Artificial intelligence and the art of keeping the customer satisfied
30 Dec 2006
As differentiation becomes increasingly difficult in a congested marketplace, vendors are turning to innovative IT solutions based on cutting-edge artificial intelligence technology to execute their precision marketing strategies and capture the minds and hearts of the discerning consumer. By Dr Kaustubh Chokshi, CEO of Intelligent Business Systems
India
is today witnessing a consumer boom that''s unprecedented
throughout its vast and glorious history. Steady and sustained
economic growth has slowly but surely roused the burgeoning
and upbeat middle class into uninhibited spending on a
wide range of goods and services. Disposable incomes have
increased, and, in a break from tradition, so has the
willingness of consumers to dispose of those incomes.
In such a scenario, it is not surprising that vendors are falling over each other in their eagerness to attract and retain customers by any conceivable means. From banking to telecom, organised retail to real estate and beyond, everyone''s out to attract customers, keep them happy, and ensure their steadfast loyalty.
Marketing
Myopia
With the fight to retain customers and expand markets,
terms like "customer-centric approach", are
turning out to be adopted endearingly, in the hope belief
that words alone would widen the revenue base and maintain
optimum profitability. One would assume that these intentions
could easily be actioned, given the expanding markets.
If markets are growing, so is the competition. Moreover, vendors operating in such a dynamic marketplace need to constantly respond to the new opportunities and threats that are continuously emerging. Alliances are being forged and broken and these further affect the dynamics of the market. The irony is that organisations in a similar domain all have similar underlying goals as their competitors; share a largely overlapping customer base (albeit expanding); and, operate in the same business and economic environment.
With the increased complexity that this brings about, business agility and flexibility take a beating. Creation of new customer value is difficult, as differentiation of the organisation and its products is difficult. In their desperation to stay in the market-share race, most organisations resort to "brute force" marketing, wherein every existing and potential customer is bombarded with all possible cross-selling, up-selling and general promotional offers under the sun through every conceivable marketing channel.
Not only is this a costly proposition for the organisation, it could result in an exodus of confused or irritated customers, while at the same time diluting the message intended for the actual target audience.
The fact of the matter, in most cases, is that customer relationship management is yet to reach any significant level of sophistication that would ensure that the "customer-centric approach" is anything beyond mere imaginary marketing hype.
But, organisations can continue to operate in this mode only to their own detriment, and ultimate demise. Unquestionably, ''precision marketing'' - in essence creating the right product-customer fit — is the biggest challenge being faced by marketers today. Traditional methodologies and strategies fall short, because when it comes to the accurate customer profiling required for precision marketing, they can only deliver profiling that''s based on demographics; or other static rules; or, at best, based on historical spends and loyalty.
Where
AI Comes In
That''s why organisations today are turning to software
that''s based on artificial intelligence (AI) technology,
where a high level of decision automation in arriving
at the right product-customer fit is facilitated. A carefully
designed AI engine, customised to the precise needs of
a client - be it a banking major or a chain of retail
stores — can be trained to offer true decision automation,
for instance in the selection of a target market from
an existing or potential customer base, with razor-sharp
profiling based on a dynamically determinable parameter-set
and action-set.
Data from the data warehouse is selected and cleaned and goes through a first level of pre-processing transformation, during which the data is normalised. Using data mining techniques, patterns in the data are established. At this stage the AI engine takes over in order to automate decisions based on the knowledge generated from the raw data.
Fig 1: From Raw Data to Decision Automation
AI-based solutions can be designed to perform context-sensitive customer acquisition, behaviour analysis, cross-selling, up-selling and retention programmes, thus enabling multi-product, multi-channel organisations to drive more efficient, cost-effective and profitable customer interactions.
With AI software, decisions and predictions can be based on actual historical data. First the software establishes, or helps establish, the criteria to make predictions from existing sets of data. Next, it generates scoring algorithms for assigning weights to different data characteristics.
Finally, it segments populations into sub-groups, enabling variable treatment for each business decision. This is predictive analytics in action — predicting likely future results from the patterns found to be prevalent in historical data. Most AI software has neural network technology at its core, wherein the system is "trained" using historical data and expert opinion to detect hidden patterns and produce the desired output.
Automation of a decision involves a derivation of the decision criteria and the actions to be taken when these criteria are met as well as when they are not met. After the formulation of criteria and actions, they must be incorporated into the system in a manner that makes them available to other relevant modules and systems as well as to non-technical users via a user-friendly interface.
The heart of analytic modelling is a mechanism to transform all variables and equations to code in such a manner that new transactions provide inputs for further predictive measurements. In addition, the sequence and timing of rule execution is of prime importance, in accordance with inputs and computed data values.
AI
in Action
In simple terms, Artificial Intelligence software facilitates
the process of transformation of the raw data that an
organisation collects from its various operations / sources
into usable information and then converts that information
into smart decisions.
Take a supermarket, for instance. Several thousand transactions are recorded at each checkout counter every day. In its raw form this transactional data provides only basic levels of information, such as what items were sold, when they were sold, and the price at which they were sold.
However, using AI-based software, that raw sales data can be transformed into actionable information, enabling the supermarket to gain deeper insights into marketing strategy and consumer behaviour. How are discounts impacting sales trends? What are the best-selling items in each department? How does shelf-positioning affect sales? What trends and patterns are emerging that might have an impact on future sales? Which marketing channels are more effective? What product-mix works best? Obtaining answers to these and other similar questions could provide valuable insights for fine-tuning the overall marketing operations.
Armed with such predictive analysis and decision automation, the supermarket management can better plan for the future. By predicting buying trends and consumer behaviour, inventory control can be totally streamlined. Moreover, insights into frequently clubbed purchases in the past (such as potato chips and salsa dips) may enable better layout of the store shelves to increase revenues. Promotions that do well in impacting sales in a test location can be replicated across the chain in order to boost sales and profitability.
From the consumer viewpoint, if the entire system is linked to a loyalty card scheme that is actually based on buying patterns in the past, then special offers, cross-selling and up-selling deals can be tailored to the needs of the individual consumer (or groups of consumers), thus enhancing customer satisfaction considerably and automatically increasing yields.
Conclusion
As is evident from the above example, a well-designed
AI-based system can empower an organisation to offer the
right value proposition through the right product mix
to the right consumer at the