Turning Data Into Action

Introduction?

The business topic on the minds of many senior executives is positioning their organizations for achieving increased profitability through enhancing their “share of the customer.”? While their approaches may vary, a common goal of many organizations is to increase their customer focus by getting closer to their customers and achieving customer intimacy through a better understanding of customer needs and preferences.? Many of these organizations are turning to their Data Warehouse to help achieve this heightened customer relationship.? These organizations are leveraging customer information to serve customers better, help themselves make fact-based decisions and take swifter actions on behalf of customers.? A Data Warehouse enables this capability by offering greater opportunities for customer-focused business transformation.?

The absolute measure of value to be derived from any information environment lies in an organizations ability to leverage this source of consolidated customer information into actionable business decisions.? Many leading edge customer focused organizations are indeed ‘turning data into action’ to achieve new found growth via customer intimacy.?

A customer intimate philosophy dictates treating every single customer interaction (inclusive from product/service activity usage through to customer complaints) with your organization as an opportunity to learn more about your customers preferences and their buying habits.? Achieving customer intimacy requires a 360° view of your customers to enable development of products and services which will increase your ‘share of the customer.’? Finally, a customer intimate approach will lead to the creation of an environment capable of answering the customer-centric business issues listed in Figure 1.?

Breakthrough Solutions and Business Views?

In many organizations, there are no established formal processes or systems in place dedicated to consolidated cross-functional decision support.? The most urgent requirement in these organizations, which drives development of a Data Warehouse, is to provide users with an easy-to-use method of accessing their legacy (some prefer ‘heritage’ or ‘foundation’) data.? Building a Data Warehouse enables organizations to leverage information to gain a strategic advantage.? In order to fully contribute to business initiatives and be perceived as delivering true value-add, a Data Warehouse must deliver functionality to users which enhances or exceeds their current capabilities.? Merely migrating existing functionality into a Data Warehouse does not leverage the significant opportunity to redress user needs and design an environment suited to their on-going requirements.??

While many Data Warehouse practitioners focus solely on building the environment, due diligence must be paid to ‘what you are going to do with it’ and ‘where do you plan to go with it’ to return maximum value to your organization.? Having provided the enabling infrastructure, your focus should rapidly shift to identify and develop information applications which maximize the opportunity.? An appropriate term for these applications is “Breakthrough Solutions,” due to their unique data-centered characteristics compared to traditional business function-automating applications.??

Breakthrough Solutions are designed to specifically resolve an identified business issue.? These issues are frequently cross-functional, spanning departments or functional business areas.? A Breakthrough Solution will enable users to carry through on their analyses without the limitations imposed by traditional stovepipe data processing systems.?? Having navigated through a prescribed path to narrow their subject or field of interest, users can begin to explore the underlying detail data to further support their analyses.? Their analyses proceed in this manner since the Data Warehouse allows for an information-oriented navigation rather than by a transactional-focused Order Entry or Billing system.??

The fundamental aspect of Breakthrough Solutions is that they are designed and developed exclusively to empower end-users.? Breakthrough Solutions manage data as a corporate asset, regardless of the source system or department owner of the data.? Breakthrough Solutions permit data to be shared among and between departments on an as-needed basis.?

As noted earlier in defining the term Breakthrough Solutions, they do not necessarily follow the same development methodology as traditional application development techniques.? An appropriate characterization of Breakthrough Solutions is to consider them as multiple Business Views.? Each Business View has different characteristics depending on its target user audience and their required scope of analysis.?

A successful approach to developing these Business Views is to realize that compared to traditional application development, your goal is to never develop a static or final version of your information applications.? Becoming comfortable with the dynamic nature of Data Warehouse Business Views will help you adapt to this on-going evolutionary process in partnership with your users to meet their ever expanding requirements.?

The issues involved in developing the components of a Business View (e.g. a Customer or Product Profile) requires identifying what the profile will contain and how that information will be presented to the user community.? In other words, you must consider both the ‘form’ and the ‘function’ of presenting information.? Figure 2 presents a logical representation of the relationship between the ‘form’ and the ‘function’ using a Product Revenue Stimulation Business View as an example.?

