The Roller Coaster: How to Go from Global to Local and Back again—the Case of a Walking Drive Model in Paris

How To Position And Present Ethnography In A World That Doesn’t Know It

February 2019, Paris, France. The Director of Innovation of the world’s second largest food retail group (€2,1 million sales annually, with 12,300 stores in 30 countries) questioned us about the value of the online retail pick-up points they’d started experimenting with in urban areas. The walking-drive model is simple: a counter, not located in a hypermarket, where online orders can be retrieved by customers. Their idea was simple: to improve this model, we needed to start from the needs and expectations of urban customers in terms of food shopping. And since not all urban customers have the same needs, the “walking-drives” should meet local expectations.

Historically and culturally, however, nothing predisposes this group to reason in this way. As a central player in the food industry since the 1950s, the company is based on this triptych: mass production, thanks to a food industry with productivist breeding…for masses, i.e. international “markets” where large-meshed typologies are addressed (“young”, “old”, “rich”, “poor”)…with massive means (giant infrastructures, organizational standardization, macro indicators—and the eternal myth of food abundance).

This logic culminated in the invention of hypermarkets in the 1960s. A symbol of mass distribution in France, this model has since been the subject of much criticism. More precisely, this all-scalable logic is regularly accused of “killing the local”. By reproducing “miniature towns” inside their shopping galleries—which are set up at the entrance to towns—hypermarkets are said to contribute to the closure of small shopkeepers (grocers, butchers, greengrocers, hairdressers, etc.). With online shopping, they would also capture part of the social life of the town centres that is based on commercial exchange. Shared and relayed by many local elected officials, this fear gradually reached the scale of a public order problem, to the point of triggering intervention by the French State in the most affected localities¹.

More recent but also more confidential, another criticism has been made by French anthropologist Marc Augé (1992) with his concept of non-places. A non-place is a space with no history, no identity, and no social relations. In other words, it is an interchangeable space in which the individuals who use it remain anonymous bystanders. According to the author, these spaces proliferate, while “folklorizing” local identities: they include airports, train stations, department stores and, of course, hypermarkets. Shopping galleries are a good example. Hypermarkets have recreated the markets of the city centres by removing its disorder (the auction, price negotiation, smells, unruly crowds, winding alleys, etc.), by imposing standards (storefronts must all be the same size for example). They however kept the city centre’s promise of localism, “authenticity” and proximity. But, in the end, every shopping gallery is the same, wherever they are.

To sum up, on the one hand, hypermarkets are accused of siphoning off that part of the social life of city centres that is based on commerce. On the other hand, they are suspected of weakening the part of socialization that is based on space. Wherever they settle, they contribute to the desertification of city centres, by proposing as an alternative trade without the social relations and local histories attached to it. With their scale logic, they install clones in the strongholds of particularism. Scaling, as Anna Lowenhaupt Tsing (2015) writes, would therefore be to “eliminate diversity”?

Here’s the set for our ethnographic work. So the position our client suggested seemed relatively new because he took these criticisms seriously. Nevertheless, his motivations were more prosaic: at the beginning of 2019, after one year of launch, the results of the walking-drive (subsequently referred to as the Drive) were mixed. Although the Drive had attracted customers, attendance and average basket size were perceived as insufficient. Numbers didn’t live up to expectations. To solve this issue, the Director of Innovation took the way the group had built itself in reverse: starting from the local characteristics of a place to propose an alternative food range or to build new non-food service offers; and not building service offers and then putting them in a place, whatever its characteristics.

Following a call for proposal, the food retailer’s innovation team selected _unknowns because the methodology we created allowed us to study a specific neighbourhood and made it possible to scale the results produced to other neighbourhoods. Our client wanted us to create 2 monographs. The first one was to be done in the Parmentier in the 11^th district located in the east of Paris because this is where the new Drive was located. The second one was in a perimeter around a store located in Villeurbanne, a city in the middle of eastern France, connected to Lyon, because a second Drive could be deployed there. We mainly describe the survey conducted in Paris.

