THE MATHEMATICS OF DESIRE

Knowing ‘You’

But now things, it seems, have changed; people have changed. With the decline in mass manufacturing in the West, the consumption patterns of the new worker are shifting. Commentators on business, marketing and sociology have lined up to describe this new individual. No longer content to be defined by the group (and to consume en mass) or assuming that State or Corporation will care for them, the new individuals represent the emergence of a more segmented and individuated market. This, observers claim, is causing serious problems for business. I discuss the ‘new’ individual later, but first I want to look at the role of technology in this change of patterns of consumption.

Television, the most powerful of the twentieth century marketing technologies, is a good place to start. Recently, the Financial Times assigned the reason for the decline in audience share of ITV, one of the five UK terrestrial TV channels, to the proliferation of satellite TV Channels. Five years ago ITV had 60% share of the market, now it is down to 22.6%, a decline that appears inexorable.12 Of course, less audience share means a smaller number of people exposed to advertising, and television advertising spend has consistently fallen as a consequence. Marketing budgets are being spent elsewhere and the internet is a big beneficiary. Despite a retrenchment immediately following the dotcom bust, online advertising and marketing is growing significantly. PricewaterhouseCoopers has reported an 85% rise in online spending in 2003,13 with the Advertising Association giving 376m as the figure spent by UK advertisers in 2003.14 This is in the context of a decline in marketing budgets generally, 11% in the second quarter of 2003 for instance.15 Of course actual usage levels of the internet are an indication of the medium’s potential to reach customers and these usage levels are high. The UK’s then telecommunications regulator, Oftel, calculated that in December 2003 50% of UK households were connected to the internet.16

As the Financial Times enthuses ‘The greater ability to personalise advertising using new technology - ranging from the internet and mobile technology to digital printing - may provide something of an answer. It allows advertisers to exploit the very fragmentation that denies them the single broad platform that old-style TV used to give.’ 17 This ability to reach individuals is often cited as the advantage of digital media, whether it is personalised e-mails or customised direct mail print outs from digital presses.

Reaching individuals may mean simply the acquisition of a list of email addresses and a bulk mail out, banner ads or the use of cookies to monitor browsing histories. At its most developed, however, such digital marketing techniques encompass a range of approaches utilising extensive information on individuals. This area is loosely known as ‘customer relationship management’ (CRM), and is seen as key to the future of marketing. It is a broad area, encompassing call centre management software, customer support systems, sales process automation (SA), account management and so forth. There are also widely differing degrees of intrusion into the lives of customers, ranging from the basic cross selling strategies of online retailers ("people who bought that also bought this") through to enormously sophisticated pricing strategies of grocery loyalty card schemes. As with all such wide ranging marketing concepts there are various conflicting definitions of what it actually means, however, as a quote from one of the North American CRM industry association sites illustrates, a notion of ‘customer centricity’ runs through most of them:

‘CRM...is a company-wide business strategy designed to reduce costs and increase profitability by solidifying customer loyalty...If customer relationships are the heart of business success, then CRM is the valve that pumps a company's life blood...It's a strategy used to learn more about customers' needs and behaviors in order to develop stronger relationships with them.

..True CRM brings together information from all data sources...to give one, holistic view of each customer in real time. This allows customer facing employees...to make informed decisions on everything from cross-selling and up-selling opportunities to target marketing strategies to competitive positioning tactics.’18

Although now a broadly encompassing concept, relationship management initially, in the early nineties, referred to the software that managed the process. This link between technological capacity and marketing philosophy is instructive. In his article Messages That Never Miss the Mark, Simon London describes the correlation between the development of ‘one to one’ marketing techniques and advances in technology: ‘Technology was a large part of the argument. Mr Edelman, who left BCG two years ago, says the consulting firm could see that the cost of data storage was falling fast. This was making it possible for companies to warehouse data in hitherto uneconomic quantities. Moreover, data capture opportunities were on the rise, with the advent of electronic point-of-sale (Epos) systems, bar-coding, loyalty cards and the like.’19

Thus new philosophies of marketing evolved because technological advances made them possible. However, to avoid, like Beringer, of falling into the trap of believing technology to be self evolving, there needs to be an explanation for the persistence of CRM and its successful evolution from a range of software packages into a marketing philosophy. A brief history of ‘Relationship’ and One to One’ marketing techniques can help us here.

