Sales Channel Analytics
We might consider sales channel analytics a subset of business productivity analytics, yet there is enough value in segmenting this area of application.
- Marketing—Both the ability to fine-tune a marketing program and the ability to determine marketing effectiveness can be derived through sales channel analytics. A typical iterative process would be to identify a marketing strategy based on an analysis of a clustering of customers by profile and then to implement that strategy. The effectiveness of the strategy will ripple through the sales channel data, which can then be used to compare the actual results with expectations. The degree to which those expectations are met (or exceeded) can be fed back into the analytical processing to help determine new strategies.
- Sales performance and pipeline—Data associated with the sales staff can be analyzed to identify variables that affect the efficiency of the sales cycle, such as individual sales staff member, region, industry, contact people, contact times, and contact frequency.
Supply Chain Analytics
Supply channel analytics are used to characterize and benchmark a company’s supply channels from its various vendors and suppliers, through internal inventory management and ultimately aspects of delivering products to its customers. Aspects of supply chain analytics involve the following:
- Supplier and vendor management—Many organizations are unable to identify who their vendors are or how many vendors are supplying products or services. Supply chain analytics allow a company’s management to track performance and reliability by supplier, evaluating and rating the quality of the products supplied, as well as help to optimize supplier relationships with respect to spending, procurement, and risk.
- Shipping—There are different methods by which a company delivers its products to its customers, each with its own cost schedule. For example, it may be more expensive to ship products by air than by truck, but the products will arrive at the destination faster if shipped by air. A company can minimize its delivery costs by being able to select the most efficient delivery method for any specific business arrangement, but knowing whether the products can be available within the right time schedule is a difficult problem, especially if your production depends on external suppliers. Therefore, merging supplier and inventory information with productivity data (see Business Productivity Analytics on page 19) lets management accurately determine the best way to move product.
- Inventory control—As discussed earlier, maintaining an inventory of commodity products that exhibit volatile pricing and limited useful life creates a market risk if those products cannot be used before their obsolescence. Alternatively, we would not want to keep the shelves empty, because parts are needed to build the products that are in the order-and-fulfillment cycle. Between the sales channel information, the productivity data, and the supply chain data, it is possible to make more precise predictions about inventory requirements. It is also possible to determine the best way to quantify and mitigate risk, especially through the development of financial products (such as barrier options) to limit financial losses.
- Distribution analysis—Imagine that your company has a large number of retail outlets, a smaller number of regional warehouses, and a very small number of factories. The optimal distribution model would arrange for the delivery of the exact number of products from each factory to its closest warehouses so that each warehouse could deliver the exact number of products to each of the retail stores. Unfortunately for both companies and customers, this optimal distribution is pretty rare. If a company can predict demand for specific products within certain areas, though, the managers cannot only distribute the product to the right locations in the right quantities, but also minimize shipping costs by ramping up product creation at the factories most economically geographically located at a rate that matches the consumer demand.
Behavior Analysis
Most of the analytical applications we have reviewed so far deal with “drillable” data that a manager can use to optimize some kind of process, such as sales, utilization, or distribution. Another area of analytics deals with a more fluid view of activity as a way to predict trends or capitalize on identifying specific kind of behaviors. In general, any behavior pattern that presages significant business events is worth noting and then seeking. This type of analytical processing makes use of historical data to look for behavior patterns that take place before the significant event (whether or not they are causal) and then try to identify those behavior patterns as they are taking place. This allows for the following kinds of analytics:
- Purchasing trends—Although many product lifecycles can easily be predicted and charted, there are apparent nonlinear trends that elude predictability, the most notable cases being toy sales around winter holiday time. Yet not being able to identify a warming (or heating!) product may result in the inability to ramp up production to meet demand or the inability to move products from factory to store shelves, which can effectively dump a glass of cold water on that hot product. Behavior analytics can be used to identify purchasing patterns that indicate a growing trend that can be used to adjust a company’s reaction to customer trends.