The Research Issue: Typical Cases x Similar Urban Characteristics = Scalability?

With socio-economist Max Weber’s teachings (1986), we have become accustomed to characterizing the city in a very different way from what scalability implies. What is striking in European cities, and arguably elsewhere, is their cultural, political, legal, and economic specificities. The same is true at a lower level—what characterizes neighbourhoods is a topography (a hill, a river), a dominant function (recreational, residential, intellectual, economic), an architectural style, a historic event (La Bastille in Paris), an emblematic personality (Authier, 2006). And above all, a subgroup of the general population (student, executive, couple, retired, etc.) with specific needs.

We had to understand those specificities to analyse their impact on the “food races”. In order to leave the high & macro scales and go down to “human height”, the ethnographic approach, by the concern it brings to the description of details, was our best ally. However, we also needed to reintegrate our teachings into a scalability scheme. The 1:1 scale of the ethnographic study suddenly seemed too narrow. In other words, we had to find a way to take the service offer imagined on the basis of a neighbourhood and duplicate it in another neighbourhood; potentially in a different city. But how could we scale a walking-drive built on hyper-local singularities? How could we scale the “non scalable”? We had to find a unit of analysis with a better potential for generalization. Figure 1 shows the generalized framework we built and our results.

Figure 1. Generalized Scale Framework ©_unknowns

Let’s explain each level.

The Macro Level

To begin with, we completely changed our focus: like an entomologist looking at an insect with a magnifying glass, we went up a notch to get a view from above, while diving into the details.

This leads to questions such as:

• Who are the inhabitants?
• What are their professions?
• What degree do they hold?
• How much do they earn on average?
• etc.

To get this macro point of view this, we identified and then examined statistical and cartographic studies from the national census, data from the Ateliers Parisiens d’urbanisme (APUR) and data on the evolution of prices per m² from the Notaires de France. This enabled us to discover, it was then a presupposition, that within the 11^th arrondissement, the population did not have homogeneous characteristics, and that consequently, subgroups were distributed differently in space. For example, on the map in Figure 2, published in 2012 by the National Institute of Statistics and Economic Study (INSEE), we see that middle class people (in orange) live side by side with executives (in yellow) along with people having several social backgrounds (in green; i.e. students, young graduates, migrant workers, recipients dependent on social benefits, intellectuals, etc.).

Figure 2. Neighbourhood cartography. Adapted from ©INSEE

However, we lacked a framework for interpreting these data. Urban sociology offered us the concept of “gentrification” which we will see later. Intuitively, we knew that this concept made it possible to read the city not only as pure physical data, flows in a topography, but as the “projection on the ground of social relationships” as French geographer Henri Lefebvre once said (1968). In short, the neighbourhood we were to study could be read as an arena where two sub-groups—the gentrified and the gentrifiers—were struggling to appropriate a neighbourhood; theirs in this case. One of our hypotheses was that food stores played a role in this struggle.

The Meso Level

Like town planners who wonder how people move around according to the transport infrastructure, we asked ourselves how people shop according to the urban characteristics of the neighbourhood where they live.

This leads to questions such as:

• Do the buildings have stairwells wide enough to carry several bags of groceries?
• Are there any level breaks on the roadway?
• How easy is it to drag a shopping cart down a crowded street?

There was also a need to find a way to observe eating routines, such as lunch breaks for employees, and to track their deployment in space—for example, the use of a park. The objective was therefore to uncover the encounter between urban space and shopping practices. This is what we called configurations, or the meso level.

But how do we capture it? In the beginning, the temptation was great to want to “observe everything”. But this dream of ubiquity, already illusory in the context of a restricted observation perimeter (a schoolyard for example), became impossible on the scale of a neighbourhood. In the manner of ornithologists who want to observe the passage of migratory birds, we did some spotting to identify observation posts. School outings, parks, subway exits, and pedestrian walkways seemed to be the most promising places in terms of feeding routines. In order to increase the hourly scope of the observations, we also took accommodation on site for the duration of the study.