The phrase ‘segment-of-one’ was created in the late 1980s by The Boston Consulting Group (BCG), the management consultancy. ‘One-to-One Marketing’ was popularised a few years later with the publication in 1993 of The One to One Future, a business best-seller written by Don Peppers and Martha Rogers, two former advertising executives. These concepts concentrated on the relationship between sellers and buyers, aiming to re-orientate companies away from the products they manufactured or sold, towards the customer. Shoshana Zuboff and James Maxim in The Support Economy, (a broad ranging look at the modern consumer and how, they claim, corporations are failing them), explain the popularity of these techniques amongst companies. By the late 1980s, as the collapse of Fordism and the ascendance of free market ideologies drove the globalisation of production and consumption, fierce competition beset many companies. Within this new global marketplace a profusion of new products fought for the customer’s attention (Zuboff and Maxim quote the statistic that between 1985 and 1989, the number of new products grew by 60%, to an all time high of 12,055 per year).20 The logic for personalised marketing, along with the increasing awareness of the importance of brand, was that amongst this blizzard of products, customers would respond to marketing that appealed to their individuality; companies would be able to learn more about their customers and be able to offer customization (however trivial) of products that this new intimacy would demand. In doing so companies could use the power of a relationship (or relationships) - loyalty, emotional investment and exclusivity - to establish a unique and compelling position in the market place. Zuboff and Maxim quote Regis McKenna from his book of 1991, Relationship Marketing: ‘In a world where customers have so many choices, they can be fickle. This means modern marketing is a battle for customer loyalty. Positioning must involve more than simple awareness of a hierarchy of brands and company names. It demands a special relationship with the customer.’21

Couched in the language of intimacy and convenience, these techniques nonetheless exhibited a familiar preoccupation with control of consumption. Zuboff and Maxim continue, critically quoting Don Peppers and Martha Rogers: ‘ " Now, even if a competitor offers the same type of customization and interaction, your customer won’t be able to get back to the same level of convenience until he re-teaches the competitor what he’s already spent time and energy teaching you." In other words, relationship marketing and its one-to-one cousin were from the start strategies to limit consumer choice.’22

This pre-occupation with ‘knowing’ the customer still persists today in the literature of software vendors and consultants: ‘A company's business processes must be reengineered to bolster its CRM initiative, often from the view of: ‘how can this process better serve the customer?’23 Again, behind the rhetoric lies a more hardnosed reality: ‘What works is the company-wide commitment to customers, the ongoing creation of customer-perceived value and 'barriers to exit' which leads to loyalty and advocacy.’ And: ‘Success will be defined by three outcomes: the highest share of customer possible, optimal lifetime customer value generation, and the lowest voluntary churn.’24

Lucrative customers are identified,25 their spend optimised and their loyalty enforced. Within the logic of enterprise capitalism26 this is to be expected, however, as I shall consider in the following section, it belies claims that technology is fundamentally changing the relationship between producer and consumer; the emphasis on control of consumption is a linear descendent of the early twentieth century marketing innovations already discussed. The objective is to know accurately what the demand for goods and services is, to produce efficiently to meet this demand and to minimise variance. It is business as usual.

In 1924, halfway through the year, Alfred Sloan, head of General Motors, went west to confirm whether his hunch was right and that the company was building more automobiles than the market demanded, despite his newly instigated market feedback mechanisms. On discovering in various mid-west cities that indeed, the forecourts were full of unsold vehicles, he slashed production and introduced a still more refined feedback process. This determination to seek new levels of linkage between demand and supply can be still be seen in contemporary business practice as it relates to CRM, such as in a text book on data mining and business intelligence that devotes a section to the application of 6 Sigma to marketing and sales,27 or a recent article in The Register, the technology news site. Reporting on the precipitous decline in CRM software revenues, the article predicts that oCRm (operational CRM) software vendors will have to partner with analytical CRM vendors (data mining) to ensure survival and growth. That is, only companies that can supply software to model and predict customer behaviour, as well as automate basic routine sales and marketing operations, will succeed.28 Here we see the integration of business metrics, sales process and product marketing. This is not a new paradigm, but computers have facilitated the integration and given unprecedented speed and reach to the project; discovering what customer’s want, but also to creating the want, of managing their desire.

