The quest for customer loyalty finds an ally in AI
24 Feb 2007
Every marketer's dream of attracting customers with unflinchingly loyalty comes true… with artificial intelligence, says Dr Kaustubh Chokshi, CEO of UK-based Intelligent Business Systems
It''s a fickle world out there. In an age when people change jobs and partners almost as often as they change hats, it''s small wonder that every marketer dreams of attracting customers who will be unflinchingly loyal to the products and services of his company. Finding, attracting and retaining such loyal customers in a competitive and crowded marketplace are an ongoing challenge. While customer satisfaction is the basic starting-point for retention, full-fledged loyalty goes far beyond — it cannot merely be bought, but rather, has to be earned.
The relationship between the 21st century customer and the vendor is a complex one. From basic marketing research, promotions and sampling, marketing methodology has evolved to encompass more innovative ways of capturing customer loyalty. In the retail trade particularly, there has been a rising trend towards loyalty card programmes, in the hope that such schemes would guarantee customers return to a particular outlet or chain rather than an turn to the competition.
Loyalty cards — a ''hygiene factor''
Loyalty programmes in some form or the other (such as saver stamps, card or sticker series collections, etc) have existed for centuries. The electronic version, based on smart cards, became popular just over two decades ago, with frequent flyer programmes offered by airlines. It soon expanded to cover the hotel industry, followed by most other major industries catering to individual consumers. Now there are programmes for frequent flyers, frequent buyers, frequent diners, frequent players, and myriad other schemes. But, in essence, most of these schemes are mere electronic replacements of their paper predecessors.
A study conducted by the Gartner Group revealed that American firms spent over $1.2 billion on loyalty programmes in the year 2003. Jupiter Research says that over 75 per cent of consumers have at least one loyalty card, and a third of the US shopping population has two or more. Other surveys by Forrester Research and the META Group indicate that spending by organisations on loyalty programmes have exploded in the last few years and show no signs of abating in the near future. As the organised retail sector expands in India, so will the number of loyalty cards that consumers carry around in their wallets.
But is there anything that differentiates those cards from each other? At the outset, when loyalty programmes were not widespread, they did have considerable utility, because they rewarded high-value customer and actually built loyalty around customer bases. But ultimately, competitors also cottoned on, and loyalty programmes became synonymous with point systems. There has been little or no concerted effort to understand the individual needs of the consumer; it''s a one-size-fits all attitude.
When nearly everyone in the marketplace offers a points reward system, the system itself no longer remains a differentiator. In the airline, hotel and organised retail sectors, point systems have become a cost of doing business, a hygiene factor--having one is not a differentiator, but not having one is damaging. Whether such systems actually build loyalty is anybody''s guess. Further, each organisation needs to ensure that the customer carries its loyalty card in preference to those issued by the competition.
Dynamic customer relationship
It would not be an exaggeration to state that the days of traditional loyalty card programmes based purely on reward points accumulation are numbered. Not only is the cost of sustaining such programmes increasing, but also the impact on the sponsoring organisation''s business model and bottom line are questionable, to say the least.
It has now become clear that rewards and benefits that are not personalised (hence often unusable) cannot cultivate loyalty; they may instead only serve to irritate or anger the consumer. For most retail businesses, consumers first select stores driven by requirement and convenience. The retailer, thus, needs to rise to the occasion and make consumers aware that their specific requirements and desires are known. This is possible only if the loyalty plan can be personalised to the customer''s tastes. The best marketers deliver this kind of simple, valuable relationship in the course of daily business, and not as a benefit the customer must "earn."
Any customer relationship management programme that includes a loyalty scheme must hence be able to go beyond customer demographics and total spends in order to develop customer categories. Both known and unknown consumer behaviour and psychographics have a bearing on the segmentation of consumer preferences. Thus the retailer needs to keep a close eye on the performance analysis of the different marketing channels used by the organisation. Further, to make such schemes effective, a complete analysis of existing and proposed product-mix alternatives needs to be made, in order to match promotional pricing and special offers with potential risk.
The primary benefit of a loyalty programme to the sponsoring organisation is not consumer loyalty per se, but the opportunity to build relationships with individual customers in gathering purchase tend and product information, with a view to achieving better focus and targeting, improving operational efficiencies and ultimately altering consumer behaviour for enhanced profitability. For this to happen, the information collected from the loyalty programme must serve as an input to improve and fine-tune core operations such as product selection, displays, advertising, pricing, etc.
AI to the rescue
Of course, with traditional, logic-based software it would be impossible to achieve the above goals and overcome all the prevalent limitations. There needs to be a high level of decision automation in the customer profiling and rewards personalisation process. Further, the profiling needs to be accurate even when there are gaps in the data pertaining to a specific customer; and, the profiling and personalisation need to dynamically evolve as the consumer behaviour, buying patterns and status change over a period of time. All this can only be achieved with software that has Artificial Intelligence technology at its core.
Nevertheless, for any loyalty card system to be truly successful, it needs to have a wide range of data pertaining to a customer. That''s why any loyalty scheme worth its salt ties together many products and services--across brands, industries, sectors and geographies. While the data on each loyalty cardholder is being built up, techniques such as Bayesian inferencing are used to fill in the gaps.
AI software based on neural networks is especially designed to recognise patterns in the transaction data. Over a period of time the likes and dislikes, attitudes and behaviour of each customer participating in the loyalty card programme emerges, and the resultant categorisation is obviously more meaningful than that based only on demographics and spends. This sets the stage for Precision Marketing exercises, which help in the anticipation of the needs and wants of customers, rewarding them for their profitable and desired behaviour. Such a loyalty programme, which delivers relevant and worthwhile communications to customers at precisely the right time in the purchasing cycle, would definitely have positive fallout in terms of customer satisfaction, loyalty and yields.
The sky is the Limit
Once the structure of the loyalty card programme is set to ensure collection of large amounts of consumer data, as well as provide for personalised and relevant rewards to ensure loyalty and continuity, there is no limit to the ways in which the AI system can be trained to automate decisions pertaining to a wide variety of marketing scenarios. For instance, in a supermarket, the raw data can be put through the AI engine, transforming it into information and knowledge that enables sophisticated levels of prediction: The supermarket can gain deeper insights such as — how discounts and promotions impact the sales trends; which items are selling the best in each department; what trends are patterns are emerging that may have an impact on future sales; which marketing channels are more effective; what product-mix works best, etc — and then use these insights for fine tuning the entire marketing operation.
Customers may indeed be fickle-minded in today''s world of myriad options. But an AI-based loyalty card programme that offers