Resources Under Pressure

Looking at INSEE’s statistical data, one of the first things that struck us was the extreme density of this neighbourhood. In 2010, 44,744 inhabitants lived there per km². In other words, there are twice as many inhabitants in this neighbourhood as in the rest of Paris on average (21,200). By way of comparison, there are 7,100 inhabitants in New York City per km² on average in the same year.

But after all, why is this a problem? It’s a problem because it means that demographic pressure is putting pressure on the resources located in this territory—the foreground space. Hence, for example, policies to de-densify the territory, as shown in Figure 3 with what urban planners call a “green tooth”, i.e. public gardens installed between two buildings.

Figure 3. Installation of a public garden on Voltaire Boulevard, in front of Saint-Ambroise Church. ©_unknowns

Said in less policed terms, it means that residents are competing for the space and facilities there: housing, green spaces, parking, and of course food stores. For example, one need only look at the neighbourhood’s public library to observe traces of saturation; such as this calendar posted at the entrance to a public library that informs about usage levels (see Figure 4). By walking around the shelves of this same library, one understands something else: to avoid being deprived of available resources, some residents bypass the commonly established rules for sharing these resources. They do this in order to capture resources before they are captured by others—in this case cultural goods such as DVDs as shown in Figure 5.

Figure 4. Poster representing the peak hours of the Parmentier media library. ©_unknowns

Figure 5. Media library of Parmentier, on level -1, in the poetry corner, a little isolated from the main aisles. The sign tells patrons not to hide the DVDs “behind the books.” ©_unknowns

Another characteristic of the neighbourhood is the relative difficulty of getting around. Coupled with the high density, the narrowness of the sidewalks as well as the level breaks make the transport of shopping a real ordeal. So much so, that when faced with a commodity, some inhabitants estimate the effort required to bring their goods home before deciding that when on foot, they will be selective about what they purchase. In the end, the answer to this question has of course an influence on the choice of store. But it also gives rise to tactics to reduce the drudgery—such as the interviewee who buys heavy goods only in the grocery store downstairs. Others divide up the carrying work, such as the interviewee who asks her neighbour to go to the store with her to help carry water bottles home. From this perspective, helping individuals with their errands means reducing a constraint produced by the meeting of demographics (high density) and urban characteristics (narrow sidewalk, level breaks).

Thinking Customer Segmentation Through Gentrification.

Up to now, we have talked about competition from residents without really specifying the identity of the protagonists. Who are these inhabitants? How do they form sub-groups? And above all, do they have different needs in terms of food shopping? In order to find the most differentiating marker possible, we used the concept of gentrification, or embourgeoisement in French. This notion comes from the Anglo-Saxon geography of the 1960s. To my knowledge, it was the sociologist Ruth Glass (1962) who first used gentrification to describe, in the neighbourhoods of Notting Hill and Islington, the transition from a working-class population to a more affluent population, the gentry.

The lens of gentrification allows us to sociologize our analysis a little more. It can now be hypothesized that competition for resources is a social competition, bringing together social groups that do not have the same characteristics and therefore may not always have the same interests.

Statistics about changing professional classes also established that gentrification occurred in the neighbourhood. If we look at them, we learn that between 1954 and 2010, the share of Executives and Senior Intellectual Professions, Business Leaders, and Intermediate Professions increased by 45 points, from 28.6% to 73.8%. This is exactly the number of points lost by the share of blue-collar and white-collar workers over the same period: from 71.4% in 1954, to 2010 representing 26.2% of the population of the 11^th district. This inversion continues today. If we compare only workers and executives and higher intellectual professions, we can see that between 2010 and 2015 the share of the former is decreasing (from 5.1% to 4.3%), while the share of the latter is increasing (from 30.3% to 32.5%). In short, managers are the majority in the district and blue-collar workers are the minority professional class.

