Collaborative CRM: The Kaleidoscope Principle


Do you have a loyalty card? The chances are that if you are an adult living in the UK you will have at least one and some people have 7 or more. The UK’s most popular retailer loyalty schemes are now over 10 years old and the majority of UK adults have at least one card. In Europe as a whole, more than 112 million retailer loyalty cards have been issued and cumulative investments made in the people and IT to support and leverage these schemes runs into many thousands of millions of pounds.

But at the same time, it has been reported widely that 80% of the expenditure on CRM systems has failed to provide an adequate return. So are these retailers just locked in a game they can’t escape from, or do successful schemes really pay back?

The fixed costs of running these schemes are very high. If a large UK high street chain store wished to set-up and run their own scheme they could easily find themselves spending well in excess of £5m just to buy the IT involved. This is before the costs of incentives, consumables, and on-going customer support. Cheap it ain’t.

Sharing facilities can reduce these costs, and companies like Catalina, Loyalty Logic, and RMS provide such services. Alternatively retailers can join coalition schemes like Nectar, which enable them to share the administration costs with other retailers whilst in principle gaining customers through the collectively higher scheme membership.

But running costs are still significant and for many retailers the gain in customer can often be insufficient in itself to pay for the cost of the scheme.

For the large supermarket retailers, having a loyalty scheme in place can be worth around 5%-10% on total sales, and that’s before using the information they gain from it to leverage sales performance through other incentives or by improving the overall store offering facing the consumer. For retailers like Tesco this means the scheme can pay back but for smaller players this is not necessarily so. In the UK, for example, Safeway dropped their scheme some years ago as the immediate financial benefits did not justify the on-going costs and they were not willing to make the investments in analytical capabilities to leverage the data sufficiently to find other areas of benefit. They believed they could get a bigger return by adopting a Hi-Lo pricing strategy.

For the smaller players the benefits have to be derived from a mixture of directly increased customer loyalty and using the data to change the store offer to make it better meet customer needs. It also means using it to communicate effectively both with its customers and internally with the trading teams. If this is not done, or not done well, perceptions arise that the company is not using the data which take root amongst the company’s suppliers and customers, and become very difficult to change.

For example, one major UK grocery retailer did not shrink from making the investments necessary in data collection and in data analysis but it did shrink from the very high cost of mailing all its customers on a regular basis. This cost was big. After two years the company had around 9 million cardholders so communicating by mail with all of them 4 times a year, at typically around £1 per cardholder per time, would have been an on-going cost of £30-40m per year with no guarantee of a return. Not easy for any Board of Directors to sign-off, no matter how big the company.

Retailers can offset some of these costs by obtaining “supplier funded” mailings from manufacturers, but unless the trading teams embrace this wholeheartedly, and unless manufacturers see advantages in communicating with the customers of individual store chains on a one-by-one basis, then the benefits will always be constrained.

So what of the manufacturers? These days you will be hard pressed to find a manufacturer who does not think that they need their own consumer-CRM database to communicate with their most valuable customers. But if you look more closely at what many are really doing, much of it is actually direct marketing (i.e. targeted couponing) rather than true Customer Loyalty Management. There are typically no long-term incentive schemes for consumers, and audits have shown that because the original data was often gathered from sources such as coupon redemption, very high percentages of the current records on such databases are either wrong or incomplete.

Coping with such issues was one of the drivers behind the establishment of manufacturer consortiums such as Jigsaw and the recent decision by P&G to consolidate its European CRM programme into a single outsourced operation in Belgium. Even so, the economics of these “go-it-alone” consumer databases will remain questionable. For example, it was widely commented that the decision early in 2003 to link Jigsaw with OgilvyOne was due to its inability to provide an adequate return to its original consortium members.

The problem is that, unlike retailers, the gains that FMCG manufacturers can make by mining this data are limited. The data is not transactional in any true sense, capturing instead primarily static information. Such information is fine when targeting new products but for existing well-established products the benefits are less obvious.

