The services sector is exploding with over 150 million customers. Keeping them happy is a herculean task. Consider the predicament of leading retail chain, Shoppers' Stop.

Consumers who walk into its outlets want clothes of different sizes, styles and colours and tailor-made promotions and schemes, unique to their buying patterns. They expect the retailer to divine their needs.

"For this we need to understand how, when, where and in what combination the customer buys merchandise," says Unnikrishnan T M, customer care associate and chief technology officer (solutions and technology), Shoppers' Stop.

Retailers in India broadly depend on three types of intelligence -- merchandise (everything from product inventory to price and marketing); customer (knowing buying patterns to ensuring customer loyalty); and operational (which ensures cost management and includes reducing retail shrinkage due to theft by employees, shoplifting by customers or vendor fraud).

In the case of Shoppers' Stop, it uses a combination of business intelligence (BI) solutions by companies such as Business Objects and SAS to decipher consumer behaviour patterns. BI tools help retailers to, among other things, drill down to see how different styles, sizes and colours are selling.

This, of course, is simply a case in point. BI solutions slice, dice and interpret the data for managers.

With today's BI tools, employees (not just the senior management) need not wait for the IT department to churn out reports. BI tools are being used by the banking and financial sectors, retail outlets, call centres, shipping companies and even hospitals to increase consumer satisfaction, which, in turn, adds to the top and bottomlines.

FROM BI TO AI Experts predict that in the near future, call-centre applications will be embedded with intelligent agents (read: Artificial Intelligence or AI) that will gauge the anger level of callers going through voice-response prompting systems. Moreover, systems would play subliminal messages to calm the caller while alerting call-centre workers what to expect before they pick up the telephone. AI has been around for the past 50 years, but extracting data from rudimentary databases has always been a challenge. AI got a boost from the concept of neural networks that are modelled on logical associations made by the human brain. Says K Chokshi, CEO, Intelligent Business Systems (IBS): "A neural network can recognise relationships and spot trends in huge amounts of data that wouldn't be apparent to humans. It's much easier and faster for a computer to determine, for instance, a pattern to security breaches of a company's computer system." In 1998, the US-based Computer Associates was the first to commercialise neural networks in software called Neugents. Chokshi's company develops software systems that can learn from experience, acquire knowledge from data and adapt themselves to new environments. "The combination of statistical, neural and fuzzy methods now enables direct quantitative and qualitative studies to be carried out," says Chokshi who has been working with, among others, the Cyprus government to create an AI-based software to provide risk management for the tourism industry.

Consider how HDFC Bank [Get Quote] manages to keep its approximately 10 million customers satisfied. The bank can, for instance, select the profile of a high-end customer, have Unica (a marketing automation solution) search for matching profiles from the database and automatically send an email to such customers.

Thus, marketing analytics enables the bank to measure the efficacy of the campaigns, experiment with creatives, messages and media.

The bank's vice-president and head, marketing, Ajay Kelkar, maintains that companies should have an "information-led" strategy and use information as a differentiator.

"In the Indian context, there has been a wealth of customer data in the past three or four years. You can address customers by name, understand their behaviour patterns and target them with focused promotions," he explains.

"Nowadays, with marketers also interacting with CFOs [chief financial officers], measurability becomes a key factor. We can measure the result of a campaign through the sales it generates and the customers that are added. There is a clear return on investment [ROI]," he adds.

Marketing analytics campaigns have reduced the cost of customer acquisition by 25-30 per cent in select campaigns, reduced customer attrition rates and increased the ticket-size spending of customers as a result of focused communication.

These campaigns also enable the bank to cross-sell to customers, thus increasing their product holdings and consequently, customer profitability.

IT's a business strategy

Today, BI has become an inherent business strategy rather than a mere IT application. Frost & Sullivan estimates the total Indian BI market size in 2005-06 at a little over $47 million.

Major players include SAS, Business Objects, Actuate, Cognos, Hyperion, Microsoft, Oracle, SAP and Terradata. BI analysis can inspire companies to launch new business ventures or hive off a business division.

BI tools find applications in any sector. Take the case of the Hinduja Hospital in Mumbai, for instance. Pramod H Lele, CEO, P D Hinduja National Hospital & Medical Research Centre, says BI can help doctors find out what kinds of diseases are on the rise.

