Decontextualised Data and Socio-Spatial Differences

Abstract

This chapter presents (personalised) data as commodities and demonstrates the related contexts of utilisation. Making concrete the possibilities of recursive reference in the data economy, we return to the physical environment: Using commercially traded data for the cities of Berlin, Munich and Essen, an exemplary and empirical study examines in what manner a data-based logic of utilisation reaches the different districts and their inhabitants and the implications this carries for socialisation and the overall social question.

One reason why it is so difficult to grasp the far-reaching significance of personal data lies in their immateriality. The zeros and ones, to name only the most basic form of their visualisation, convey a harmless abstraction that stands in stark contrast to the influence the translations of this code can have in the real world. What makes the process of data collection and addressing more tangible is the physical visibility of the devices that accomplish this and by the immediate compensation that they offer. A digital voice assistant that can convert commands into information and services, a Skype call that makes the conversation partner immediately visible, or the use of a digital address book are examples of how with the help of technology, data are expressed in useful services.

In the types of recursive acquisition described so far, the relationship between entry and use was still partly visible as both happened through the same medium. In cyberspace, personalised advertising, news or search results sorted by their relevance appear in the same virtual environment into which a large part of the processed data flowed before. Similarly, location-based services, access control via RFID chips or advertising insertion through facial recognition in networked real space are often still recognisable as interacting systems.

By contrast, the possibility of isolating personal data from the technical context of their collection removes any attributability that would imply at least some form of control. As an immaterial commodity, individual attributions “go on a journey” and are then again presented to the user in the form of a service. In the meantime, they are subject to analyses, additions, copies and recombinations undertaken by decentralised management systems, crossing borders and legal jurisdictions. Their fluidity is not immediately apparent in real life. Generally, users do not know what they have lost, nor to whom. The various sources remain as abstract as the different utilisation contexts. Most people notice this surveillance only in the form of personalised advertising and do not consider this cause for great concern. In fact, the utilisation contexts of personal data produce a tight grid that predetermines opportunities for participation and access to resources.

The goal is therefore to concretise the diagnosed mechanisms of recursivity for traded data and to examine the social and space-constituting consequences of their capitalisation. Two sets of questions are of interest here: How does the data-based addressing of the target person or group take place and which attributes come into play in the context of Bourdieu’s field of social positions? And second, what structural influence do the traded data have on space and neighbourhood?

To answer both questions, it was necessary to gain empirical access to the data market. Its offer structure becomes apparent via the portfolios of data traders on the internet. A more in-depth analysis of the available customer data is achieved, by way of example, by extracting an extensive data set comprising the households of major German cities.

5.1 Data as an Unequally Used Commodity

Over the years, the collection, processing and utilisation of data has grown into a market worth billions of euros, with well over 1,000 participating companies in Germany alone (Goldhammer and Wiegand 2017, p. 21). In addition to companies that collect customer data online and offline, a large number of data brokers specialise in the processing of personal data of specific target groups for the credit industry, advertising, insurance, consulting or job placement, among others. The data can be sold directly to interested parties; they can be viewed for a fee or sold as an analysis result abstracted from the data set. In each case, it is once again the increasing transparency of the customer which is rediscovered as a prerequisite for economic success in the age of digitalisation.

The power to create this transparency is thus not necessarily limited to the immediate environment of the data-generating technologies or those who commission a personal address for commercial reasons. For the data trade, the centre of this power is where data acquisition and utilisation are coordinated. With many providers, especially large platforms such as Facebook or Google, the collection of data and their downstream trading lie in one hand.

As the data can be substituted, data traders can derive missing attributes from an existing data set or compensate for them with higher-level target group references. As a consequence, data subjects are addressed with data that were not collected directly through them, allowing for the data stock to be further enhanced and supplemented by characteristics that are essential for a personalised categorisation. At the same time, individual addresses can be substituted by the behaviour of other individuals as long as they belong to the same target group. Above, the term collaborative filtering was used to describe one form of this collective address (“Customers who bought this also like that”). Because of these substitution practices, even people who are very careful about the data they share are addressed more or less directly. Most importantly, the mutual exchange and purchase of data provides data traders with comprehensive portfolios. While they achieve size and the associated economies of scale through exchange, what this cross-border trade means for the data subject is a diffusion of data and total loss of transparency. The specific utilisation of the available data can then have an impact in all areas of life. Individuals are confronted with “ubiquitous data” coming from many different, loosely coupled, and sometimes overlapping sources in an asynchronous and decentralised way (Hotho et al. 2010, p. 62), addressing them individually or collectively, making it impossible for them to gain control over these data. The described interplay of online and offline now has to be understood inter-temporally and across spaces: Anything that has been stored about an individual can be used to address them anywhere, with no transparency whatsoever.

Consequently, there is a strong asymmetry of information between the data subject and the recipient or distributor of personal data. Based on the revenue generated, the trader can accurately estimate the value of the data. For consumers, on the other hand, it is almost impossible to understand in what form and context their data are being traded. To determine the value of their data, consumers would need to know, among other things, at what points and under what conditions the data are monetised, how long they are stored and economically usable, they would need to take into account risks of misuse and their consequences and they would need to be able to assess all indirect derivations from the existing data sets (cf. Wiewiorra 2018, p. 464f). Furthermore, economically harmful effects on different areas of consumers’ lives because of market entry barriers or price discrimination (e.g., loans, insurance) would have to be considered.

Since consumers do not have this information and do not share in the profits generated with their data, economic value creation is distributed very unevenly. Although there is a general idea of the value of individual data in parts of society, so far there are very few paths toward an alternative valorisation of personal data that would bypass the databases of platform operators and traders. Some health insurance providers and insurance companies do offer discounts for data, but they fail to reveal to their clients how much they actually gain.

Most consumers agree to receive nothing more than the access to digital services in exchange for their data—a fact that can be explained with the double novelty of these goods of exchange. While individual data capital as a new currency is something that needs to be discovered and assessed by everyone individually, the data economy also offers specialised services. These are commonplace by now, but at the same time still seem fascinating; to some, they are considered indispensable, and in general they are perceived as an important novelty, their actual value hard to determine. In this perspective, the special opportunity for digital participation more than outweighs the abstract way in which information is disclosed (Acquisti et al. 2013).

