A version of this article first appeared in eB-21, published 10 times a year in Europe by TBC Research .Based in London and San Francisco, TBC Research helps senior business professionals make more informed technology decisions through its magazine, research, and events portfolio.
British Airways, despite massive investments in a Web site that receives 30 million hits a month, wouldn't consider it. And even Amazon.com, the Young Turk of the online world and renowned for its technological advancement, is only making tentative and exploratory approaches. But just what is holding back the Internet pioneers from sending out marketing messages to customers in real-time?
Despite using analytics software from White Cross Systems in a wide ranging project to better understand its on-line customers, British Airways.com is wary of fully embracing real-time customer interaction. John Mornement, BA's head of
e-commerce operations, says the sheer amount of integration needed between different information systems made real time to all intents and purposes impossible.
Echoing these sentiments is John McKean, independent management consultant and sometime IT advisor to the US government. While acknowledging Amazon.com's ability to use previous click history to make purchase suggestions to customers as the closest anyone has got to real-time interaction, he adds that it still faced a huge technical challenge. And Amazon has no legacy systems at all.
Real-time marketing requires an almost instantaneous response to a customer interaction. For this to work some level of integration between off-line data, held in a data warehouse, and incoming online data about customer interactions from front-office systems is required. A return link is also needed from front-office information to ongoing marketing campaigns. Many are worried about the effect on their well-trusted existing systems and this can inhibit the degree of experimentation. In addition, companies are worried about pestering and deterring customers by making real-time offers.
Throwing customer analytic applications into the mix can remove much of the guesswork and risk, allowing for greater cohesion for any real-time marketing campaign. Analytics expands on traditional data mining by applying further reporting using information on demographics and customer behaviour. If the analysis is fed back into the real time campaign, it allows you to make the right offer at the right time.
If real-time marketing does not take place within this context it can easily become product push or lead to customer annoyance. Analyst Giga Group states: "Companies must first understand customers and develop customer strategies to determine what type of real-time capabilities are most appropriate." This may seem obvious, but in the rush to exploit the opportunities the customer has often been a secondary consideration and tactics, such as including business rules and protocols to ensure you are not annoying your customers, have often been left out. As Diana Woodhead, director of consultancy group Marketing Best Practice, says: "People don't really have any measure or vision of the extent to which they annoy their customers as well as please them."
The tendency among many enterprises to keep on-line and off-line data separate is based on some well-founded reasons. The nature of legacy databases makes it very difficult to perform meaningful analysis on millions of customer history records in the time required for online feedback, and marketers are understandably reluctant to trust their best customer information to a computer algorithm.
But Peter Dorrington, business solution manager at SAS Institute, warns that keeping off-line analytics separate from the new real-time information means it is applied retrospectively and, to a certain extent, intuitively. Dorrington says bringing the two together allows a marketing manager to be more objective when making decisions, rather than relying on assumptions.
He points to the plague of viral e-mail marketing as an example of the dangers of relying on information which is not targeted: "Bombarding customers in the knowledge that only 0.001 per cent will respond is very flawed. Yes, you can in theory mail millions but you are assuming an infinite market which is infinitely growing when in fact the opposite is happening. Plus, those who don't want the message will simply install rules to block its arrival."
The approach is also intrusive and even marketing guru Philip Kotler has no sympathy for peers dealing in spam. "Every consumer should have the right to have his name and information deleted from the file of marketers in whose companies he has little or no interest. I am not sympathetic with the money this will cost marketers to clean up their act."
A customer backlash movement against this very modern plague is already surfacing in the US where customer intelligence is being embedded in e-mail applications which automatically deletes a message if it is unsolicited
This is where the value of real-time or near real-time marketing can be realised. But the approach to date has often been flawed. The tendency to ask a list of questions in which to collate information, is increasingly seen to be of little value. The questions can be poorly designed and people are certainly not averse to lying on such surveys. The key, says Dorrington, is to not to ask questions but rather to steer customers, depending on their behaviour, in certain directions. "Customers are far happier with this proposition," he claims.
Real-time analysis of trends gives a company the opportunity to present real-time options, while analytics can help a company identify characteristics or actions that spell out customer interest. This in turn can feed into the formation of business rules that can be applied in real-time allowing a better interaction with customers by monitoring such things as abandoned shopping carts, e-mail or survey response and changes in account balance.
US retailer Target store has a sweepingly ambitious plan to implement real-time marketing. The company is part of a Compaq-inspired pilot called the zero latency enterprise, which attempts to integrate real-time transactions with online processing. This means whether a customer is shopping in a store or online any kind of business intelligence could be applied with instantaneous results. For example, a retail point of sale could easily spot an attempt to use a stolen credit card or a distributor could be immediately notified of plunging stock levels. At any point in time a customer's needs can be responded to.
Following a 20 year partnership with Compaq, Target store has announced it will adopt a new architecture and a real-time server which caches all application data and creates a central repository of business intelligence. It's a hugely ambitious project that, if successful, promises to revolutionise the business. Watch this space.