The ‘form’ side of the equation requires understanding what each individuals Business View will focus on and how it will vary between the different domains (Business Information Views) of the profile.? The views are constructed based on end-user information requirements and the level of detail required for each individual users analysis. Access is packaged in such a way that end-user delivery is tailored based on the user profile of how much or how little detail is required to complete an individuals Business View.? In reference to the Product Revenue Stimulation Business View example in Figure 2, each User Focused View is developed looking up from within each analysis.? This is a viable approach since it is unlikely that all users will require the same detail view for all analyses within all domains.?

The ‘function’ element of a Business View requires a definition of the scope for all domains of a profile and then further exploring the scope of analyses within each domain. Identifying domains for a given Business View requires an understanding of related business issues and the benefits to be achieved from developing a specific domain.? Domains can be considered as the point where an organizations business issues are related to the corresponding information requirements.? The individual analyses within each domain can be considered as the point where data is directly related to resolving the business issues.? Analyses can be structured for ‘ad hoc’ user access, ‘pre-defined’ user access, ‘EIS’ user access, results from Exception Reporting or access to pre-defined views of results from a Predictive or Analytic model.? Analyses routinely combine various levels of detail data which incorporate multiple access types.? In all cases, there should always be an access path back to the detail to enable an iterative drill down within the topic of interest.??

So how does a Data Mart approach relate to the Business Views concept?? A Data Mart is the physical delivery infrastructure for a particular class or domain of analyses within a Business View.? Data Marts are intended to provide easy access to information primarily for reporting purposes.? A Data Mart typically contains an extract of a function or subject-based subset (e.g. Revenue reporting from the Finance departments perspective) of data in the Data Warehouse.? This enables users to rapidly manipulate a collection of data using tools designed for multi-dimension analysis (e.g. hyper-cubes, cross-tab and table pivoting tools).? In building this physical architecture, a critical success factor remains; the facility to drill down to the detail must always be present.?

Data Navigator?

The first step in fostering user acceptance of their actionable decision environment is by using a presentation technique called a Data Navigator.? Developing a Data Navigator will assist in the creation and manipulation of a Business View by making the contents of the Data Warehouse easy to understand.? The Data Navigator enables a user-oriented categorization of the different dimensions within the Data Warehouse and breaks down each dimension further by its supporting elements.? Upon completion, the Data Navigator offers a highly effective technique for visualizing the contents of the database.? Figure 3 shows a sample Data Navigator for the Transportation industry.?

The Data Navigator can also be utilized to assist in the creation of ‘non-scientific’ predictive models.? The key to building ‘non-scientific’ predictive models is to ensure that business users are contributing to the definition of all aspects of the model as well as its refinement.? Their input will identify all appropriate dimensions and the respective elements which will contribute to the predictability of a given model.? After the targeted elements are weighted appropriately, the model is ready to be tested on a sample of data.? The process of building a predictive model in this manner is highly iterative and requires on-going measurement and continual refinement to ensure that its design is optimized while it accommodates changing business conditions.? Though less accurate than a detailed model, a ‘non-scientific’ model is a far more cost-effective solution that is easily adjusted to reflect changes in data structures or the need to incorporate additional data elements or data sources.? When appropriately deployed, this technique has proven to be effective by enabling more diverse and rapid development of predictive models when compared with more rigorous scientific modeling.?

Evolution of Knowledge Workers and Data Analysis?

An underlying premise for early adopters of Data Warehouse enabling environments is that information is the foundation for building an effective organization.? In order to maximize the ability for an organization to fully leverage their resources they must subscribe to the fundamental requirement of delivering readily available, accurate and timely information for fact-based decision making.? They must be committed to enabling their knowledge workers to meet the challenges of the current business environment and successfully compete in the future.?

Shifting from tradition-based, folklore and intuition-based decision making to fact-based decision making requires that knowledge workers evolve their skills in analyzing the data.? There are typically three phases in this evolution which are described below;?

??? Preliminary Focus - Preliminary queries focus on basic information about customer and product activity; primarily "counts and amounts".? An example from the Communications industry is as follows; How many customers order Call Waiting and Call Forwarding at the same time; By Postal Code?? By Wire Center?? By NPA_NXX?? This line of questioning is the first step towards becoming true knowledge workers.?

??? Advanced Focus - After the preliminary questions have been answered, users typically advance to iterative information analysis to derive insight and knowledge from the data.? This level of data interrogation involves repeated analysis of data -- drilling down into the detail further and further to uncover hidden or underlying relationships.? For another example we’ll stay with the Communications industry; Develop a profile of high revenue customers using multiple dimensions, including demographic and product/service activity.? Compare and contrast the resultant characteristics of this profile against segments or bands of lower revenue customers to identify potential target candidates for revenue stimulation.