In the current system of the data economy, an objective economic profit can only be achieved if the disclosure of data is accompanied by an economic benefit that exceeds the costs incurred. In this context, the opportunities for acquiring capital by gaining social and cultural capital in cyberspace have been examined. The preliminary conclusion is sobering: Even in digitalised real space, profits in the data trade are based on a monetisation of knowledge which is then used by businesses to create individualised offers. While on the supply side data capital can thus be converted into economic capital, on the demand side it grants access to use that can be converted into material profit only to a limited extent. Access mainly leads into mirrored spaces, and this makes a transformation into economic capital more difficult for the segments of the population lacking capital. In other words: Data capital can be monetised very well under recursive effects, which means that the data traders benefit, and users lose out. If, because of the characteristics that haven been described, data are collected without any compensation for the data subject, all the profit generated by the digital business goes only to one side.

The databases (portfolios) of the data traders should reveal this unequal utilisation of digital capital. With the permanent establishment of a data-aggregating system, these portfolios should reflect the returns of many years. At the same time, the associated possibilities of a personalised address reveal the extent to which data trading is the cause for the recursive logic.

5.2 Data Traders and Their Portfolio

In their range of services, data and address traders have different focal points. Their overriding interest lies in comprehensive information about each customer, which allows targeted addressing, increases the probability of conversion, protects companies against default risks, enables cost reductions and impact measurements and, last but not least, allows for the planning of processes and behavioural predictions (Pentland et al. 2021 for an overview). What all these offerings share in common is an extensive data set that, depending on the business partner’s specific needs, makes a customised selection possible.

Large providers on the international market (such as Acxiom, Experian, or CoreLogic) maintain customer profiles, company addresses and e-mail contacts amounting to several million each. Numerous additional features on consumer behaviour, socio-demographics or housing and living situation are systematically processed. They offer precise query options ranging from “likelihood of switching health insurance” to “affinity for bargain shopping” and “travel type” and provide detailed information on sales-relevant categories such as education, income, health, lifestyle and class. Information that can be retrieved immediately includes household income, credit score, social class and corresponding interests. In addition, data on different product fields can be obtained, and in many cases preferences are already pre-sorted.

A look at the extensive portfolios of data traders on the internet also reveals the possibilities for addressing customers (e.g., www.az-direct.com): Options include postal or e-mail address, display marketing campaigns such as banner advertising as well as ads in search engines and social networks. In addition, they offer forms of mobile marketing, where a commercial message appears based on location or weather, for example, via the apps installed on a smart phone. Numerous possibilities of linking online and offline appear here: As part of a CRM onboarding, for example, the intersection between existing customers from the company's own database and the users of various internet platforms can be identified in order to reach them with a targeted digital campaign. This can be supplemented by an offline campaign, such as catalog mailings. The networked real space in turn lends itself to digital out-of-home campaigns, where digital outdoor displays are already linked to various targeting criteria such as local reference, weather or time of day, and segments can be programmatically controlled via addressable TV (cf. www.acxiom.de/).

As expected, all these providers stress the conformity of their services with current data protection regulations. Aside from refraining from the collection of specific types of personal data (such as race or ethnicity, political or religious views, information about gender or sexual orientation) as required by law, data traders refer to consents given or to the legitimate interests of them and their partners and clients. In Europe, regulations are comparatively strict; however, the abundance of data available makes it hard to believe that laws can limit the volume of data significantly. In countries with fewer regulations on data protection, data vendors’ offers surpass the European market:

As early as 2014, the federal consumer protection agency of the United States, the Federal Trade Commission, conducted an in-depth investigation of nine of the country’s largest data trading companies. The results revealed a nearly complete coverage of all U.S. households with diverse labels. The Commission’s summary of the enormous volume, most of which was generated without consumers’ knowledge, includes the following statements: “Of the nine data brokers, one data broker's database has information on 1.4 billion consumer transactions and over 700 billion aggregated data elements; another data broker's database covers one trillion dollars in consumer transactions; and yet another data broker adds three billion new records each month to its databases. Most importantly, data brokers hold a vast array of information on individual consumers. For example, one of the nine data brokers has 3000 data segments for nearly every U.S. consumer” (Federal Trade Commission 2014, IV). The report also revealed an intensive exchange of data between individual traders, which greatly expands the data sets to the benefit of all parties involved. As in the German market, specific groups are derived from all the data obtained, ranging from individual characteristics such as “High-End Shopper”, “Diabetest Interest” or “Home Ownership Status” to specific categorisations such as “Rural Everlasting” (single men or women aged 66 and older with a low educational level and small assets) (cf. Ibid., pp. 24f and V). Such groups can be used for the creation and sale of valuable lists for various usage contexts. A constant update of preferences, activities, memberships or address data is guaranteed at all times.

A closer look at the target groups reveals quite different customer groups. One of the references of AZ Direct in Germany, for example, is a fashion manufacturer for tall women whose market entry was facilitated by data analysis. Other customers include a foreign direct bank which advertises loans tailored to the individual financial situation or a TV station which was able to further differentiate its advertising for the male target group. Acxiom Germany lists an automobile manufacturer as one of its references that was able to improve the marketing of its SUVs by matching the customer list with the users of a social media platform as well as another online platform, and a cruise company that acquired new customers via a target group selection by using online and offline channels. These examples make it clear that the stored data assets reach individuals in virtual space as well as in real space in different contexts, targeting them individually or as part of a customer segment.

One important form of customer segmentation that is offered by almost all data traders is via geographical criteria. In addition to macrogeographic divisions (states, cities, or municipalities), data traders often have microgeographic segments, which are spaces of living and everyday life shared by people with the same values or lifestyle or from similar milieus. Campaigns can then be targeted to location-based features. The available data allow for a space-based target market selection and also provide street and building directories for whole areas. Individual characteristics can be assigned to these, so that spatial clusters with specific characteristics can be addressed collectively or individual units can be addressed individually.

Overall, the offers clearly reveal the possibilities offered by data trading companies for targeting consumers. This applies to the scope of the data, their precision and the available analysis capacities, as well as the many professionally mastered channels of an address.