This is because even when the costs appear bearable, the basic hypothesis behind the manufacturer establishing the database – namely, to better manage and understand their most profitable customers – is far from satisfied. Scan through a few issues of any direct marketing magazine and you will soon find a manufacturer proclaiming their database of thousands or even millions of valuable customers. However, in reality these lists represent only a tiny minority of the company’s or the product’s actual consumers, and very often they are the most price sensitive ones who responded to a promotion, making them very far from being highly valuable brand loyalists.

London brewery Fullers used to run a 5000-strong Fine Ale Club to whom they sent regular communications. In addition, competitions and coupon sources added several thousand further names. However, Fullers is a nationally distributed brand with a good reputation amongst ale consumers and simply by counting the number of cardholders who we know have bought Fullers beers in one major supermarket chain, we identified it has an off-trade customer base of at least 260,000 from that retailer alone. That’s 50 times bigger than its own database, and as someone else is bearing the cost of updating and managing it, it’s a shame not to use it.

My Customer is Your Consumer

Most retailers with loyalty card schemes have segmented their databases according to their customer’s needs, as revealed by the mix of products they purchase. Labels such as “Traditionals”, “Foodies”, “Organics”, “Price Sensitives” and so on are used to summarise customer needs so that suitable offers can be developed for them. And the major food chains have moved on to targeting these customers on a one-to-one basis, with Sainsburys, for example, now sending its customers special “treats” at times such as their birthdays and in recognition of their continued store loyalty.

Manufacturers, by contrast, seem on the whole to continue to see these retailer databases as a means for distributing targeted coupons. They believe that their view of the consumer is holistic for their brand and that their objectives are better served by communicating direct with their own consumer base.

For example, both P&G and Unilever have recently announced significant investments in the production and distribution of consumer magazines that promote their products to their own consumer databases. The aim is to promote a variety of brands and to use the vehicles to recruit names to their marketing databases. Obviously the cost of doing this is huge but just imagine how much more powerful these pieces of activity would be if P&G & Unilever could also track the impact they had had on subsequent purchase behaviour. By using retailer loyalty card data they could achieve this gain.

For many years manufacturers have complained (with some justification) that retailers had cajoled them into spending ever increasing sums of money below-the-line. So much so, it seems they have taken the message to heart! This is surprising because the costs of running these databases, producing and distributing the communications, and then tracking and handling the responses etc are huge whilst manufacturers have no real idea of how loyal or otherwise the consumers on their databases actually are.

Certainly it is possible to run surveys of the consumers on the database to find out more about their brand loyalty, or to do as L’Oreal have done in the health & beauty category and establish a web site that asks consumers questions about their brand usage and shopping behaviour. These devices are helpful but again only a very small percentage of consumers will ever be reached via these routes and even for those who complete the surveys the information is still claimed behaviour, not actual, is never up-to-date and only describes “what?” they are doing, not “why?”. This limits its use.

L’Oreal Consumer Health & Beauty Website – Tailored by Country

Consider for a moment, however, what a retail loyalty database provides. It identifies the cardholder. It has a record of their purchases on the card at the store that is, by definition, up-to-date. Some would argue that there are pitfalls in the data. They complain that the cardholder is not necessarily the end-consumer. They argue that the data is not a complete record of the customer’s activity as it misses all transactions made elsewhere and is not even a complete record of transactions at the store itself.

These people are missing the point. They are so hung up on the data they cannot see what it is telling them. These missing transactions are just as much a description of the person’s attitudinal interaction with the scheme, and with the store, as the transactions they do make. Studying what is not purchased is often more valuable than studying what is, because it identifies opportunities for cross-sell and up-sell. Moreover, it identifies opportunities for manufacturers and retailers to work more closely together for mutual benefit. Even where the cardholder is not the end-consumer, one can make a reasonable assumption that some exchange must in all cases take place between them otherwise the purchases would cease. This exchange can therefore be leveraged.

Manufacturers should therefore be clambering over themselves to help retailers mine this data for insights that will enable them to devise promotional mechanics that simultaneously require consumers to be loyal over time to both the brand and the retailer in order to obtain the incentive. Instead they remain fixated with coupons.