If we take it down a notch further, at CSP level 2 which are trades professions, this means that garage owners, masons, craftsmen, cobblers, upholsterers, printers, and metalworkers have gradually given way to artists, production managers, association leaders, theatre company administrators, nurses and secondary school teachers. This was what we could call the first wave of gentrification. In a second phase, senior executives arrived in the neighbourhood: they were more likely to be professionals (lawyers, doctors, company directors) or private sector executives (consultants, senior managers, etc.). In 2015, higher education graduates represent 2/3 of the population (61.5%). These sociological changes are modifying the supply of catering and food consumption in the district. Figure 6 shows an organic grocery store on avenue Parmentier which replaces a low-cost Franprix market.

Figure 6. Bio c’Bon organic grocery store near the Drive on avenue Parmentier. ©_unknowns

If we zoom in a little more, this time at the individual level, we can see that these changes are assessed differently. On the one hand, gentrified people castigate these changes because they see small traders disappearing in favour of restaurants. Figure 7 exemplifies a gentrified space that’s opened in the neighbourhood. A resident in Voltaire who is a receptionist at the Maison des Associations describes this change:

“We used to have a new food trader [greengrocer, butcher, fishmonger] every week—[She turns around and shows me the shops in front of Maurice Gardette Square]—now we have no more shops. Now it’s just restaurants, look: 1, 2, 3, 4, 5. We’ve got more than that. And rue Saint-Maur is just that. Only bars and restaurants.”

Figure 7. Cocktail bar and open space La Popina at rue Saint-Maur. Inside, a white man of about 40 years old in a white V-neck t-shirt consults his iPhone. A Mac decorated with a sticker is placed in front of him. ©_unknowns

Clothing wholesalers crystallize the opposition. For the poorest gentrified, they are an opportunity to buy affordable clothes—that is to say, to control their spending. As a saleswoman in a jewellery shop explains, “It’s tempting because there are some interesting items, eh? I’ve tried, but no, no, no, we don’t buy retail. But they have some nice stuff.” On the other hand, for the gentrifiers, these wholesalers have to close down to make way for shops more in line with their taste, that is to say with their social position. As the director of a business school in the 12^th arrondissement explains:

“On Boulevard Voltaire, all the Chinese wholesalers are leaving. I hope I don’t have to tell you this, but they are being replaced by shops. We are very curious to know who is moving in. […] In fact the Marais, finally the transformation of Beaumarchais must come to Voltaire. [What shops do you like in Beaumarchais?] It’s clothes shops, it’s APC, the Blend restaurant […] all the brands we like Maje, Bonpoint for children.”

This was the first interview of the study and it seemed to us emblematic of the more global process of modification of the sociological composition of the neighbourhood. In the end, it could have been called “extraordinary gentrification calculation” because our business school-educated director finally had a winning speech: not only did he want to take advantage of the effects of gentrification, but also to multiply them.

Obviously, the installation of these new populations is not without opposition. By settling, the newcomers also install new rules: what can be allowed in the neighbourhood or what is no longer possible. These new rules are sometimes denounced by the former inhabitants, who feel like they are “dispossessed” of their former stronghold. An association leader described her outrage that the bourgeois call the police when young people play football outside the hours set by the town hall: “[And the population, you’ve seen it change in recent years…] But yes, even in the square, Maurice Gardette, there are obnoxious people. They are the bourgeois who want order.”

But what do food races have to do with it? In fact, the establishment of this type of business may give the gentrifiers hope of attracting their fellow people, i.e. other executives, other engineers, or other lawyers. And thus, strengthen their presence in the neighbourhood by multiplying the small bastions in order to occupy the space. Here they will be able to live out their social status.