Consider the notion of personalization. This connection between knowing an individual and using this knowledge to offer specific and appropriate goods and services lies at the heart of accounts of the technologies’ effectiveness. In 2001 two BCG consultants wrote: ‘Ten years ago marketers discovered they could narrow their focus and create products for specific customer segments. Now a segment can be trimmed down to an individual.’ 29

Two years later an advert from IBM, the UK’s number one retail technology vendor: ‘A sense-and-respond retail environment, for instance, would know every time its best customers entered the store. It would be able to respond to what each valued customer was shopping for that day and suggest appropriate cross- and up-sells. Products would be in stock, promotions would be relevant, sales associates would be experts, check-out would be instantaneous.’ 30

A study of data mining, the technologies and methodologies that create this ‘personalisation’, reveals a different story. Data mining is based in standard statistical operations and the models of econometrics, but mediated by computer algorithms and adapted for analysis of very large amounts of data. That these techniques are powerful and effective in a limited way is, debatably, true (see the following section), however, they only allow manipulation of marketing data within familiar segmented models. In the same way that, due to the impossible (potentially infinite) profusion of data involved in the creation of true individual narrative paths, such interactive narratives can only fake interactive effects,31 data mining must link individual profiles to prior known groupings, themselves constructed from data exploration approaches such as clustering and association. Thus when a customer buys a book on Amazon, they are offered a deal including another book or when a supermarket emails an offer to you, say for Australian wine, the personalization is the result of identifying which segment/group, or combination of segments/groups that you the customer belong to.32 What is unique about the technology is the ability, at speed, to receive the individual’s information, process the information (i.e. associate to groupings and compute the combination value) and output the relevant information. Clearly, such strategies have use for the company in terms of pricing, sales optimisation and the maximisation of profit. There is, debatably, also some limited use to the customer, in that unfamiliar products may come to their notice. However, the reality of this species of machine ‘intelligence’ falls a long way short of the ideal: it does not offer absolute and unfailing knowledge of the individual consumer, or present potentially transformative insights into who a customer is, or what they want now and in the future. ‘One to one’ or ‘relationship’ marketing in this sense is technical fakery.

That the work of data mining can only be founded on the assumptions and goals of the retailer, and involves the needs of the individual customer only tangentially, is confirmed by its epistemology. As the author of the Principles of Data Mining states: ‘It (choosing the appropriate analytic model) involves a number of steps:..deciding how to quantify and compare how well different representations fit the data (that is choosing a score function.)’33

This emphasis on defining goals and creating business contexts runs through much of the operational literature on data mining. That the technology aids institutional cognition, rather than challenge it is illustrated by the following: ‘The relationships and structures found within a set of data must, of course be novel. Clearly novelty must be measured relative to the user’s (the corporate operative) prior knowledge (my emphasis). Unfortunately few data mining algorithms take into account a user’s prior knowledge. For this reason we will not say very much about novelty in this text. It remains an open research problem. Whilst novelty is an important property of the relationships we seek, it is not sufficient to qualify a relationship as being worth finding. In particular, the relationships must also be understandable.’ 34

This technical fakery is acutely dissected by Simon Schaffer in his essay OK Computer a brief history of what Schaffer calls cerebral metrology, or the measure of intelligence. His key insight, quoting Hugh Kenner, is that ‘..authentic human capacity and specifically mechanical capability develop in tandem.’ 35 Schaffer’s consideration of the philosophy of machine intelligence has implications for our analysis of marketing technology. He goes on to describe Babbage’s party trick of programming his calculating engine to, at a predetermined moment, suddenly break from the increment of integers from zero to one million, to advancing in ten thousands. This discontinuity came as a surprise to the observers, understanding as they did machine’s original procession as law-like, but not the transformation, which of course was prior programmed by Babbage. We see a similar sleight of hand with the issue of novelty in data mining, where newness can only, unlike the claims of the propaganda, be defined prior to computation. ‘Aping the street hucksters and wizard impresarios, Babbage’s house party tricks, Jevon’s logical piano and Loebner’s e-mail Turing tests are, precisely, bits of showmanship designed to insinuate through their histrionics the humanity of machinery and the machine-like aspects of human behaviour.’36 For Schaffer these shows represent rival modes of the representation of human and machine capabilities. In constructing machines to ‘know’ people through inflections of statistics and, in the end, fairly crude patterns of probability, we shift our expectations of what it is to be human.

In CRM we find a hollow flattery of a customer’s individuality, purporting to offer something for ‘you’, when in fact it disguises that ‘you’ are offered the familiar products of mass manufacture, but in a way that increasingly limits choice and maximises expenditure. Furthermore, the de-limited ‘self’ these technologies ‘know’ is one the individual is invited, obliged even, to inhabit.



Next