With the trading of geo data and forms of address in real space, the residential environment again becomes a focus of attention. The negative effects of disadvantaged neighbourhoods on their residents in terms of socialisation and the ability to constitute space were discussed above. In this context, access to cyberspace was considered as a form of individual empowerment within an alternative opportunity structure, which then revealed its limited potential because of a new form of recursivity. As data traders refer to residential areas, are consumers again defined in ways that hinder their advancement?

So far, the commercial translations of individual characteristics have been described as recursive effects, which—as shown clearly by the offers of the data traders—individuals are confronted with in quite different ways. The place of residence is relevant here as the address that makes it possible to contact the (potential) customer with offers in real space. Beyond that, however, address and neighbourhood also provide information about dispositions and preferences of their residents, allowing for—as underscored by the data traders’ advertising—more far-reaching deductions. For example, geo-scoring can be used to attribute specific needs or assess someone’s creditworthiness. However, via the additional possibility of tracing action spaces, the acquisition of geo data also influences the constitution of space. The deliberate placement of advertising is one example, but it goes even further.

Taking a closer look at how customers are addressed in particular, but also at the significance of a person’s address and neighbourhood in the context of data trading, the following section will directly apply purchasable geo data to individual neighbourhoods in German cities. The nature of the data and the traders’ associated offers point to utilisation contexts that have different social effects in each neighbourhood.

5.3 Empirical Findings in the Socially Segmented Urban Space

For a more detailed definition of the empirical focus, we will return to the starting point: Above, we looked at the interdependence between available capital and residential area. While the residential location and situation depend on economic resources, these resources also help finance the conditions which enable or limit residents: Physical, social as well as image-related circumstances go along with certain socialisation influences (contextual and compositional factors), and they determine to what degree further capital can be obtained. There exists a fundamental social inequality of opportunity which unfolds fully only in a spatial perspective.

The fewer resources a spatial opportunity structure offers, the greater the disadvantage of its inhabitants in relation to neighbourhoods rich in capital. But how polarised are cities currently? A look at the economic indicators alone reveals a clear polarisation of income in many cities even in wealthy countries, accompanied by an increasing segregation of residents according to their economic situation (Musterd et al. 2015). Some segments of the population simply cannot afford to live in certain neighbourhoods, or even certain cities. Their only option is to use those spaces which are affordable to low-income tenants.

However, the argument about enabling and limiting conditions in a spatial context would not hold if these differences applied only to economic capital, as, following Bourdieu, a wider concept of capital has been used here, tying social advancement to education, taste and social capital. Therefore, a structural disadvantage in the neighbourhood is primarily the result of a lack of access to social and cultural capital. Also, the term milieu itself implies that residents of a neighbourhood share further things in common which constitute their socio-cultural identity (e.g., taste regarding the aesthetics of everyday life, values, consumption preferences).

Contextual and compositional factors play an important role in the constitution of the milieu. At the same time, the residents of a neighbourhood in their action spaces contribute to the constitution of this space. Löw also supposes “milieu-specific operations of synthesis” that produce the urban space.

But beyond income, education and occupation, which other characteristics apply to a larger number of residents in a neighbourhood?

This question is of central importance for a data-based form of recursive address. If, in addition to economic characteristics, there are further items with a spatial concentration that can be found in the portfolios of data traders, it would follow that the personalised address based on them contributes to the solidification of neighbourhoods. A uniform address with similar financial as well as consumption offers would, over time, contribute to the homogeneity of the population of a neighbourhood. Added to the influence neighbourhood residents have over each other, there is then an externally controlled influence by profit-oriented companies with messages that are related to the dispositions recorded about this locality.

A closer look at the portfolio of data traders could therefore reveal to what extent the captured data reflect the shared characteristics of neighbourhood residents. This goes for economic characteristics, which generally show the degree of socio-spatial segmentation, as well as for further consumption-related characteristics that reflect broader commonalities among the neighbourhood population. In addition to the data collected, it is then of interest by whom and how these data are monetised, as they recursively influence both capital-poor and capital-rich inhabitants.

5.3.1 Acquisition of Data and Procedure

Another large German data trading company is the Bonn-based Nexiga GmbH, which focuses on the monetisation of geo data. It offers site plannings, market analyses and an optimised customer management via the space-related analysis of single characteristics with regard to their distribution, density and combination with other characteristics. Like the other providers mentioned, Nexiga has a large portfolio of stored characteristics on the population, from socio-demographic and economic information to data on sales psychology. The company’s market data are composed of more than 280 characteristics with a total of more than 1,000 specifications and can be projected onto different scale levels. In their most detailed resolution, data can be related to individual buildings, of which the company claims to have recorded well over 20 million in Germany alone (www.nexiga.de).

For a neighbourhood-based data analysis, up-to-date data sets for the German cities of Berlin, Munich and Essen were obtained from Nexiga’s portfolio in July 2018 and December 2022. Only two criteria guided the selection of these cities: On the one hand, large cities were chosen because neighbourhood formations are more evident there than in small and medium-sized cities. Berlin and Munich are both cities with several million inhabitants and are likely to have a large social spectrum. On the other hand, they were chosen because segregation is not too extreme, making it possible to explore the data economy’s reproductive logic in general and independently of already existing polarisations in the social sphere. A study by Helbig and Jähnen (2018) documents an increase in segregation between 2005 and 2014 for Germany’s 74 largest cities. Neither Berlin, which is affected by rising rents (17th place in the segregation index), nor Munich, dubbed “Germany's most expensive major city” (61st place), nor Essen, traditionally a dualistic city (25th place), occupy leading positions here (cf. Ibid., p. 30).

The information obtained on the three cities is composed of aggregated data that can be assigned to single street sections as well as data that can be assigned to a neighbourhood. At Nexiga, the neighbourhood level comprises small-scale statistical units, the size of which depends on the number of inhabitants.

At the street section level, several individual data (affinities, indices) with multi-level characteristics were obtained. At the neighbourhood level, a residential environment typology (rating in school grades), information on social class (from lower class to upper class in five levels), income classes (households with monthly net income in five levels) and other neighbourhood-related characteristics are added. The selection thus included characteristics of households that can be combined spatially with social criteria or housing quality. The information was issued as geo data in shapefile format.