In the long term, it is a question of increasing the added value of their residence when they sell it. In other words, from this point of view, gentrifiers have every interest in ousting businesses that do not inspire confidence among future buyers who look at the type of store in a neighbourhood to decide whether or not to invest there. This is an indication of the progress of the gentrification front. From this point of view, an organic store is a favourable index; a discount store is an unfavourable index.

But this eviction should not be total. The geographer Anne Clerval (2013) points out that newcomers to the working-class districts of Paris also need to stage their anchoring in the neighbourhood they have newly moved into.

“The frequenting of small shops gives the gentrifiers the impression that they are participating in the sociability of the neighbourhood (164) […]” despite the social and cultural differences that they import there. In other words, it is a means of capturing symbolic profits, those offered by the reputation of being “open-minded.”

On the nice side, however, they fear that they can no longer afford to live in their neighbourhood because the price of housing and various goods is rising faster than their wages. Staying in the 11^th arrondissement therefore forces them to invent different schemes. As far as food shopping is concerned, we met gentrified people who simply stopped shopping in the neighbourhood. They now have to get their supplies elsewhere in Paris, i.e. where gentrification has not yet arrived:

“The last purchases I made on special offer was dishwashing liquid; in normal times it’s between 1.60€ and 1.80€ and on offer it’s 3 for 3€—so if there are bargains in Auchan [in Bagnolet], often it’s on Wednesdays, I go there…I also look on the Internet every Sunday, I look at all the signs. That’s how I do my shopping.”

If we think of food shopping sessions as “acts”, then we can say that the gentrified are gradually becoming deprived of a means of asserting their belonging to their neighbourhood every time they give up.

From A Business Point Of View

The adaptations observed in the typical cases were problems to be solved in people’s lives—the basis of the future Drive offer. It was seen that some properties were difficult to access; because of the number of people who wanted them. What’s the big deal? What can a food distributor do with this information? In fact, it means that one of the possible ways to expand the Drive’s offer would be to offer the goods that are locally the most under pressure. It could be workspace, the very one that’s taken over at the library; or, to use our example of cultural goods, books and DVDs. Thus, while generating additional traffic for the Drive, it would make the resources that are most in demand at the neighbourhood level a little less scarce. To paraphrase Hobbes, it is a competition of all against all that the new offer could help to “relax” and thus be a solution to the problem of hyperdensity.

We also saw that the transport of groceries was difficult. Another solution could have been to offer home delivery services for the heaviest and/or most bulky goods. Another would have been to lend shopping carts, cargo bikes, or even an electric scooter, in exchange, for example, for a subscription to a loyalty card and/or a deposit. Here, the Drive allows to delegate the carrying of shopping or to equip the customers to reduce the drudgery of this task.

Concerning the problem of gentrification, several options were possible. On the gentrified side, we saw for example that, because of the closure of their shops, the poorest people had to go to the outskirts to find affordable prices. To enable them to stay in their neighbourhood, the Drive could have offered a range of “essential” products sold at lower prices than in the new shops. Similarly, services could have been devised to increase their income; for example, by mobilizing their assets. These could have been services to facilitate the seasonal rental of their property, be it the handing over of the keys to the tenant, a cleaning service, home improvement to be carried out, etc.

On the gentrification side, it could have been to move upmarket, with organic products, for example—capable of distinguishing them socially, while affirming their belonging to the neighbourhood. Another manifestation of gentrification could be to install the iconic markers of the neighbourhood’s history (a zinc counter, for example, typical of the former workers’ bars in the neighbourhood). In addition, services facilitating their identification with the neighbourhood could have been imagined; for example, by positioning themselves as trusted third parties between them and different trades, especially craftsmen. A mixed offer could also have been imagined, aimed at both gentrified and gentrifying people, so as to position itself as a place of “cohabitation,” meeting the needs of both groups would be an eating area on the small square facing the Drive.

