—he put a great deal of energy into the Keedoozle project in the hopes that it would succeed again. Through a gutsy maneuver to corner the Piggly Wiggly shares, Saunders risked much of his fortune and lost almost all of it. He insists that the Board of Trade changed the rules on him—an outsider to the New York financial world — and had they not done so, he would have become incredibly wealthy. Saunders did not get over this easily and his development of a “fleet” of Keedoozle formats in order to mount an “invasion” of the Northeast seems as much expansion strategy as revenge opportunity.5 The ambitions of the company demonstrate Saunders’s optimism in his new enterprise. His new store format would allow him to expand his territory in a dramatic fashion: 40 stores a year over five years for a total of 200, up from 2 in 1949. Saunders had an earlier version of Keedoozle that he tried to get off the ground in 1938, but various technical difficulties made it far to expensive too keep open.

Mas issue networks of architecture 04

Saunders describes Figure 45 from his patent application for an automatic store as a “geographic layout, illustrating a proposed method of use of the invention.” © United States Patent and Trademark Office.

In Figure 45 of U.S. Patent 2,661,682, Saunders includes a drawing of a generic territorial condition that suggests he was thinking beyond the prototype put forth in the plan. In the center of the drawing is a small square at the intersection of two roads and a railroad. The square is labeled “Distributing Store and Display Room” and is surrounded by a larger square simply labeled “City.” The roads leading from the central intersection lead to other smaller squares labeled only “Display Room” and each is in a different context: one is isolated along the arterial, one is surrounded by an eroded ellipse designated as “Village,” another by a rounded square also called “Village,” and a fourth set in a square with sharp corners labeled “Town.” Saunders’s proposal here is that the showroom and distribution hub would act at a territorial level—customers would visit the store, use their key to punch in their order, and place it with the cashier. Their request would automatically be for warded to the local “Distributing Store” to be consolidated with other orders and delivered to the store from which the order was placed. The communication between the display rooms in the local network would ensure that the distribution hub was kept well stocked. This local hub-and-spoke system suggests a much larger interwoven network of distribution centers and display rooms spread across the country. It also anticipates the model of territorial control adopted by Walmart.

Saunders’s patrons saw his price policies as political acts of resistance against an overbearing set of regulatory policies. While the low prices were an affront to the business community, they were a boon for his customers. According to the chairman of the Memphis Consumers’ Advisory Committee, “Members of the committee and other housewives of Memphis highly approve of any methods that will bring down prices of their groceries…. The new Keedoozle system seems to do this. To us, Mr. Saunders’s efforts seem more truly American than any blanket law which protects business from losses caused by its own inefficiencies.”6 In this case, it is not the business model per se that is being praised, but rather the store itself. Automation was not just a way to save money on groceries; it was a path to freedom. While Keedoozle never flourished the way Saunders hoped, the ideas that it materialized have maintained some currency. Most relevant to the discussion here is the building format itself and its implied geo-political aspirations apparent in Saunders’s plans for an aggressive invasion of northern states. The Keedoozle stores were highly mechanical and semi-automated. They were also designed to be identical and to then be relentlessly deployed. A similar use of architecture to invade hostile territory is evident in Walmart’s process of building stores in and around the state of Vermont. Vermont was the last of the United States to have a Walmart store within its borders. Largely because of local conviction that the retailer’s presence in the state would increase traffic, threaten local businesses, and encourage diffuse suburban development, opponents waged a tenacious policy and media campaign that kept the company out of the state for several years. This struggle between the small state and the large corporation was seized upon by the news media whose coverage of the conflict consistently painted it in the colors of war by using headlines like: “Battle of Vermont: Walmart Plots its Assault on Last Unconquered State”; “Walmart Lost Battles, Won the War: Vermont Store Opens”; “Waging War on Walmart”; etc.7 More than journalistic histrionics, the use of such analogies and metaphors illuminate the military approaches adopted by both sides in pursuit of their aims. In spite of resilient opposition, Walmart continued its high-profile policy-based efforts to gain purchase in Vermont. At the same time, the company proceeded to systematically build a physical line of stores along the Vermont border. This blockade of retail outlets proved to be more potent than policy negotiations because it effectively saturated the market without ever entering it. By the time Walmart was allowed access to the state, the real battle had already been won.

Faced with Walmart’s imminent arrival, concerned citizens, flatlanders, “Ecotopians,” and even the Vermont government mobilized their resources to prevent the company’s entry into the state.8 Most of the usual approaches were adopted, including petitions, demonstrations, and the strict enforcement of design guidelines. However, in the case of Vermont, other more inventive measures were taken. For example, in an effort to raise awareness of the situation, The National Trust for Historic Preservation—a private non-profit organization dedicated to the preservation of historic places—included the entire state in its annual list of “11 Most Endangered Places” in both 1993 and 2004.9 Though this inclusion has no immediate policy impact, it nonetheless holds significant sway over public opinion. At the governmental level, Howard Dean, former presidential primary candidate and the governor at the time, flew to Arkansas to meet with David Glass, the CEO of Walmart. According to Dean, “We had a good meeting. I don’t think they’d had many governors come to meet with them. I wanted them to understand that we’re not against Walmart, but that we’re just against suburban sprawl… They agreed to consider downtown locations in the future.”10 As if seeking to broker peace with a hostile invader, Dean’s role as ambassador is significant because it implicitly elevates the status of Walmart beyond that of a mere retail operation. The governor’s focus on property and territory is also revealing. It asserts that the state has less opposition to Walmart as a retail enterprise, but instead takes issue with its choice of sites, which suggests that the conflict has less to do with ideology or aesthetics than with simple location.