The objective is to visualise socio-spatial correlations using various combinations of characteristics at the street and neighbourhood level and to illustrate concentrations of characteristics in an exemplary manner. The key point here is that the data are obtained primarily to gain commercial access to the population in question. Based on the sources of supply described, they provide clues as to where and how the different interests and business models of the data buyers can be realised. They express which segments of the population are being targeted when a particular characteristic they share is relevant for the address. The visualisation of selected features thus selectively reveals exploitation templates to which data buyers recursively refer with their personalised offers.

Admittedly, different resolutions and diverse characteristics can be chosen for this representation of the economic perspective. By aggregating the characteristics at the street segment level, a common form of presentation was chosen that lends itself to an overview of consumers and the living environment of individual cities. Taking up the data set unchanged, among other things, park, water or traffic areas are not shown separately and the area-related feature carriers are not set in relation to the area-related total. The data set centres the frequency of an economically relevant characteristic in a spatial segment. In accordance with the formulated interest in the connection of economic attributions (class, income) with consumption-related characteristics in the neighbourhood context, the corresponding characteristics are illustrated in excerpts, each in pairs.

5.3.2 Berlin, Munich and Essen: Exemplary Attributions in Urban Space

In recent years, Berlin has gained nationwide attention as a contested arena of urban displacement and gentrification processes. Due to growing immigration and rising rents, the city is now showing clear signs of segregation. The social transformation of the metropolis picked up speed after the fall of the Wall in the 1990s, when, as a result of delayed structural change, increasing unemployment was registered in the eastern but also in the Western part. Due to the social decline of the laid-off workforce and as a result of growing international immigration and selective mobility processes, a stronger concentration of low-income and educationally disadvantaged segments of the population developed in the urban area. Berlin evolved from a divided city to a “multi-fragmented city” (Krajewski 2015, p. 77). The city's diminishing options for shaping housing policy were exacerbated by the sale of municipal housing units, which was intended to relieve the budget deficit of the debt-ridden metropolis. As a result, private-sector influence on the rental market grew. In addition to privatisation and redevelopment, especially in the areas near the centre of Berlin-Mitte, Berlin's post-Fordist transformation brought about infrastructural improvements in many parts of the city. After years of population stagnation and belated suburbanisation processes on the outskirts of the city with further segregation effects, the economically resurgent city experienced sustained population growth in the years that followed. On average, the city has grown by almost 40,000 inhabitants annually since 2010, which has so far been insufficiently accommodated by new construction (Wetzstein 2018, p. 39f). In addition, the national and international acquisition of real estate has been increasing sharply and, together with a boom in tourism, is exposing residents of centrally located neighbourhoods to growing rental pressure (Holm 2011; Schnur 2013). The trendy neighbourhoods especially, all located in the Wilhelminian Gründerzeit belt, form the centres of a gentrification process that is spreading out from there. Improvements in construction and rent increases then encourage the often-described segregation, in which high-income residents with a specific capital endowment successively spread out at the expense of the less wealthy previous tenants, and the latter have to deal with increasing peripherisation. In Berlin, for example, a stronger spatial concentration of socially disadvantaged segments of the population could be observed from 2010 onward, among others, on the Western outskirts of the city in Spandau and on the eastern outskirts in Marzahn-Hellersdorf (Plate et al. 2014, p. 298f.). If the segregation processes described at the beginning of this study are reflected in Berlin (see, e.g., Holm 2016), it will be interesting to see how the data trade reflects these segregation processes.

We will first take a socio-spatial survey of the data trader as a basis (Fig. 5.1). The data cover a total of 2,168,383 households in Berlin. The area signature is used to highlight all neighbourhoods that have a larger number of lower-class households. Of the available 5 levels of social class, only the lowest level is reflected with the absolute number of households per unit area. At this scale level and taking into account the different reference areas, Berlin presents itself as relatively mixed according to this characteristic. Nevertheless, concentrations of poorer parts of the population can be easily traced via the small-scale dark signatures. In the outer part of the urban area, the large housing estates stand out, such as those found on the above-mentioned city outskirts in Spandau and Marzahn-Hellersdorf. The “Märkische Viertel” in the north and “Gropuisstadt” in the south also stand out. In addition to a clearly discernible share of underclass residents in the Western district of Wilmersdorf, the densely urbanised inner-city belt in particular is characterised by a mosaic of different shades of disadvantaged neighbourhoods. Those neighbourhoods, with their large numbers of low-income residents, would be highly vulnerable in case of further rent increases. By point signature, households with a high affinity for instalment loans are indicated. Of the nine levels between “very low” and “very high” affinity, only the highest level has been selected. All street segments with more than 10% of residents with a very high affinity are marked with dots, which was the case for 404 out of 134,267 street segments.

Fig. 5.1

Berlin—Lower class and instalment loans

Fig. 5.2

Berlin—Holders of academic degrees and business magazines

Clearly, prospective borrowers are heavily concentrated in neighbourhoods with a large number of socially disadvantaged households. A direct address of the neighbourhood residents on the basis of these data could then come from banks, for example, which advertise their financial products in a direct translation of the item. Above, such a typical utilisation context was cited as a reference of a data trader. With regard to the channels described, there are numerous possibilities for making contact: Potential customers can be addressed on the websites they visit, through direct mail or e-mail or collectively. With the clustering of potential customers in a neighbourhood, a direct mail piece to all residents in that cluster might be a good idea, or neighbourhood-specific billboard advertising might seem appropriate. In other utilisation contexts, an affinity for loans signals a lack of creditworthiness. In combination with other items, this implies excluding neighbourhood residents from targeting with respect to this characteristic. Consequently, customer segmentation selectively interferes in the urban space on the basis of a comprehensive data base, addressing individuals or groups with the offers appropriate for them. As the distributions of the different characteristic bearers clearly show, offers and information reach the inhabitants of Berlin in a highly unequal manner. They correspond with economic status and residential location, making an exchange in these exact social and spatial contexts more likely.

Whether a company can benefit from targeting a spatial segment with an out-of-home campaign, for example, depends on the concentration of the relevant characteristic bearers. Even if a resident does not share the relevant attributes, the fact that he lives in a certain neighbourhood can determine which information and offers he receives. As a result, the recursive reference is not directed specifically at the person but at the feature cluster of a larger section of space.