Devising such mechanics would have numerous advantages for both the manufacturer and the retailer. To start with, manufacturers should be seeing retailers with loyalty card schemes as outsourced agents, responsible for managing a part of their consumer database on their behalf. Unilever claims that it has a UK consumer database of over 10 million names. Moreover, when announcing the start of a recent major initiative for Birds Eye, again a multi-product campaign, it set one of its key aims for the programme as “…to build a large core of loyal customers.” Yet the activity was centred around 150,000 consumers drawn from two million who had previously completed questionnaires, so the database was static and would have needed costly validation. Yet between them Sainsburys and Tesco alone can access 15m active and potential frozen food consumers, and their data are complete and constantly refreshed. All it would take would be for the two sides to work together to devise a scheme that benefits both of them and the consumer continuously over time. These perspectives are really worth manufacturers considering more fully.

In return for incentivising consumers to use one channel the retailer would need to allow more branded communications, closer to, for example, print advertising perhaps, but they would still hold the actual names and addresses of their customers confidentially. If you think about it, there is actually no need for a manufacturer to personally know the names of their consumers. All brand communication can be completed through the agent, as it has been for at least the last century and a half.

The manufacturer can then utilise differing retailers in the same way that they utilise differing media channels in order to gain coverage of as many consumers as possible. The losers in this approach are the stores without customer tracking schemes who will suddenly find that their marginal consumers are now much less likely to visit their store. Manufacturers could set-up separate “collectable” promotions for these stores but the benefits would be untargeted, costly to administer and probably less effective.

For the smaller retailers who have loyalty cards, working cooperatively on the scheme with manufacturers therefore makes sense. By identifying “anchor” products for each of their key consumer segments across a number of product categories that have strong consumer pull (i.e. what some retailers refer to as “destination” products) and strong “halo” sales effects, they can work with manufacturers to create incentive schemes which work for both retailer and manufacturer. The result will be increased category value for the retailer and manufacturer, increased brand loyalty for the manufacturer, increased store loyalty for the retailer, and a more satisfied consumer.

Finally, for both sides there is the opportunity to spread the costs of both their CRM programmes. The retailers have already invested heavily in the IT required to communicate with customers and they need to maximise the returns that they get. Manufacturers have the ability to leverage this communication channel to the hilt to get their messages to the consumer whilst minimising their financial outlays and liabilities by paying only for the level of use they make of it when they make it. The paradigm shift they need to make is to stop thinking “direct mail coupons” and start thinking more creatively about how they can use the retailers technology to drive loyalty jointly to the store and to their brands on an on-going basis.


At this stage you may well be thinking that this sounds fine but everyday practicalities may prevent such a vision coming to fruition. For example, the funding for these schemes will almost certainly need to come from existing short term price discounting. The creativity of the offer will need to overcome the transfer of, say, one 3 for 2 or extra value promotion per year which is aimed at all customers but which does not generate long term loyalty into a targeted long term incentive scheme which does.

This will require joint data analysis by the retailer and manufacturer, both to identify the best promotional mechanic and to identify which consumers to target. This is a specialist task, as it requires bringing together data from both the manufacturer and retailer on a confidential and independent basis. The retailer and the manufacturer will have to have confidence that the analysts working on the project are doing so objectively to serve both their interests with equal vigour.

This is because market research information and the manufacturer’s own database of consumers need to be merged with customer transactional data from the retail loyalty scheme in order to derive the product areas to focus on and the algorithms needed to select the right cardholders to target. Manufacturers typically segment the market using consumer attitudes whilst retailers typically use shopping behaviours. Bringing these two together in a way that will achieve the necessary win-win requires joint understanding of the drivers of both store behaviour and brand choice.

Knowledge of how to do this exists at research agencies such as RedRoute and we also have the necessary independence and confidentiality to be able to provide the objectivity and vigour. For example, data fusion and data projection methods can be used to combine market research data with transactional data. From this we can understand the motivations behind actual consumer purchasing and our research has shown that behaviours can be accurately predicted by bringing together knowledge of consumer attitudes and consumer circumstances. This enables us to uncover which products or categories and which consumers should be incentivised in order to create increased loyalty and profitability for both the manufacturer and the retailer.