Spatial concerns have been a significant aspect of Walmart’s approach, as it has consistently relied on a territorial strategy to expand its operations. As the company originated in rural areas serving a dispersed clientele, it adopted a procedure of peripheral market saturation. According to Walton, “We figured we had to build our stores so that our distribution centers, or warehouses, could take care of them, but also so those stores could be controlled… each store had to be within a day’s drive of a distribution center.”11 A claim like this supports an understanding of Walmart’s operations as a dynamic totality rather than as a collection of isolated retail locations and is significant because it helps illuminate how highly calculated their operation is. Walton went on to write, “We never planned on actually going into the cities. What we did instead was build our stores in a ring around a city.”12 This statement is supported by a 2006 study that found 49 percent of Walmart locations are within 500 meters of a city boundary, and 18 percent of stores are within 100 meters.13 This same geographical precision of property acquisition played no small role in Walmart’s efforts to enter the Vermont market.

In response to the intense opposition from within Vermont, Walmart adopted an aggressive siege strategy and proceeded to systematically surround the state with outlets in an attempt to lure its inaccessible target market across the borders into New York, Massachusetts, or sales-tax-free New Hampshire. One reporter even suggested that Walmart was building a “Maginot Line of four open or soon-to-open stores along the state’s border.”14 If Walmart could not enter Vermont, it would get as close as possible and distribute its locations to ensure saturated border coverage. There are seven Walmart locations within 5 miles from the border (two are even less than 2,000 feet away) and another six in a slightly larger ring around the state.15 Taking a standard 20-mile radius as an index of coverage, The Vermont border is effectively sealed by Walmart stores. If one of the stakes in Vermont’s “battle” against Walmart is a kind of authentic “Vermont-ness,” then Walmart’s spatial tactics would, according to its opponents, threaten this quality. By encircling the state with precisely targeted retail locations, Walmart effectively acquired the market territory it was pursuing without entering Vermont itself. The state border that served as a political boundary is trumped by the “catchment areas” of the store locations and their strategic constellation effectively inscribes a new kind of elastic border within and around Vermont. Faced with the increasing migration of its tax-base, the state eventually agreed to allow Walmart entry into its domain.

The four stores that were allowed to be built—in Williston, Berlin, Rutland, and Bennington—are themselves variations on typical Walmart formats, if only in small degrees. All four are located in towns at crossings of significant state roads or interstates to avail themselves to as much consumer traffic as possible. In Rutland, Walmart seems to have taken Dean’s request seriously and opened at one end of a shopping center in the center of the town and as part of a city revitalization project. This location is promoted in the company’s 1997 Annual Report as evidence of its interest in maintaining healthy and vital small towns. The company’s record might suggest that its interests are elsewhere, but this particular arrangement is an exception to their standard model of growth.

A last piece of evidence in support of the company’s precise location process is the store in Hinsdale, New Hampshire. A supercenter (not permitted in Vermont), the store is located almost immediately over the border—any closer and it would actually be in the river that separates the two states.

Mas issue networks of architecture 05

Walmart in Hinsdale, New Hampshire. © Jesse LeCavalier / Google Earth.

The deliberate and systematic discipline required to execute Walmart’s space-based takeover strategy is significant because it implicates architecture in practices of power and control. The single retail unit, when understood collectively, becomes an insidious territorial instrument capable of securing space within an established logic of market control. Architecture can be understood here not as a system of isolated buildings, but instead as an interlinked network united in a common purpose. In this sense, rather than developing the symbolic content of its buildings, Walmart emphasizes their symbolic presence and in doing so, asserts the importance of architecture within a territorial practice. By using buildings to create and enforce its own policy, one that eclipses federal divisions and desires, Walmart demonstrates that architecture is not the result of politics; it is politics.

Mas issue networks of architecture 06

Walmart locations in Vermont and its vicinity. © Jesse LeCavalier.