??? Actionable Decisions -? The goal of all Data Warehouse environments is to facilitate actionable decisions.? As your environment matures, knowledge workers and system support personnel develop an appreciation for the power of and an increasing level of expertise in managing and manipulating this new found knowledge base.? Actionable decisions offer a means of transforming the way organizations interact with customers, the ways in which information is used to drive decisions and actions and the ways in which customers perceive their suppliers.? Achieving this level of capability is the first step towards spanning the gap between decision support and operational system environments. The process of linking rapid actions to emerging changes in the marketplace has been proven to be a critical operational advantage in the marketplace for companies whose efforts have evolved leveraging this strategic capability.? Successfully merging strategic initiatives with a decision making process based on timely and accurate information will have a ripple effect throughout the organization and offer opportunities to positively transform the business.

Another important factor which will allow organizations to maximize their Data Warehouse value is to align the goals of the organization with a decision making infrastructure designed to achieve those goals.? Knowledge workers must be encouraged and empowered to rapidly act on their findings and transform their job functions from being ‘task oriented specialists’ to ‘results oriented generalists’.? They must have an environment which stimulates and rewards their innovation and creativity in leveraging corporate data.? Organizations must continually ‘raise-the-bar’ and challenge their knowledge workers ability to manage their area of responsibility within the operation.

In other words, organizations must provide knowledge workers with the responsibility for self-discovery of business trends and the authority and accountability to act on those discoveries.

A caveat to empowering knowledge workers is blind faith in the ability of your knowledge workers to make the right decisions.? Rich Teerlink, CEO, Harley-Davidson says, “If you empower dummies, you get bad decisions faster.”? While this statement reflects a valid concern of Senior Management, there needs to be a program in place to train the uninitiated in methods and techniques to fully leverage corporate information to produce better decisions and ultimately better actions.? Merely putting a PC with access to actionable information on a knowledge workers desk and expecting more frequent and better actions is analogous to putting someone behind the wheel of a sports car and expecting them to win ‘The Indy 500’.?

To summarize, the technology required to carry out large-scale decision support is merely an enabling tool. To truly capture the rewards of a Data Warehouse environment requires a well thought out and organization-wide plan to facilitate and support an actionable decision making infrastructure which incorporates the supporting actionable business processes.

Types of Data Warehouse Business Views

A Data Warehouse enables an organization to capitalize on the wealth of data which are typically created as a by-product of their business operation.? Moving to higher stages of actionable information requires a vision and a plan for how to leverage this base of information.

In general, there are three stages of actionable information.? The first stage is the most basic and will arise as a by-product of building a consolidated Data Warehouse environment.? While it has been shown to deliver value and quantified benefits, it does not contain significant potential for business transformation.? As characterized in Figure 5, the next two stages offer the greatest opportunity to return maximum value from any Data Warehouse environment.? The first of these stages involves profiling your customer base across multiple criteria to identify how your customers map to an ‘ideal’ or best-in-class customer profile.? The second of these stages employs sophisticated modeling to predict customer propensity based on past behaviour or tendencies.? This approach is favourable as it provides a mathematical scoring of the target sample, however, it is more costly to implement and needs to be robust enough to be employed in a repetitive fashion.

While the approach of these two latter stages differ somewhat, they both offer great potential to positively impact the business and therefore, may be considered equivalent value initiatives.

This section provides a grouping of three closely related concepts.? The preliminary topics detail the different types of Business Views or Business Categories.? They are followed by a look at the types of analytical techniques which can be integrated into each supporting analysis.? The section closes with a discussion on some potential uses of the results of Business Views.

Types of Business Views or Business Categories:

Corporate Performance Measurement

Responding to a general lack of data availability and accessibility among knowledge workers, many organizations consider it a major accomplishment merely providing their users with access to the information that had previously seemed so far away.? The most common preliminary deliverable in most Data Warehouse environments is providing end-user direct access to data.? This gives end-users the ability to self-generate standard historical reports from a centralized and consolidated source focusing on specific details which? have transpired over a period of interest (e.g. last week, last month or last year).

While there is value in providing this capability, it remains minimal compared to potential gains of more advanced initiatives.? Typically this type of capability provides value via cost avoidance and cost displacement metrics based on the alternatives of capturing information within large organizations.?