Our model then evaluates each potential proposition on 5 key dimensions: relevancy, brand identification (for both retailer and manufacturer), accessibility, value and consumer confidence. The latter aspect is important because it really reflects the level of equity in the manufacturer’s and the retailer’s brands that exists amongst the target group. For example, McDonald’s may be a perfectly acceptable location to buy a hamburger but may lack credibility as a location for buying a fresh salad.

The model therefore enables propositions to be developed which have the right mix of these dimensions in order to create the maximum sales effectiveness from the scheme. An example of the relative difference can be seen in the diagram below which is taken from our work in the DIY sector. The original promotional mechanic was heavily value-based (money-off), had no repeat purchase component and was only targeted at the most recently active customers. The new promotion had a slightly lower level of discount, was targeted at customers who would probably be most interested based on their attitudes and circumstances, and rewarded repeat purchase behaviour.

Benefits from Using Motivational Modelling to Develop More Effective Promotions

The net effect was a 50% increase in sales effectiveness for the same promotional cost. Moreover, the people attracted were far more likely to continue buying both the brand and shopping at the same store as a result of the activity. A win all round.

Retailers may be concerned that this is not scaleable across all their suppliers but for two reasons this is not a real issue. Firstly, the “anchor” products and categories are going to be those only analysed by category champions and, even then, only when they appeal to a sufficient critical mass of consumers to make the effort worthwhile.

Secondly, by using Motivational Modelling the diagnostics identifying how to proceed are easily repeatable and the costs can be tailored to meet the needs of the customer group. This would mean, for example, that a smaller manufacturer who is a category champion within a niche category such as organics would not be excluded from participating, and nor would secondary manufacturers in any other category.


Manufacturers should be taking a far more active role in putting pressure on retailers who have customer-tracking information to enable them to use that data to jointly drive brand and store loyalty. The benefits for both parties are real and substantial.

This requires a change of thinking within manufacturers whereby they now consider that retailers are the custodians of differing sectors of their own consumer base and that the retailers own the technology to enable manufacturers not only to communicate with their consumers but also (and unlike media owners) actively and continuously manage their loyalty to the manufacturer’s brands and to the retailer.

There is no need for mass-market manufacturers to personally know the names of their consumers and they will never have databases that are either comprehensive enough, or up-to-date enough to be able to use them to really drive consumer behaviour. They should not, therefore, be wasting millions of pounds investing in universal one-to-one marketing technology. It would be far more efficient, both for themselves and the retailers, to utilise retailer customer contact systems in ways that benefit both parties. Not doing so misses a profound opportunity for both of them that is truly global, not just a UK phenomenon.

By bringing this about manufacturers will be able to utilise their own consumer CRM databases far more effectively, concentrating their resources and costs on reaching the consumers that cannot be contacted via the retailers and using them as “hubs” from which to manage their use of the retailer databases. The diagram below shows how this works and we call it using “The Kaleidoscope Principle” for obvious reasons.

The Kaleidoscope Principle

For retailers faced with falling customer loyalty and the need to maximise the return from their CRM investments, manufacturers represent a known source of revenue to help achieve both these objectives. However for their part they must adopt a more cooperative approach and work with the manufacturers to develop more corporate strategies for customer / consumer management which achieve both sides objectives, not one or the other. Motivational models of customer behaviour already exist to enable this to be achieved that use these data sources intelligently and without undue reliance on unstructured data mining techniques. The latter typically uncover operational efficiencies but rarely identify new creative opportunities.

In sum, manufacturers and retailers are standing on the threshold of a potentially enormous leap forward in their joint ability to identify and meet consumer needs and it is in the interests of both parties, and especially the mid-sized retailers, to make this happen. The benefits are the ability of manufacturers to connect with individual consumers and be able to position their brands relevantly in the minds of each of them. The retailer’s customer segmentations are far more reflective of an individual consumer’s overall needs than are the manufacturers attitudinal segments but merging the two through a Motivational Model identifies what to say to whom, when and how, one-to-one, to make both brand and store relevant to each and every one of them.