Perhaps the greatest shortcoming of Business Intelligence (BI) technologies such as data warehousing and data mining is that we often fail to make the most of their potential. When fully utilized they can be powerful tools for improving customer loyalty, customer service and resource allocation. They can even alter the way a company does business, offering previously unconsidered solutions and strategies based on in-depth analysis.
However, learning to fully exploit these technologies requires that employees know more than basic ‘how to’ skills, such as what keys to press or how to set up a query. It means educating them to ensure that any initiatives are viewed not in isolation but as part of the overall business strategy. As well, employees must understand how to effectively use the vast quantities of information that BI solutions provide.
To this end, it is imperative that knowledge workers be educated on the three key steps of business transformation, in-depth analysis, and decision optimization, and how they relate to the enhancement of a data warehouse.
Step One: Business Transformation
Data warehousing and data mining initiatives may need to go hand in hand with significant re-engineering and process change. Business transformation allows an organization to reap maximum advantage from a sophisticated BI solution.
For example, banks tend to organize themselves along traditional lines. Typically, they have different groups of employees focusing on maximizing profits from such areas as chequing accounts, credit cards, estate trusts, mortgages and consumer loans. The result is that each branded product area thinks of itself as a silo, distinct from the others. While each department is intent on maximizing its own profitability, their parallel but separate efforts can cause overall bank profits to fall short of full potential.
To help rectify the situation, many organizations turn to service providers to analyze the situation and help develop a solution. Often the first step has very little to do with computers or software. At the outset, consultants may work with a client to determine if there is room for improvement in the way people make business decisions. In those first steps the consultant functions more like a process engineer than an IT consultant. IT comes later. It is at this point that business transformation is suggested.
Business transformation takes a ‘whole relationship’ approach. A traditionally organized bank, for example, often views its customers as individuals using certain of the banks products (e.g. a credit card, a chequing account, and an auto loan). Despite having this valuable cross-section of data, the bank typically fails to maximize its relationship with that person because its marketing efforts originate from several single-purpose product silos. In a ‘whole relationship’ approach, the individual or Valued Customer (V.C.) and his or her family are looked at comprehensively, as they relate to the full service spectrum. This gives the bank a chance to cross-sell a product such as a home mortgage, or up-sell a product such as a gold credit card.
This approach demands substantial change. For example, the bank could offer a free chequing account if the V.C. accepts a gold credit card. Experience says that the V.C. will spend more on a card with a higher credit limit, enabling the bank to come out ahead overall. However, unless reforms reflect process changes, the Chequing Department will show reduced profit and possibly a loss.
Process engineering, therefore, may identify areas where an organization’s operations as a whole can be more profitable, but the organization has to match its increasing business intelligence sophistication with educated process change. At the very least, a major IT initiative on this order of magnitude requires that a company have an executive with the authority to coordinate change.
Step Two: In-depth Analysis
Many organizations with data warehouses are action-oriented, but are light on analytics. They often have data warehouses full of invisible potential, which only comes to light as a result of sophisticated in-depth analysis – the underpinning of successful implementations. Assuming the appropriate data is warehoused, an organization should be able to unearth those hidden opportunities. It takes clever mathematics and pattern detection techniques to mine those valuable nuggets and bring them to the surface.
Returning to our financial services example, a lot of banks already have massive data warehouses which are being put to less than optimal use and could be fine-tuned to generate a higher ROI. A bank may track in excess of 400 attributes on its customers. In addition to basics such as age, gender and income, it stores such attributes as: the number of cash advances in the past 40 days; the state of the balance over the past several months; uses to which a card is put; the number of consecutive days that zero dollars were spent in restaurants; whether the card holder belongs to an air travel club; average usage compared to the general population. By properly cross-linking and data-mining those 400 attributes, the IT department can facilitate access to an enormous amount of buyer behavior data.
With this much information already collected, most of the labour intensive work involved in creating and implementing the data warehouse has already been completed and paid for. At this point it’s simply a matter of exploiting it properly.
Researchers and marketing analysts are no longer restricted to working with that 1950s axiom of prosperity: a husband, a wife, 2.3 children and a Chevrolet. Today, data is analyzed and referenced across an enormous range of attributes: everything from air miles to product purchases by category, intervals between credit card use, family income, entertainment choices, average bank balances, investment portfolios, and so on.
Step Three: Decision Optimization
The BI task does not end with successful data mining. Along with the instantly useful nuggets of information discovered through mining, managers are unearthing knowledge which looks valuable, but cannot be ‘actioned’ without more work.
This third process is called decision optimization. At this point, the necessary data is already in place supporting the conventional warehouse. The next step is to run it again in new, strategic ways designed to help managers make key decisions, typically about resource allocation, profitability or cost reduction.
The service provider can help IT managers subject the data to more programming and complex analysis combined, again, with research analytics. The tools discovered in this largely unexplored field let managers assess millions, perhaps billions, of decision trade-offs in order to arrive at a single correct, strategic decision.
At IBM, we and our customers have discovered some surprising answers when information is mined at this level. IT managers need to be prepared for unexpected results and should be ready to understand and accept the answers. In most cases, managers have never looked at their data in such depth before and should be prepared to change processes to exploit this new learning. The information discovered as the result of this third process step may well change the way a company thinks about its business.
When an organization mines its information to this degree, results sometimes appear that seem counterintuitive or incorrect. However, this third process often uncovers the real footprints of an organization’s customers, and sometimes its competitors, letting managers know exactly what is going on. So managers should be warned that when a warehouse is subjected to such analysis there may be surprises.
The attitude of senior managers to the importance of IT in their businesses is changing, and changing fast. Three years ago, the majority of CEOs saw no reason why IT needed to be an integral part of corporate strategy. The attitude was: strategy is strategy; IT is a tool to make it work. Today, most of those CEOs realize that information technology is tightly woven into the fabric of business strategy and decision-making. Organizations are data rich, but information poor. As such, senior managers are searching for solutions to effectively deploy Business Intelligence solutions in their organizations.
Taking full advantage of a data warehouse may be as basic as this: be prepared to change the way you do business. True value depends on being able to spot, and act upon, those unexploited gems of opportunity.
Steve Del Zotto is Solutions Executive, Business Intelligence Solutions, IBM Global Services, IBM Canada Ltd. He can be reached at sdelzott@ca.ibm.com