- Web activity—In the world of e-commerce, the ability to draw and maintain customers to a Web site and then encourage them to commit to purchasing products is not only critical to success, but also much more difficult than doing the same in a brick and mortar environment. Different kinds of content presentation may lead different kinds of consumers to behave differently. It is interesting to identify patterns that lead to committed business (e.g., product purchase)—let’s call them “success patterns.” Then perhaps including some personalization (see Customer Analytics on page 18), the content presentation can be crafted to direct the Web site visitor into these success patterns, which in theory should improve the probability of making a sale.
- Fraud and abuse detection—Fraudulent (or abusive) behavior frequently is manifested in patterns. For example, there are many popular health insurance fraud schemes involving making claims with inflated charges or practitioners prescribing expensive medications or procedures that may not be necessary. Behavior analytics can be used to seek out patterns of suspicious behavior by provider, geographical region, agent, etc.
- Customer attrition—Another serious problem for many businesses is customer attrition, when a company’s customers decide they no longer want to remain affiliated with that company. In competitive industries, it is much easier to convince a customer to stay with the company before the decision has been made to leave rather than afterwards. For example, offering a long-distance telephone customer a better offer than can be gotten from a competitor can recapture that customer, but it is not to the company’s benefit to make this offer to (higher valued) complacent customers. Therefore, it is important to recognize the signs that a customer is ready to cease being a customer. This can be done by evaluating patterns of behavior before previous attritions (such as a history of customer service complaints) and then using those patterns for ongoing customer behavior analysis.
- Social network analysis—Sometimes it is important to identify relationships between specific entities within a system and to analyze their behavior as a group. For example, a component of criminal intelligence is finding collections of individuals whose individual behavior may be nondescript yet who act suspiciously as a group. This kind of analytical processing is valuable to law enforcement, regulatory compliance (think of insider trading), marketing (consider viral marketing, which is a strategy that encourages individuals to pass your marketing message to all of their contacts), as well as sales optimization (by finding a contact path of people to find the right audience).
The Intelligence Dashboard
A key performance indicator (KPI) is some objective measurement of an aspect of a business that is critical to the success of that business. Such KPIs are a component of the conceptual scorecard for a business and can be associated with a number of different business activities, such as customer satisfaction, productivity, supply channel performance, and profitability. In fact a large number of KPIs can be defined in terms of measuring performance associated with many of the BI analytics functions that we described earlier.
Another conceptual value of BI is the ability to capture the business definitions of the key performance indicators, manage those definitions as part of the corporate knowledge base, and then provide a visualization dashboard that reflects those KPI measurements, presented in a form for management review. This intelligence dashboard displays the results of the analytics required to configure the KPIs in a succinct visual representation that can be understood instantaneously or selected for drill-down. An intelligence dashboard will not only provide real-time presentation of the selected KPIs, but will also hook directly into the BI components that allow for that drill-down.
The following are some example KPIs:
- Regional sales figures by sales location
- Personnel statistics
- Real-time supply chain reports by supplier
- Customer satisfaction measurements
- Factory productivity
- Average customer profitability
Business Intelligence Adds Value
We can confidently say that knowledge derived from a company’s data can be used as an asset, as long as senior managers understand that an investment in turning data into actionable knowledge can have a significant payoff. It is important to recognize that this problem cannot be solved solely by the application of technology. In truth, the technology must augment a more serious senior-level management commitment to exploiting discovered knowledge and having a way to measure the value of those activities.
There are a number of BI analytics that provide business value. Selecting and integrating these analytic functions depends on the ability to effectively build the underlying information infrastructure to support the applications as well as the ability to configure reporting and visualization of the discovered knowledge.
For a more in-depth discussion of information valuation, I recommend “Measuring the Value of Information: An Asset Valuation Approach,” Daniel Moody and Peter Walsh, ECIS 1999.
David Loshin is the president of Knowledge Integrity, Incorporated, a
consulting and development company with expertise in the areas
of business intelligence, knowledge management, data quality,
data mining and scalable system development. Knowledge Integrity
helps companies address common (and not so common) business
problems arising from the collection, migration, transmission
and analysis of large sets of data. David is the author of
"Business Intelligence - The Savvy Manager's Guide"
(Morgan Kaufmann 2003). Contact him at (301) 754-6350, or
via email at .
|