As their lack of capital is possibly reproduced through specific offers, residents of individual streets or entire neighbourhoods must increasingly deal with offers detrimental to them. At the other end of the spectrum, potential resources reach the privileged segments of the population due to the characteristics recorded about them. Figure 5.2 provides an overview of how academic degree holders are distributed across the city. First of all, the overview reveals that this form of symbolic capital coincides with economic capital in many neighbourhoods. A concentration of holders of academic degrees is clearly visible in the affluent district of Steglitz-Zehlendorf, the districts of Charlottenburg and Westend as well as in the southern parts of the Mitte district. The existence of this category as well will play a role in different utilisation contexts, generally promoting a form of address with further appropriate offers tailored to the educational level.

Superimposing the affinity for business and financial magazines (the highest and second-highest values on the nine-point scale) on top of this plot, shown here in points denoting more than 20% of all households per street segment, again reveals numerous overlaps. As it makes little sense for publishers to invest in contacting populations with no affinity, the identified target group is predestined to receive information with which they can potentially increase their capital. By contrast, the high target group transparency of the data economy prevents the population segments which are not included from obtaining such information. While obtaining this type of information at a later stage is not impossible, it is certainly made less likely, be it via online channels or via addresses in real space.

A closer look at the data for Berlin also reveals that rough interpolations were made for several neighbourhoods (e.g., numerous streets in the Prenzlauer Berg or Tempelhof districts are marked with the same characteristics). Data traders transfer mean values to neighbouring districts, neglecting a more precise description of the target group. Similarly, neighbourhoods that extend across large areas or have a small population are described uniformly on the basis of a small number of feature carriers. The large area at the outskirts of the Western part of Berlin is a good example for this. It includes the Spandauer Forst, which has residential buildings only in its south and east. The overall low social status of its inhabitants is disproportionately represented over the large total area. As a consequence of such generalisations, other people’s data can lead to individuals being put at a disadvantage. Conversely, a lack of precision can also mean less rigid personalisation, allowing an individual to receive offers and information not associated with their milieu. Ultimately, each company determines the resolution, the target group accuracy and the quality of the data that it uses to reach its goals.

On the whole, in spite of imprecisions regarding the selected scale levels, patterns are emerging in Berlin's urban space that represent selective targeting of individuals as well as groups of people in the neighbourhood. These patterns are based on personalised data which, through trade and economic valorisation, fortify that which they describe in terms of content and space. The battle over social sovereignty within individual neighbourhoods, as it is currently being waged in Berlin, is thus also related to how a neighbourhood is recursively addressed.

The Bavarian capital of Munich has been growing steadily for decades, allowing the city and the region to become one of Europe’s most successful locations. The fact that this city is less affected by social polarisation processes than other big cities in Germany is largely owed to its economic structure. The upheaval caused by the structural and social transformation processes of the late 1970s affected Munich to a lesser extent. As Munich was incorporated into Germany’s industrialisation and modernisation process fairly late, it succeeded in attracting crisis-resistant high-tech companies which mitigated the decline of the old industries. Until today, the economy in the Munich metropolitan area is characterised by a mix of companies, some of which are interconnected, of different sizes and from a wide range of industries (the so-called Munich Mix). While the city’s economic prosperity has positively impacted its job market and also its options regarding social policies, the high cost of living is a big problem for large parts of the population. Already in the 1990s, the Munich Report on Poverty pointed to a rising poverty risk as a result of higher costs of living, a circumstance which the city is still dealing with in the present (Martens 2011, S. 177). Rents, in particular, are placing a growing burden on people’s budgets in this rapidly growing city. Similar to Berlin, current challenges for housing policy are linked to the real estate boom in Germany, due to which rental prices have been increasing even more since around 2010. Presently, rents across Munich have become so high that most low-income residents have no choice but to move further and further away from the centre.

Figure 5.3 shows all households which are assigned the worst housing location (level “poor”). In addition, neighbourhoods are highlighted in which a larger number of residents is considered to belong to the lower class. Overall, the data include 813,491 households. Both characteristics reflect a relatively balanced socio-spatial distribution and correlate only to a limited extent. A poor residential location cannot be applied to larger neighbourhoods across the board; rather, it is heavily trafficked streets, especially arterial roads and parts of the Mittlerer Ring, that stand out as disadvantaged residential locations. Households that form part of the lower class are equally divided across different urban areas. Nevertheless, Munich also has neighbourhoods where poor residential location and a high number of lower-class households coincide, such as the districts of Hasenbergl and Am Hart, which are characterised by larger apartment blocks. Like in other disadvantaged neighbourhoods, residents are faced with various location-related disadvantages, such as those described above. With respect to the contextual factors, the material equipment as well as the symbolic connotations of a place are to be emphasised. Hasenbergl in particular has over the decades become known as Munich’s problem neighbourhood. The negative connotations of the neighbourhood can have a detrimental effect on residents’ self-esteem, and the stigma associated with it can put them at a disadvantage in professional and everyday life.

Fig. 5.3

Munich—Lower class and quality of residential location

With the intelligence collected by the data-processing service providers, discriminating targeting becomes systemic: Companies offering exclusive goods or advertise certain cultural events or even such that take place elsewhere will hardly target any of the residents of such a neighbourhood as they do not see a market here. Consequently, the corresponding content does not become part of the socialisation context or neighbourhood exchange. Class-specific habitualisations are challenged by external offers to a lesser extent and social advancement is made more difficult.

