Introduction
In recent years, the evolution of technology and simultaneous growth in data availability have enabled a revolution in information management and customer-focused decision-making. Decisions have become increasingly datadriven and the time frame for these decisions continues to accelerate. We refer to the extent of a company's deployment of processes that capitalize on these capabilities as Analytic Maturity (AM). Fundamentally, AM entails the compilation of inputs from all parts of the business, the full processing of all relevant linkages and the use of this new information to drive optimal business decisions. Although the appropriate components vary across industries, it is clear that companies that excel at Analytic Maturity are the most likely to succeed. So, the important question for any company should be how to get there, rather than whether it is desirable.

Surprisingly some business observers have suggested that this trend may be unimportant or unnecessary for many companies (see for example, Raden at http://www.hiredbrains.com/Davenport_Rebuttal.htm, commenting on Davenport, www.tomdavenport.com). We believe that Analytic Maturity is compatible with both highly centralized, expert-driven solutions and decentralized distributed decision-making. Although AM impacts entire companies, it needs to be highly customized in order to be effective.

Who Should Own Business Intelligence?
Should it be quantitative experts who analyze patterns in the data to assist in big decisions and direct strategy OR is it at its best when used by a broad range of people and processes at the operational level, marginally improving performance, repeatedly and often?

I suggest that we need both the forest and the trees! Organizations and employees make many thousands of big and small decisions every day. Each of these needs to be supported by appropriately scaled BI. But the whole is clearly more than the sum of the parts: The provision of appropriate data, and execution of appropriate local decisions based on that data (e.g. Should John Doe's expense claim be approved, Is the manufacturing process for the last 1000 units within six sigma specifications, should prospect segment 5 be offered the 5.9% introductory APR) does not automatically ensure that bigger issues will be dealt with in an optimal analytically mature fashion (for example, optimal integration of direct marketing of introductory rates for new customers, promotional programs for existing customers and mass market communications). True Business Intelligence ensures that each individual and department within the organization is supported with the information they need to make the right decisions; and some decisions require both sophisticated models and appropriately trained experts.

Does Software make PhD's Obsolete?
This question could also be cast as the tradeoff between a centralized team of quantitative experts versus more democratic non-expert departmental decisionmaking based on user-friendly tools.

There has been great progress in analytic software over the last 15 years, including tools for business intelligence, campaign management, customer relationship management, statistical modeling and data visualization. The collective impact of this evolution has resulted in much greater efficiency in accomplishing tasks that could previously be done only manually and in "democratizing" many specialized functions. Today, nonexperts can be trained to use these tools to accomplish tasks that previously required programmers or statisticians with advanced degrees. However, this is not a fool proof solution since these "software operators" are always vulnerable to the rule that "you don't know what you don't know" - in other words, there are data analytic challenges where expert judgment is still required and no software can be relied upon to make these judgments or even identify these situations. For this reason, the ideal solution is to combine a central (or external) expert group for the really hard problems with a broad dissemination of efficient tools to non-experts throughout the organization.

Centralize or Decentralize?
In other words, should data and expertise by centrally controlled OR do agile organizations require highly decentralized decision making with local BI tools?

Analytic Maturity often entails the existence of a centralized data warehouse or expert modeling group or the presence of a BI system that makes information linkages transparently and flexibly available throughout the organization. However, such systems and processes are not synonymous with centralized control of decision making. In fact their essential value lies in maintenance of data quality, accurate discovery of business process drivers through modeling, and provision of flexible access to data relationships and information through BI systems. When effectively implemented, this results in a high quality support system that enables accurate local (i.e. departmental) decisionmaking, to the benefit of both the department and the whole enterprise.

Does the CEO Need to be On Board?
Putting it another way: Does success on the road to Analytic Maturity require senior executive commitment OR is much of the value in analytics really an issue for local improvements?

Any initiative to advance a decisionmaking process along the Analytic Maturity scale requires appropriate leadership commitment. Of course the necessary level of leadership depends upon the scope of the initiative. Some important improvements can be quite local - so, for example, the adoption of more sophisticated BI software may be the responsibility of a VP reporting to the CTO. However, some organizations may require a far-reaching realignment of information management and decisionmaking processes. It would be difficult to imagine that such an initiative could succeed without collective C-level leadership.

Different for Product and Service Businesses?
Is Analytic Maturity only for customercentric service businesses or does it have broader application, even to manufacturers?

Consider the cases of a consumer credit card issuer and a home appliance manufacturer. The credit card company can readily be identified as a customercentric business:

• Dependence on direct marketing through tailored communications to individual consumers
• product design as a flexible and readily customized process
• production is largely a matter of tuning the rules for individual customers (interest rate, rewards programs, credit lines etc.)
• key to a profitable relationship is the customer's ongoing utilization of the product, while a successful "sale" or customer acquisition is not in itself profitable
• clear and direct link among marketing communications, product management and individual customer relationship management

On the other hand, the appliance manufacturer has a much less immediate relationship with its customers:

• Most revenue occurs at the moment of the sale
• little if any ongoing relationship
• transactions are intermediated by an independent retailer
• greater proportion of the manufacturer’s marketing budget is dedicated to mass market communications, building brand and product awareness
• manufacturer needs to dedicate substantial resources (including analytics) to the challenges supply chain management, facility planning, product design and manufacturing quality control.

Nevertheless, it has been many years since manufacturers could operate with a “make it and they will come” mentality. It should be clear from the above that manufacturers can benefit just as much as credit card issuers from access to integrated and timely data that spans financial performance, product design, production, marketing and customer understanding. Similarly the appliance makers’ decisioning processes benefit from being tailored to leverage this broad span of information as needed. As manufacturers progress along the Analytic Maturity spectrum, the relevance of this process becomes even more evident as they develop new service opportunities – for example, for customized service contracts and credit plans.

Conclusions
Analytic Maturity impacts information management and decision making throughout any organization. But it is not a one-size-fits-all solution. The key dimensions of an analytically mature solution for companies in different industries are quite different with respect to the optimal degree of integration of data across departments and processes, the need for decisions to be centralized vs. decentralized and the extent and time scale in which operational issues need to be closely linked to one another.

Future articles will provide more information about Inductis’ vision for Analytic Maturity and the process steps required to achieve it.

About The Author

Dr. Martin Ahrens, Vice President, Methodology & Quality Assurance, Inductis

As Vice President of Methodology & Quality Assurance, Martin develops analytic solutions to business challenges. He advises all Inductis project teams on the selection and application of scientifically sound methodologies and also oversees the Quality Assurance process. In addition, he plays a lead role in communicating our analytic solutions to the broader business and technical community. During his career, Martin has provided technical leadership in risk analysis, consumer and business marketing, product rating development, survey design and new business development. He served as a Senior Engagement Manager and led the development of analytic services for Mitchell Madison Group, a global management consulting firm.

Previously, he held the position of Director of Statistical Services with Consumer Reports, the world's largest independent consumer magazine and product testing laboratory. Martin has served many leading firms in the publishing, online services, training, financial and data management industries. He combines broad business understanding with advanced statistical knowledge. Martin earned his Ph.D. in Ecology at McGill University in Montreal.