From A Design Point Of View

Taking those local specificities into account could avoid building “non-places” devoid of any identity, such as hypermarkets, because of the scalability model they are based on that erases local specificities. For example, we could imagine designing a store integrated into the local history of the neighbourhood, or an offer adapted to its main functionality (for example, commuting to Parmentier).

As to the impact on the client itself, we tried to shift

• posture; no longer thinking of its walking-drives only as the end of a supply chain, but as a neighbourhood business, with social issues at stake,
• reading grid; rethinking its customer segmentation in terms of gentrifiers & gentrified, in Paris, and adapting its range according to their needs,
• design method: offers of services thought from a typical cases x urban characteristics grammar), and finally,
• deployment method: if typical cases x urban characteristics of district A = those of district B then we can deploy.

Study Protocol & Analysis

First, not all observation units were comparable. French national statistics take the household as a unit, whereas the interviews conducted had each member as the unit. The former obliterated the distribution of domestic work within the couple (especially shopping) but not the latter.

Second, although the articulation of macro-meso-micro scales makes it possible to make otherwise scattered facts intelligible, it sometimes resembles a cosmogony that is too coherent to be true (De Sardan, 1996)! This raises the question of the place of scale comparison in the research process: is it the end or the beginning of research?

In other words, I think the scaffolding has been effective in making assumptions. It was “heuristic” as sociologists say. But to think that all the scales easily fit together, that you only have to pull the ball of wool to see all the dimensions of a subject—in this case demographic, urban, social, and finally political—is to give too much credit to the idea that the phenomena we observe are all consistent with each other, whatever the scale on which they occur. In any case, the interweaving of these different levels of analysis is as much a datum to be explained as it is an explanatory datum.

Another disadvantage linked this time to the problematization in terms of gentrification. If the advantage is not to be naïve about the social issues behind the census figures, one pitfall is to take the side of the gentrified or the gentrifiers without realizing it. We are tempted to be miserable when we talk about the gentrified, we are tempted to be accusatory when we talk about the gentrifiers. In other words, without being careful, we can pass imperceptibly from a judgment of fact to a judgment of value.

The Implementation Of The Scale Model

The scalability model we have proposed (typical cases x urban characteristics), forces us to work with clones, i.e. similarities in similar contexts. This puts considerable weight on variables from the first monographs (e.g. hyperdensity + gentrification + walking). That said, these variables can be verified fairly quickly—provided, however, that there are agencies that produce these statistics. In countries where such agencies do not exist, it may not be possible to collect this information.

Receipt Of Results At The Client’s Premises

We have presented our results to several directorates. However, the higher we went up the hierarchical levels, the more the macro reference scales resurfaced, with the representativeness of the study as a banner. One of our mistakes was to think that the ambition of the Innovation department was shared by all stakeholders. And that everyone around the table was ready to integrate into their usual client typology, a typology inspired by the Marxist geography of the 1980s. This is probably forgetting that our interlocutors were not politically neutral. Even taking all possible precautions to de-politicize this notion, I am not sure that our client has appropriated it as a tool for description and analysis. In other words, an analysis in terms of gentrification was probably too “radical” to be accepted—at least we should have formulated our ideas differently.

Another issue was the unit of analysis. As a unit of analysis, the neighbourhood seemed too small compared to the units of analysis that were usually used; typically a city or even the region. However, it was the highest macro level of our framework. Even when proposing a scalability model to deploy the new offers, this spin-off logic was too different from the usual model of scale they use; i.e. to duplicate the same offer everywhere with a few adaptations at the margin according to local specificities.

In the end, perhaps everything was too new in this study and we should have said from the very start: “This is the first time they’re going to hear about social science, gentrification, a swarming scale model, etc.” We should have better adapted our discourse to our interlocutors. To do this, we should have consulted more with our innovation interlocutor.

Marc-Antoine Morier – Anthropologist & Strategist at _unknowns, innovation consultancy marcantoine.morier@gmail.com