To trace the selective targeting of neighbourhood residents in more detail, we will take a more differentiated look at a section in Munich’s southeast, specifically the neighbourhoods of the Ramersdorf district and the Neuperlach district (Fig. 5.4). The core of eastern Ramersdorf is an established single-family housing development with middle-class residents, which in the map section only changes to a multi-story row house area with a larger number of social housing units in the south up to Ständlerstraße. Connected to the east by the Ostpark is Neuperlach, one of the largest post-war housing projects in West Germany. Neuperlach was erected in the 1960s and 70s as an extensive large housing estate to meet the rapidly growing demand for housing in Munich at the time. Originally built for middle-class residents, Neuperlach has today become one of Munich’s poorer districts. In a citywide comparison, the social indicators reveal higher scores with regard to poverty density and receipt of government support (Stadt München 2017). Although the social differences between the neighbourhoods are relativised in the overall social spectrum of Munich, the neighbourhood structure still points to different socialisation contexts within a distance of only a few 100 meters. The residents of northern Ramersdorf have relatively large, individually different house and living spaces as well as their own use of gardens. The neighbourhood has facilities such as a restaurant, a bakery, clothing stores, a Montessori preschool and a centrally located park which foster contact between residents. In contrast, Neuperlach’s large housing estate is characterised by standardised construction, offering considerably less room for individuality. For shopping and leisure activities, residents only have the estate’s central facilities. Furthermore, both neighbourhoods belong to different school districts, which means that the children do not have a chance to come into contact with each other during school. While Munich’s southwest is generally wealthy, it can be divided into different areas of acquisition for the population. While upper-class neighbourhoods further away (such as Nymphenburg or the suburb of Grünwald) are certainly characterised by a higher degree of social distance, even in a small geographic area there are clear differences regarding the conditions for the acquisition of capital and the above-mentioned barriers regarding access, perception and social exclusion.

Fig. 5.4

Munich, Ramersdorf-Perlach—Purchasing power index and concern for the environment

With regard to the external influence of data traders the question is whether these small-scale differences between Ramersdorf and Neuperlach show up in the provider’s portfolio. As an example, the purchasing power beyond the Munich average (index value > 125) and concern for the environment is shown in its effect on consumer behaviour. While the purchasing power relevant for the sale of products is exists in some of Neuperlach’s households, the potential sales area is diminished by other characteristics, such as concern for the environment. A marked concern on the part of the residents of eastern Ramersdorf stands in stark contrast to the neighbouring residents, with clear consequences for the targeting. As a consequence, offers and information related to the environment are more likely to reach eastern Ramersdorf, providing content for everyday activities and social exchange.

Since the data-based selection is derived from personalised information, commercial addresses always include messages related to this information. In relation to a neighbourhood, however, this does not mean that a campaign will be addressed exclusively to that neighbourhood. Companies aiming to place a specific ad, a credit offer or an insurance policy will also consider characteristics relevant that are present across spaces and milieus. In the end, the provider’s goal is to maximise customer potential. Numerous characteristics, such as “gender”, “customer card user”, affinity for home delivery food” or “distrustful financial client”, which are offered by Nexiga for target group segmentation are not attributable to an economic status. The available diversity of characteristics, which affects rich and poor residents alike, could result in a targeting that provides disadvantaged segments of the population with the same content that wealthier households also receive. In other words: With reference to items that apply to all milieus, the argument of recursive targeting could be countered. The characteristics that are then relevant in this context are the ones which allow access to services and goods with which their recipients can successfully engage in processes of distinction. One such characteristic could be “significance of product novelty”.

Figure 5.5 shows the highest value for the attitude “The product novelty/innovation factor is decisive when buying consumer goods” (the highest and second-highest values on the nine-point scale) and

Fig. 5.5

Munich—Upper class and product novelty

the frequency of upper-class households. There is no clear correlation between the two characteristics. However, an accumulation regarding this interest stands out in the city centre, even if one takes into account the relatively higher density of settlement in the Gründerzeit neighbourhoods close to the city centre. The interest in product novelty may partly coincide with the spatial distribution of younger urban residents, for whom the desire for an urban environment in a central location is particularly important. Even if this again means a disproportional representation of expensive and prestigious districts such as Schwabing-West, Maxvorstadt or Au-Haidhausen, inhabitants lacking capital are also linked to attributions that carry distinctions in this example. A discriminating form of address again comes into play when cross-milieu characteristics are combined with other characteristics in data processing. Someone who markets an expensive watch, for example, will want more than a target group that appreciates the novelty of the product. To minimise wastage, the addressed target group should above all be able to pay for the watch. Similarly, a basic “affinity for home delivery food” or a “distrusting financial customer” still results in a fuzzy customer profile, the specification of which can be achieved precisely via additional economic characteristics. Last but not least, the widespread use of predefined milieu classifications (Sinus milieus, Geo milieus) underscores the high relevance of economic capital in the data traders’ portfolio.

The comparison of the Hasenbergl/Am Hart district and the expensive Bogenhausen district (Fig. 5.6) shows how the affinity for new products changes when economic criteria, such as the affinity for instalment loans, are considered: Both have high percentages of residents receptive to new, innovative products per street segment (highest and second highest on the nine-point scale). However, if only customers with no interest in instalment loans are considered (also the highest and second-highest score on the nine-point scale), a provider will direct his offers almost exclusively to Bogenhausen.

Another argument against the social bridging function of cross-milieu items lies in the diversity of address. By using personal data to present the customer with ever-new commercial messages, he is confronted with his own dispositions in a variety of ways. Overall, a highly individual form of acquisition must be assumed which distinguishes between residents of a neighbourhood in many respects. This does certainly not imply social change. If the existing dispositions of every individual are mirrored, the same is true for the population of a neighbourhood. The influence coming from the outside may have a less homogenising effect on the social conditions, but what reaches an individual in various combinations is again a reflection of individual dispositions.

Munich as a city less segregated than others is affected by such a stabilisation of existing social conditions through a reflexive form of address just as much as other cities.

Fig. 5.6

Munich—Product novelty and instalment loans in a comparison of neighbourhoods

As a former centre of the German coal and steel industry, Essen is fundamentally different from the above-mentioned cities in its economic development and socio-spatial structure. The city’s boom during industrialisation ended with the decline of mining in the late 1950s. The long-term structural crisis led to a significant decline of the population since 1965. Not until 2012 did the city again record a population increase. Essen’s socio-spatial structure is directly linked to the development of mining: The more densely built-up north, with numerous working-class neighbourhoods in former coal-mining areas, was directly affected by the northward migration of mining and by the crisis in coal and steel, while the south established itself as the preferred residential area of industrialists and other wealthy sections of the population. A few exceptions apart, Essen’s north has long been characterised by a disproportionately high number of welfare recipients and a high proportion of immigrants (Grabbert 2008, S. 139ff). By contrast, the south of Essen has only a low number of low-income residents. This inequality of opportunity is reflected in the field of education, among others. In the north, a comparatively low percentage of students attends a school that prepares them for university, while the numerically smaller number of children from the south mostly do their Abitur (the prerequisite for studying at a university) (Strohmeier 2006, S. 13). In recent years, a gradient between outside and inside has been added to Essen’s polarised spatial structure. Some neighbourhoods on the outskirts of the northern city area show signs of structural upgrading. By contrast, the northern belt of the city centre, with its post-war apartment buildings, is one of the city’s main areas in need of development.