1 According to business historian David Magee, the Piggly Wiggly method of organizing its inventory was highly influential on executives from the Japanese automobile manufacturer Toyota. After a visit to a local Ford plant, the delegation visited a Piggly Wiggly store and were influenced by the manner in which inventory was restocked automatically, as it was removed from the shelves, and how merchandise was replaced only as customers depleted it. Magee asserts that this system was very influential on Toyota’s “pull” approach to material in which products are developed “just in time.” Source: David Magee, How Toyota Became Number One (New York: Penguin, 2007) 36-37.
2 “Keedoozle’s Robots Fail, Men Get Jobs: Saunders Throws Mechanical Men on Junk Heap, Clerks Will Work Behind the Shelves,” Jack Bryan, Memphis Press-Scimitar, February 02, 1938.
3 Woody Forbes, former mechanical engineer for Keedoozle, recounts in an interview, “I doubt seriously if it would ever work in the grocery business… I’ve thought of other things he might have made with that unit so that it could distribute … a kind of warehouse type thing.” Interview of Woody Forbes by Mike Freeman, January 04, 1984. Mississippi Valley Collection, University of Memphis, 438-03-16.
4 Howard Sowell quoted in “Saunders Ready to Go Again with a New Store: Keedoozle System Open in March at Union Extended and Poplar,” Clark Porteus, Memphis Press Scimitar, January 23, 1948.
5 According to a article written in 1949, “The New York Times tells of the ‘invasion’ of New York, northern New Jersey and the New England states by Keedoozle. Keedoozle Eastern Corporation has been formed in the east as an affiliate of Saunders’ parent company, Automatic Systems Corporation, in Memphis. Franchises to some five stores will be issued this year in New York City. The New York Times says the Eastern organization plans to have 100 stores in New York City and another 100 in the rest of the territory by 1954, but Mr. Saunders said he hoped the system would spread much more rapidly in the East.” From “Keedoozle Stores Spreading Out: Corporation in East Plans 100 Units,”Memphis Press-Scimitar, January 20, 1949.
6 Quoted in “Thousands Get Free Food: Keedoozle Against Tennessee Law,” Memphis Press-Scimitar, September 24, 1948.
7 These articles include Frederic M. Biddle, “Battle of Vermont: Walmart Plots Its Assault on Last Unconquered State,” in Boston Globe, July 18, 1993; Malcolm Gladwell, “Walmart Encounters a Wall of Resistant in Vermont,” in The Washington Post, July 27, 1994; John Greenwald, “Up Against the Walmart,” in Time, August 22, 1994: Ross Sneyd, “Walmart Lost Battles, Won the War: Vermont Store Opens,” in St. Paul Pioneer Press, September 20, 1995; Pam Belluck, “Preservationists Call Vermont Endangered, by Walmart,” in The New York Times, May 25, 2004; George F. Will, “Waging War on Walmart, “ in Newsweek, July 05, 2004; and Alex Beam, “Walmart and the Battle of Vermont” in The Boston Globe, September 12, 2007.
8 “Flatlanders” is a term used by native Vermonters to describe outsiders who have moved to the state. “New Ecotopians” is a term established by the marketing firm Claritas to describe the demographic group made up of “consumers with above-average education who are technology-oriented and civically active. They are more likely than other Americans to make bread from scratch, drive a Jeep, watch the Learning Channel and read Outdoor Life and American Health (Source: Malcolm Gladwell, “Walmart Encounters a Wall of Resistant in Vermont,” in The Washington Post, July 27, 1994).
9 The other entries for the 2004 list of “11 Most Endangered Places” include: 2 Columbus Circle, New York; Bethlehem Steel Plant, Pennsylvania; Elkmont Historic District, Tennessee; George Kraigher House, Texas; Gullah/Geechee Coast, South Carolina; Historic Cook County Hospital, Illinois; Madison-Lenox Hotel, Michigan; Nine Mile Canyon, Utah; and Ridgewood Ranch, Home of Seabiscuit, California; and Tobacco Barns of Southern Maryland, Maryland. The 1993 list also includes the following: Brandy Station Battlefield, Virginia; Downtown New Orleans, Louisiana; Eight Historic Dallas Neighborhoods, Texas; Prehistoric Serpent Mound, Ohio; Schooner C.A. Thayer, California; South Pasadena/El Sereno, California; Sweetgrass Hills, Montana; Thomas Edison’s Invention Factory, New Jersey; Town of Ste. Genevieve, Missouri; and Virginia City, Montana. (Source:
10 Sally Johnson, “Vermonters are up against the Walmart – effort to stop retail chain from entering Vermont,” in Insight on the News, Jan 10, 1994.
11 Sam Walton and John Huey, Sam Walton, Made in America: My Story (New York: Bantam, 1991), 140-141.
12 Ibid, 141 (emphasis added).
13 Matthew Zook and Mark Graham, “Wal-Mart Nation: Mapping the Reach of a Retail Colossus,” in Wal-Mart World: The World’s Biggest Corporation in the Global Economy, ed. Stanley Brunn (London: Routledge, 2006) 23.
14 Frederic M. Biddle, “Battle of Vermont: Walmart Plots Its Assault on Last Unconquered State,” inBoston Globe, July 18, 1993. Perhaps it is worth noting that the comparison, however evocative, is misleading because the Maginot line of bunkers and fortifications was designed to serve protective and preventative purposes.
15 Though there are Walmarts in Canada, there are currently no locations within 20 miles of the Vermont border.