Other tangible benefits cover issues like "Is the bill ready when the person is discharged" and "Did he get his laboratory reports on time".

Adds Lele, "One can analyse the pattern of medical reports, find out which drugs are used most and during which season. This data can help doctors by increasing the stock of the required drugs during these seasons. Many such conclusions can be arrived at. This is where analytics helps us."

It helps even in the shipping sector. Changes in oil prices, fluctuating exchange rates, different vendors, geographical locations, and products to be shipped all have huge impacts on shipping lines.

When Great Eastern Shipping decided to bring analytics on board, the first task was to make databases "talk" to each other. "Ships are moving targets. Our BI solutions has helped us forecast various scenarios and have solutions in place beforehand," says R P Dumasia, general manager, IT, Great Eastern Shipping.

What's needed

It has been estimated that the amount of information in the world doubles every 20 months and the size and number of databases increase even faster. Hence, properly integrating BI with operational applications remains a challenge.

BI applications should also be compatible with the existing enterprise resource Planning, supply chain management and customer relationship management applications, failing which enterprise application integration tools and middleware applications can be brought in to resolve the issue and harvest the data into the data warehouse.

Second, BI solutions generally cost upwards of $10,000. So, it's best to clearly define the ROI even before you start. Outline what you expect to achieve, then do regular reality checks. The ROI depends on the application for which the BI is used: some areas pay rich dividends.

"Birthday mailers are sent to loyal customers and the response has been phenomenal. We also use BI during the festival season to increase traffic to the stores. Here, too, the results have been encouraging," says Unnikrishnan.

Shoppers' Stop uses the information and analysis from the BI tools to help it study the demographic details of the target markets and position merchandise in retail store locations.

The organisation has also used various predictive methods and models to find out which of its customers are likely to buy again and how much. This has helped create focused campaigns and incentives for its high-value customer base.

An important point is that although the idea of BI for the masses has been talked about for several years, in most companies it is still deployed to only a small percentage of employees.

Companies also need to understand the different ways people work and fit into office applications and e-mail. In a recent study by Ventana Research, e-mail was the most important means of information redistribution, cited by 74 per cent of those surveyed.

Operational BI applications must be integrated with e-mail to facilitate joint management of goals, forecasts and plans. The most important ad hoc capability, cited by 81 per cent of those surveyed, was the ability to export to Excel (or a spreadsheet), notes Ventana.

One way to achieve this: deliver reports already in Excel formats to assure consistency of use and re-use.

"BI is not magic. Nor is it a mathematical model. It is getting the right information when you want. It's better to start small. BI is not instantaneous. And it' is never over," cautions IT consultant Raj Chaudhuri.

"We use extensive qualitative and quantitative evaluation criteria. Every solution is graded on these criteria," emphasises Unnikrishnan.

IDBI Bank has its own internally-developed framework for BI. The bank implemented BI in a phased manner by first studying the customer segments the bank wanted to target and then sifting the profitable customers from the database. It reportedly took IDBI around six to eight months to create the BI framework.

BI can also help companies understand their customers better, which, in turn, can help in better decision making. But there's no right strategy, as Dumasia puts it.

"Two years ago, we had a centralised management information system. Now we have moved towards analytics. At the end of the day, analyses of data is not just a spreadsheet," he cautions.

Where previously, only the top management would use such tools, they believe that this is now filtering down to lower levels, such as information workers and managers.

In a recent worldwide survey by market research firm IDC, some 70 per cent of firms polled say they plan to increase the number of internal users of BI tools over the next year. Let's hope companies can use that intelligence to retain existing consumers and new ones.

HOW TO GET BI RIGHT Define ROI clearly before you start. BI solutions generally cost upwards of $10,000. Outline what you expect to achieve (live reports or will 90-day reports do?), then do a reality check every quarter or six months. BI vendors position technology ahead of management intent. That's wrong. Don't buy BI software because you think you need it. Deploy it with the idea that there are numbers out there that you need to find, and know roughly where they might be. BI assumes clean data; take stock of the data you have and ensure it's clean. Spend time to analyse the data collection process and correct it if necessary. Train users effectively; ensure management commitment to make it happen.