The ad campaign that has your name on it

Marketing: sophisticated databases are transforming the targeting of customers into a precise science

SUPERMARKET shopping trips can be stressful enough as it is. But soon the screaming of children and other noises could be accompanied by the sound of talking trolleys. Tests are under way in the US to assess shopping carts that can - on the insertion of a charge card or credit card - direct the shopper to aisles that stock the items most likely to be of interest to him or her.

For the shops the value of this technique is that it could lure the shopper to the promotions and special offers.

So how does the trolley know a customer will, for instance, be interested in a few pence off a certain brand of coffee or natural yoghurt? Because the card will reveal past spending habits and hence provide what marketing folk call "strong predictors" of future consumption.

Because it is based on fact, its promoters say, this approach is more effective than those that use external sources of data, such as the census, to build up a broad picture of customers. People do not always conform to type, says Judi Gehlcken, chairwoman of the Direct Marketing Association and head of her own database marketing operation, The Computing Group. Under the new system, "Information about sales made to actual individuals replaces suppositions with fact," she says.

To drive the point home, Martin King, who runs the UK operation of MarketPulse, supplier of a leading software package enabling instant analysis of vast amounts of data, tells the story of a seed catalogue company that assumed flat-dwellers were not interested in its products. That was based on census data, but in fact significant numbers in a given area had allotments and were very interested.

With powerful processing systems and a bit of subtlety, "facts" such as this are not so difficult to obtain. In North America, the beer company Labatt has apparently resorted to an "interactive beer can" that gives consumers the impression they are taking part in a quiz as a way of discovering information.

Every time we fill in a bank customer survey, or a store questionnaire, or a card inside a CD asking what type of music we like, or when we tell the holiday booking agent where we saw the advertisement, we are adding something to somebody's computer files. We are often persuaded to give this information because we are getting something - information about products or special offers - in retum.

The most common incentive is the loyalty card, whereby the customer receives a discount or money-saving voucher after so many purchases. The best-known of these schemes is the Air Miles programme, which claims to have given half a million people flights, holiday discounts or weekend breaks last year alone, and later this year provides an update on its sophisticated database. But many stores operate them in one form or another.

The information gleaned in this way is responsible for your junk mail becoming more tempting. Depending on who you bank with, the chances are your statements come complete with inserts that only a few hundred account- holders see. Messages promising such advantages as special deals on life assurance can be printed on statements.

Similarly, those offers of Caribbean cruises - so beloved of the direct marketing pioneers - are no longer sent to couples in their mid-thirties with two children and negative equity. Some travel companies know that their regular users take holidays only at certain times of the year - and send them brochures when they are most likely to be thinking about their holidays.

This is database-driven marketing. And signs are that it is becoming the moving force in the advertising field. The DMA estimates the market is worth about £9bn. Others reckon it could be far bigger, depending on what is included in it. Financial pressure on even the biggest companies' marketing spends is forcing a shift away from grand "brand awareness" advertising of the type that has traditionally graced peak-time television schedules, towards specifically targeted campaigns.

Heinz, for example, created a stir last year when it announced it was increasingly using direct marketing in the drive for better cost-effectiveness. Likewise, Tesco could use its recently launched loyalty card to target its own-brand baked beans at customers known to buy branded beans. Or it could find out why some buy alcohol at the supermarket and others do not.

Indeed, one of the most useful attributes of database-driven marketing is in establishing patterns, according to Mr King. For example, a large retailer could discover that people who buy expensive suits have a high propensity to buy mail-order wine. If it wanted to boost sales of the wine, it could target people buying suits who did not already buy wine. In all likelihood it would achieve a high response rate at lower cost in leafleting or other forms of marketing. "It's turning junk mail into smart mail," Mr King says.

But there are other ways in which the technology can be used to get close to the customer. Chains can also use the database to find out which lines sold best in which individual stores and stock accordingly.

Alternatively, the data can be used to influence strategic decisions, such as the siting of stores or car dealers' service centres. Or it can be used in a public service way, to carry out speedy recalls of faulty products or to handle complaints.

Even within marketing, it can play a complementary role to above-the- line advertising. After all, many television advertisements now carry direct-response telephone numbers. For example, The Database Group, an information technology support company in Bristol that helps clients analyse data, has done some work on profiling for the Volvo car company. It has been building up a picture of the sort of people who might buy a Volvo, so that the advertising agency can draw up a campaign that might appeal to them.

Much of the information being analysed has been available in one form or another for years; it just has not been possible to analyse it in a cost-effective way. According to Les O'Reilly, chairman of The Database Group, many companies are "data-rich but information-poor".

The most notable example of the attempt to remedy this is the way banks - which have notoriously poor images with their customers - are using information derived from questionn-aires to translate all the data they have held on account holders into information about real people.

Now ever more information is becoming available through sophisticated cash desks or electronic point of sale (Epos) machines. As the careful shopper will know, it is commonplace now to have till receipts that detail exactly what you have spent your money on - which brand of beans, book title, whatever.

Such information enables the store to target customers much more narrowly, so as to invite a select group of fans of a particular author, say, to a book signing or a small group of women to a fashion evening. Even comparatively small retailers are investing in this sort of technology in the battle to find an edge.

Whether they will achieve that remains to be seen, but Olivetti, the Italian company that operates many Epos systems, has been sufficiently impressed with developments to take a stake in The Database Group.