An example of cost displacement is the value in providing a consolidated view of all previously disparate Decision Support System (DSS) environments.? This offers substantial impact when considering the effort expended on behalf of most organizations by the numerous Information Providers who have seemingly made themselves a career through ‘data’ empire building.? There are countless people in all large organizations, who have carved out a niche for themselves because they have taken the time to understand the organizations data environment.? This should not be viewed as a negative development.? In many cases, these same people are fulfilling a valuable service by sharing information with others in the field which ultimately benefits the entire organization.? Unfortunately, this is typically thought of as the only means for the end-user to get at valuable information.? It provides a vehicle for end-users to go ‘around’ the I.S. organization which can lead to a potentially enormous duplication of effort and a deviation in focus of the charter of individual departments.? In all cases, it leads to multiple versions of information being used as the basis for decisions and it robs the organization from the potential collective capability available if all of these individuals had the opportunity and the forum to harmonize their efforts.

The ability to effectively leverage the full value of a Data Warehouse is tied to an organizations ability to move up the proverbial ‘food chain’ and implement the types of Business Views which follow.

Customer Profiling

What is a Customer Profile?? This question evokes as many different responses as asking how an organization answers the question - “What is a customer?”?

An effective Customer Profile (or similarly a Product Profile) can be the cornerstone for building a customer focused organization.? Developing effective Customer Profiles can be achieved by harnessing the collective business knowledge about your customers and combining this knowledge in such a way as to foster a corporate-wide common view of understanding customer behaviour.? Building a profile can take on many different approaches, many of which may be beneficial.? The key is in determining how your organization wants to characterize its customer data and to clarify the current and future intent of its use.

The most useful Profiles are those which incorporate multiple cross-functional views of your customer which prior to simple user access to a Data Warehouse were simply unavailable.? Facilitating the analysis of all customer interaction with an organization will lead to developing customer oriented products and customer oriented processes to serve your customers.? They are certainly the first step towards transforming an organization into a customer focused business.? The best Profiles are those which provide multiple methods for correlating seemingly unrelated customer activity.

Analytic Modeling

Analytic or Predictive Models are becoming increasingly prevalent in Data Warehouse environments.? The category of modeling techniques includes; Neural Network technology, more traditional Statistical Analysis (e.g. logistic regression), Induction methods and Data Visualization.? These techniques are used to uncover previously unidentified inter-relationships, to model customer transaction data or to score or predict future customer behaviour. The process commonly results in mathematical algorithms applied to internal operational transaction data combined with external demographic data. These algorithms are iteratively refined until the highest level of accuracy is achieved.? Analytic Modeling techniques are frequently employed to provide detailed insight into a targeted customer segment to identify which customers are most likely to respond to a targeted or specific future initiative.?

In Analytic Modeling circles there is considerable debate as to the merits of traditional Statistical techniques versus Neural Network technology versus other non-traditional Induction algorithms.? Statistical modeling proponents state that “the performance of a technique is highly dependent upon the nature of the data set under study.”? Neural net proponents claim that based on the ability to deal with non-linear relationships, “ a well built Neural Network will always equal and will usually outperform a statistical model by 10-15%.”? The third category of advanced forecasting technology includes new and emerging ‘off-the-shelf’ products.? These products are commonly built upon proprietary induction algorithms which they claim produce consistently better results than Neural Networks.? You can conclude that in real world applications each technique may present certain advantages.? Factors to be considered in choosing the right technique, include; the cost of implementation of the solution, the re-usability of the solution, the ability to tune the model as your needs evolve, the practical impact and the scope of the implementation along with your desired result set.

A preliminary sophisticated Analytic Modeling is often referred to as ‘Scoring’.? ‘Scoring customers employs mathematical algorithms to rate an individuals propensity or likelihood to acquire a specific product. This can lead to a Targeted Marketing sample or the results may be used to embellish a Customer or Product Profile Business View.

Analytic Models are created by identifying positive or negative characteristics that may influence customer behaviour.? The next step involves weighting the contributing factors effectively in alignment with the desired response.? While these steps are similar to the non-scientific method of predictive modeling discussed earlier, Analytic Models are more complex to develop and routinely incorporate a greater number of variables with increased accuracy.?? Analytic Modeling has proven to be beneficial for employing programs to acquire, retain, stimulate, cross-sell or winback customers.?