Essen’s large-scale polarisation is clearly reflected in the traded geo data, which includes 312,439 households. Households with a “good” or “very good” residential location are located in the neighbourhoods that also have the highest number of upper-class households (Fig. 5.7). While few upper-class households can be found in Essen’s north, the south has a high concentration of wealthy residents. The discrepancy between inside and outside mentioned above is reflected particularly in the concentration of upper-class households in the northwest, where the neighbourhoods of Frintrop and Gerschede also have a very high-quality residential location. In comparison, the centre of Essen is characterised by areas with a low number of upper-class households.

Fig. 5.7

Essen—Upper class and quality of the residential area

The map clearly reflects residential location and social class even in small-scale concentration; a notable cluster of upper-class households is always tied to a privileged residential location. Essen is a prime example of the congruence between social and spatial inequality, at the same time pointing to the socially reproductive effect associated with the social, symbolic and material dimensions of the neighbourhood.

Their causes can be traced back historically in a relational perspective to, among other things, the powerful constitution of space by the mine operators and the economically justified arrangement of infrastructure and workers’ settlements. Numerous external actors are also involved in the institutionalised space of the mining industry, for example, as buyers of raw materials, who later, through their reduced demand, contributed to the industrial area’s decline.

To conclude, we will once again relate the polarised social structure to the selective targeting of the data-processing economy. Analogous to the cluster of upper-class households, Essen’s lower-class households are shown in their spatial distribution. They are combined here with “affinity for TV shopping” (highest and second-highest values on the nine-point scale) (Fig. 5.8). By way of example, the distribution of the characteristic carriers again points to the spatial connection between economic and consumption-related characteristics. The affinity for TV shopping can be found almost exclusively in Essen’s north with its prevalence of lower-class milieus. More than 1,000 households here are considered to be receptive to TV targeting, a result that is certainly relevant for many businesses interested in individual forms of address. Via smart TV, this demand can be met in an increasingly personalised manner. Conversely, a higher relevance of TV consumption and the messages from advertising feed into consumer socialisation. Unless mobile platforms are used, this way of obtaining what one needs ties consumers to their homes, replacing trips to the places that are in demand across milieus. Over the long term, such places and their resources disappear, as perception becomes more and more habitualised. The forms of social recognition including the possibilities of distinction from the products obtained via TV shopping take place exclusively within the milieu.

Fig. 5.8

Essen—Lower class and TV shopping

Superimposing the slides shown with their attributions of the population, the selection made already reveals a system of utilisation that reproduces inequality. This happens within cities via the unequal allocation of information and services, which were diagnosed above as resources for capital acquisition. The unequal allocation takes place at different scale levels according to economic selection criteria and has the potential of social entrenchment.

So far, references to space and neighbourhoods have been made deliberately, although the traded data can reach individuals also outside of their place of residence. In addition, the data feed into other utilisation contexts directed at the neighbourhood, influencing opportunity structures there.

5.3.3 Externally Driven Site Effects

The campaigns mentioned target individual or collective attributions from which those interests and needs are sifted out which the business exploiting the data can meet with specific products and services. In addition to this usage which involves a direct address of the relevant feature carriers, there are other forms of data utilisation. Product providers can use the transparency of a city’s inhabitants to align their local presence with the customer.

Let us return to the findings on Munich’s southeast, which revealed an increased demand for organic products in eastern Ramersdorf and a lower demand in the adjacent neighbourhood of Neuperlach. For someone planning to open a new store based on these criteria and in search of the right location, such comparisons are crucial. Thanks to the extensive database, the sales and customer potential, frequencies, market environment, accessibility, consumer and residential typology, product-specific purchasing power or competing locations can all be analysed for each location planning, as offered for instance by Nexiga. Nexiga refers to its own case study on an organic supermarket, for which the company analysed, among other characteristics, the number of households with the “environmental” consumption style, retail-relevant purchasing power and age groups (https://www.nexiga.com/?s=Case+Study). This helps the respective provider to estimate which investments are worthwhile in which neighbourhoods.

For neighbourhood residents, this results in a utility infrastructure with placements and orientations that are a result of recorded attributions. While purchasing power may only influence the number of shops, the translation of consumption preferences contributes to a greater differentiation of living contexts. Milieus with high levels of education and capital can expect a different supply structure in their neighbourhoods than socially disadvantaged ones. Citing Bourdieu, the ruling class’s “taste for luxury” could be contrasted with the lower class’s “taste for necessity”. Interestingly, the different supply structures do not result primarily from the actions of the neighbourhood residents, but from the decision of businesses operators, who create personalised spaces based on a comprehensive database. What Bourdieu failed to see is a constitution of space that comes before actions and practices of distinction, in this case resulting from the rational calculations of a data-based site planning.

This externally driven design of space can have far-reaching consequences: Based on the products of data trader Nexiga, site planning involves optimising all operational processes, selecting suitable territories and sites according to sales potential or identifying suitable personnel. In addition, the neighbourhood residents’ numerous items can be used to align products, prices and assortments with regional and local needs as well. What the data economy does is thus nothing less than providing an opportunity to rationally shape space in those areas that are not regulated by the public sector. Even if this implementation is gradual, many data-based decisions do not necessarily have anything to do with personalised dispositions and persistent structures are constraining it, a growing impact on the neighbourhood must be considered.