Types of Analytical Techniques

Unrestricted Ad Hoc User Queries Access

This category of Business Views includes one-time stand alone requests and iterative or culling type analyses which are commonly referred to as “data mining”.? Providing unrestricted access to detailed data facilitates iterative analyses where the end-user typically produces a number of successive queries.? Each of these successive queries are developed with insight acquired from the previous answer set.? An increasing number of very sophisticated tools are available today to facilitate the ad hoc analysis process (for both structured and unstructured ad hoc data discovery).? Many organizations have realized significant returns where users have the means and the incentive to seek out underlying trends and patterns within a topic of interest.

System Monitored Exception Identification and Reporting

Exception Reporting analyses are the basis for Business Views which can be high value and relatively easy to implement.? Exception Reporting systems are designed to monitor an identified subject or condition and trigger a flag or alert when data indicates a pre-defined threshold is being approached.? These types of systems can be automated for continual incidence monitoring or they can be incorporated into an analysis within a Business View.? Other types of Business Views which can be included in this category are Executive Information Systems (EIS) and other leading edge business transformation type initiatives such as unattended notification to operational systems.

The term Executive Information System along with many other terms in the technology field evokes vastly different interpretations from one organization and one person to the next.? With the alternatives in production today, the term EIS is far too generic to describe a single type of system.? In actuality, EIS’ are designed for different purposes and take on many different flavours.? The intent of early EIS environments was to provide senior executives with a snapshot of operational performance delivered via on-line easy-to-interpret (versus easy-to-use) results.? Some organizations have evolved their EIS to include gateways to pre-defined specific performance data along with the ability for the executive to engage in very high level data mining.? This evolution leads to re-categorizing EIS using the Business Views approach discussed in the Breakthrough Solutions section on page 2.? This states that Business Views take on different functionality depending on their target user audience and their scope of analysis.

Data Mining

The term Data Mining is quickly becoming a generic reference used to describe a number of vastly different things.? This was certainly the case with two of the previous ‘hot’ topics in the Data Warehouse world; Metadata and Data Warehousing itself.??

Contrary to what many tool vendors would like you to believe, ‘real’ Data Mining entails far more than typical multi-dimensional analysis capabilities (e.g. OLAP) and traditional iterative SQL queries (e.g. ROLAP).? A complete (yet somewhat lengthy) definition of Data Mining which clearly articulates both its purpose and the process is ‘the automated analysis of detailed operational customer transaction data for the purposes of discovering hidden, unidentified or underlying patterns, trends and inter-relationships to understand, score or predict future customer, product or process behaviour.’? Data Mining provides the opportunity for knowledge workers to go well beyond assumption-based hypothesis testing or to draw conclusions based solely on intuition in discovering inter-relationships within the data.

A robust Data Mining environment is based on analytic or sophisticated mathematical modeling and is becoming increasingly necessary to maximize the value of a Data Warehouse.? It serves as an ideal method to understand what is driving or likely to drive a particular customer behaviour, product activity or event enabling organizations to be proactive rather than reactive to changing conditions.

The most common Data Mining techniques employed today include; Neural Network technology, traditional Statistical Analysis, Induction methods, Time sequencing and Clustering.? These techniques are spawning new powerful Data Mining applications which span multiple industries.? These applications include:

Market Basket Analysis or Product Affinity Analysis (What products or services are most frequently purchased as a group?? What are the ‘on ad’ products which drive the highest value affinity sales?)

Customer Retention/Vulnerability (What are the factors or characteristics which predict that a customer is on the verge of cancellation?)

Customer Acquisition Lifecycle (What is the sequence of products (events) that will be acquired (take place) over what period of time?)?

Price Optimization (What price will they pay?? What price will move the optimal number of products while providing the optimal return?)

Risk Management (What are the over riding factors or characteristics to consider in predicting the risk level of a particular customer scenario or portfolio?)

Target Marketing & Segmentation (Which customer and/or channel characteristics are indicative of producing the most profitable customers in the shortest period of time which have the longest tenure?? Who has the highest propensity to acquire a given product?)

Proceed with caution if you intend to ‘mine’ data which has been highly aggregated or too summarized.? Your results will be probably be too generalized and certainly be very ‘average’.? You must ensure that you are mining a data set which has sufficient breadth (detail) and depth (history) to produce meaningful results.? With tongue planted firmly in cheek, one Neural Network developer once noted that…..“Models are usually wrong, however, they can make a customer a lot of money!”

Potential Uses of Business Views

Operational System or Business Process Assessment

A Data Warehouse often enables a view of the data that may not have been possible prior to its inception.? This can be due to previous inaccessibility of data or the lack of interest in viewing the data in any particular manner.? The owners of operational systems and processes can initiate a data quality assessment to indicate a possible breakdown somewhere in operational systems and processes.? Combining multiple data demographic analyses such as, table group-bys and referential integrity checks can be eye opening when conducted for the first time.? In some cases poor data quality and poor data integrity can negate the value of developing targeted user-driven Business Views.??

A detailed data demographic analysis can also be a significant contributor to a business process re-engineering initiative.? Data demographic analysis results can contribute to a cross-functional prioritization of business objectives, a review of operational system data quality and data entry processes and procedures.

When consolidating information from several source systems care must be taken to ensure that common content from disparate systems exhibits the same format and that all occurrences within a given field conform to the same convention.? This is essential for effective and meaningful data analysis.? A process called Scrubbing & Matching is employed to identify both instances.? This process is critical since inconsistent and inaccurate data will produce conclusions which are at best suspect and likely negate any potential benefit gained from this enabling environment.

The greater challenge at the outset is to ensure that all data that fails to conform to a consistent specified format of a given field is identified and adjusted appropriately (e.g. (inaccurate data capture & inconsistent convention) date fields, last name/first name flips; (multiple occurrences of the same instance) Customer Name: Revenue Canada, Rev. Canada, Revenue Service, Cdn. Gov’t - Rev. Can).? This leads to two simple alternatives.? Do you adjust the non-conforming content once it is stored in the Data Warehouse or do you go back to each source system and adjust the field ‘at the source’?? This decision is dependent on a number of factors.? Among the contributing decision making factors, first and foremost, is the scope and magnitude of the problem and the degree of difficulty in correcting the situation.? Other factors to consider include the volume of historical errors, the challenge of going back to all archived historical information and the policies and procedures of the given organization.? The ideal solution would be to focus upstream on the data flow and if feasible, adjust all historical operational data and ensure operational conformance as you move forward.? Unfortunately, this may not always be practical.? A go-between solution is to adjust the historical non-conforming information in the Data Warehouse and ensure that operational conformance is adhered to going forward.

Business Transformation

Companies which were early adopters of Data Warehouse principles have leveraged this advantage into a key differentiator.? In many cases customer to supplier to vendor relationships have been re-designed or completely designed from scratch with a business approach which leverages the new found knowledge they have acquired.?

An early adopter example of a Business Transformation application includes a supplier monitoring inventory levels in the retailers stores.? The responsibility to restock the shelves has been shifted from the retailer to the supplier.? The supplier can benefit from this relationship since they can now rapidly see comparative sales results from multiple regions.? Another example includes providing the customers of a Telco with access to their traffic patterns across multiple product lines.? A customer can conduct their own analyses and provision new or re-deploy existing products based on a greatly enhanced view of actual traffic and results.

Looking ahead, a future Business Transformation application could come in the form of ‘user agents’.? These agents may be enabled through the development of a standard method of interaction across multiple database and Data Warehouse systems.? While on-line via the Internet a potential customer can initiate an analysis searching for a product or products which match their search criteria (e.g. Search all target databases for a no-load mutual fund which has provided the highest yield over the past 5 years with the least fund manipulation where the fund manager has remained the same).? Interface products and services which are likely to emerge include sophisticated user agents which can transparently interface across several subject databases where access is enabled by vendor organizations categorized by subjects of interest.? The ‘user agent’ will accomplish the search and report back the instances which match with the users search criteria.

The User Interface

Data Warehouse Business Views can be based upon a custom interface design or standard tool set depending upon organizational preference.? They should provide a level of functionality which leads the user through a pre-defined set of questions to narrow or expand on the subject level of interest.? It is critical to provide a level of functionality which enables a user to drill down into the detail after navigating through to their specific domain of analysis.? The interface should provide much greater functionality than a base user level point and click access environment while shielding the user from having to know every detail of the underlying data model and how to generate appropriate queries.