If this form of spatial constitution is viewed as a socialisation context, we are again facing the problem of inequality of opportunity, in this case dependent on the neighbourhood. As has been shown for Berlin, Munich and Essen, in each case a different social spectrum results in spatial structures which provide unequal resources. In an extreme case, on one end of the spectrum we find a spatial structure that provides its distinguished residents with rare and expensive goods, supporting them in their practices of distinction. This is taking place in an exclusive sales environment, which at the other end of the spectrum is contrasted with a low-priced standard range in simple stores or discounters. Even when the differences are less pronounced, it is evident that the lifelong experiences in these environments shape young people for different life paths. In the first case, clear opportunities exist to increase the cultural, social and symbolic capital, while in the second case, there are hardly any guidelines for connecting to the rules, codes and manners of the privileged class. The private-sector’s interest in a profitable spatial design produces new site effects.

Data-based constitution of space reveals another form of recursive acquisition. While so far freely traded data in the different forms of their capitalisation represented a direct form of recursive address, it is now the transformation into spatial structures which reaches the consumer recursively in an indirect form. The neighbourhood is constituted through corporate action, expressing inhabitants’ dispositions within an economically relevant segment of space. By mirroring the dispositions of the spatial segment it creates a recursive acquisition.

To summarise, it is evident that real-space structures do not necessarily need a technological infrastructure with sensors, cameras and displays to reflect social conditions based on data. Added to the types of recursive acquisition, as illustrated by numerous examples covering all basic functions of existence as well as in the context of the sentient city, there is the indirect effect of a commercial infrastructure. Mixed forms of such data-based spaces are also conceivable, as for example when a supermarket uses personalised data both for site planning and for ads on their own customer displays. Clearly, we are dealing with different contexts which, in addition to virtual forms of personalised influence, also have an effect in physical real space. In view of growing data markets, which are able to guarantee economic success through customer transparency more and more convincingly, further translations of individual characteristics into commercial offers are to be expected. Economic actors are indirectly perpetuating social inequality.

5.4 Interim Conclusion: Bourdieu in the Context of the Data Economy

Commercial data trading has a strong social influence, which is not based on the data traders’ privilege to alone cover the need for personalised data. As has been shown, a targeted capturing of data is now taking place in many different contexts, and they are passed on in various forms outside of the official markets. What is remarkable, however, is that it is data trade that can systematically externalise in aggregated and economically processed form all that has accumulated about individuals in a wide variety of contexts by means of modern information and communication technologies. Increasing transparency of the target person can be achieved in a planned manner through data trading from manifold sources and can be put into value economically.

The data subjects themselves hardly have an economic benefit as a result of this cross-space trade. But since the characteristics that are traded are immaterial, the loss remains abstract. It becomes concrete, however, when the ways of monetisation are analysed.

The form of analysis chosen here focuses on the major data traders in Germany. Their comprehensive portfolio first of all testifies to the commercial significance of the data in general. For an interpretation focused on social stratification processes, the way they are used is also instructive: The customer can be reached through numerous channels via cyberspace or real space by linking the respective information, service or product offer with the stored characteristics of the target person. Familiar and fitting messages are recombined economically and brought to the customer recursively. According to Bourdieu, the milieu-specific acquisition of knowledge, taste or culture contains the reflexive element which consolidates the habitus, reducing the chances to advance to other milieus. Correspondingly, the recursive adoption of habitualised characterisations by data buyers, transmitted by personalised messages and appropriated via products, information and services, must be understood as a barrier to advancement.

Bourdieu tied social positions that people occupy in social space depending on their capital to lifestyles, relating consumption and leisure preferences, occupations and income to each other (Bourdieu 1984, S. 262f). The different indicators used for the classification of milieus now appear in the data traders’ portfolio in a greatly expanded form. A wide range of information with detailed subdivisions is available on large parts of the population—in some cases even in real time. With this information, it is possible to differentiate groups in a more refined and comprehensive way than Bourdieu was ever able to do in his empirical studies in France.

At a superficial glance, this finding could be taken as an argument for a far-reaching differentiation of society, in which social differences are blurred by the diversity of available lifestyles despite unequal capital endowments. In this view, an enormous range of features could be combined with an enormous variety of offers.

In the hands of data traders and their customers, however, attributions represent the possibility of deliberately moderating the fundamentally free choice of lifestyle. In partly automated allocations, content is directly linked to existing preferences, while the range of interests that are addressed often depend on the economic capital. This characteristic, which is a central component of almost all offers and typologies, makes it possible for the provider to assess the ability to pay and to offer a more differentiated service according to the interests recorded.

While Bourdieu might find it more difficult today to distinguish lifestyles and social positions by means of distinctive characteristics, there is much to be said for the fact that addressing each individual in a precisely tailored way systematically counteracts a differentiation that is independent of class. The heterogeneity of the target groups can be managed by flexible combinations of the richly available data, hampering social change.

From a spatial perspective, the reproductive logic must be particularly evident if a data-based approach also reaches the majority of neighbourhood residents. Then, the “mirroring” of that which is one's own is supplemented by a one-sided address of the neighbourhood. A closer look at commercially offered geo data reveals exactly this influence on the different neighbourhoods of Berlin, Munich and Essen, among others. Depending on the available income, population groups are not only distributed differently across the urban area, they also sometimes show similar affinities in their respective clusters. Thus, the selective targeting of individual population groups reproduces their neighbourhoods beyond economic characteristics.

Since the economic circumstances of neighbourhood residents often play a special role in practice, as is clearly demonstrated by almost all of the data traders’ reference projects, economic criteria come into play again and again as a principle for structuring customer groups. As an example from Munich Bogenhausen in comparison with Hasenbergl/Am Hart illustrates, the neighbourhoods relevant for the address can then be grouped together even more closely via the material dimension.

In addition to the various selection processes used by companies to narrow down target groups for direct contact, there are also selections for location and sales planning. The supply intended for each population segment, such as supermarket offering certain products, is again based on economic possibilities and other preferences, which are expressed physically, resulting in overarching (dis)advantages as a result of how spaces are equipped.

Bourdieu (especially 1991) is always guided by the actions of the privileged who take possession of a space and also occupy it symbolically. Under the data economy, however, the different spatial profits also result from the data-based decisions of businesses, whose sole focus is on the distribution of market-relevant actors in the urban space. In combination with the different ways of directing messages at each target group, they have a reinforcing effect on social milieus in space. Paradoxically, this is taking place under the technological conditions of the digital age, once praised as an opportunity to overcome social stratification processes.