In most environments you must provide capabilities for a diverse range of end-users.? Typically, multiple tools of varying functionality will be required.? By incorporating this user functionality within your Business Views architecture you can deliver an effective access strategy meeting the business needs of the user community.? Adhering to a Business View framework ensures that even as available tools evolve, your deployment strategy is consistent.?

The Organizational Interface

An essential component in positioning an organization for long term benefit from their Data Warehouse environment requires a process for on-going user support and analysis to be addressed.? Issues which are commonly associated with a Data Warehouse include;

1.?? Users that lack the necessary skills or data knowledge for information drill down analysis.

2.?? Users who don’t have the time to undertake detailed iterative analyses.

3.?? Users who don’t have the desire to leverage the opportunities presented by a Data Warehouse.?

In order to maximize the value of a Data Warehouse, the needs of these individuals must be considered by addressing the organizations support and analysis environment.? Some organizations have established ‘information analysis centers’ which function in partnership with I.S..? Functionally they resides between the users and the operational I.S. environment.? They are responsible for:

·?????? Mining the data.

·?????? Evolving and enhancing the user access environment.

·?????? Ensuring that users are served through evolving the data model to user needs.

·?????? Delivering information to those who are unable or unwilling to initiate information access.

Trying to accommodate the needs of end-users via typical infrastructures proves to be unnecessarily challenging and does not capitalize on the full opportunity enabled by a Data Warehouse.

Closely aligned with end-users, these ‘information analysis centers’ most closely resemble the traditional role of an Information Provider.? The main difference is in their charter to provide pro-active solutions based on the on-going prioritized needs of end-users.? They fulfill a discovery center and analysis engine role which provides the ability to fully leverage a Data Warehouse.? They are also responsible for generating and maintaining user enthusiasm for using the Data Warehouse as their primary source of information.? User participation has proven time and again to be the absolute measure of success for continued sustenance of a Data Warehouse.? I.S. must embrace a new paradigm and regard the Data Warehouse as a wholly user-driven environment.? Organizations must continually seek out, promote and encourage user involvement.? Failing to do this the Data Warehouse may generally be regarded within the user community as “just another I.S. sponsored initiative” and may never deliver on its potential value to the business.

Common Data Warehouse Challenges

There are numerous factors which inhibit successful information access for many organizations.? In general, knowledge workers lack access to current, actionable, quality information which enables informed decision making and actions.? The following are some inhibitors to timely information access.

Information Currency & Availability

Many business users are unable to get timely answers to fundamental questions, if they get to the information at all.? In cases where users have access to data it is typically months old and not very meaningful.?

The age old paradigm required knowledge workers to be fortunate enough to receive their requested reports with the answer sets exactly as required.? Of course this often was not the case.?? Following weeks of waiting, requests are routinely sent back to I.S. for clarifications or modifications.? If the subsequent changes are approved, it could take weeks or months for I.S. to turn around the request.? This leads to an environment where the users are reluctant to even ask for changes leading to an environment where people make decisions based on folklore and intuition.

Information Accuracy

“We have multiple sources of information and the information is different..........Stop the confusion.”? Multiple sources of information can snowball within an organization leading to vastly different interpretations of results which yields ineffective reporting.? An effectively implemented Data Warehouse enables actionable decisions based on certainty, not guesswork.

Information Accessibility

Empowering knowledge workers to make decisions implies that users have access to the information necessary to make those decisions.? Many companies can be classified as? possessing a wealth of data, while their knowledge workers are clamoring for information and starving for knowledge.? Effective information management dictates that users must have access to an appropriate level of detail information, on a timely basis, in an easy-to-use access environment.

A Data Warehouse overcomes many of these hurdles and over time will become indispensable to an organization.

Summary

The road to customer intimacy is wrought with twisting and winding curves.? It will be a challenge for most organizations to extend beyond their traditional perception of business systems to one where their business systems are viewed as a foundation for enabling true business transformation.

While the journey begins with consolidating all of your customer information, sustaining value from a Data Warehouse requires a user-centric approach as you move forward.? Long term value can be achieved depending on the involvement of your user community in the preliminary design of the system through to the ways in which you architect a solution to meet end-user access to the information.?

Traveling on the road to customer intimacy is a never ending journey.? By nature a Data Warehouse is dynamic and evolutionary.? It is one which offers many organizations the ability to leverage their customer information environment to heights previously reserved for the risk takers or early adopters of Data Warehouse technology.? Effectively harnessed many significant customer focused business transformation opportunities are waiting to be discovered.??