Achieving Continuous Improvement in Complex Supply Chains Today - Part 2

In Part 1 of Achieving Continuous Improvement in Complex Supply Chains Today, I reviewed Robert Bowman's article in GLCS "Achieving Continuous Improvement in Complex Supply Chains Today" and took a closer look at the P&G example cited there. In this post, I intend to focus on the application of Business Intelligence and its role in continuous improvement.
Robert brings up the notion of Managing by Dashboard in the latter half of the article. He introduces the following:

A crucial element of any such technology is the executive dashboard, software which allows managers to monitor at a glance a series of key performance indicators (KPIs). HighJump has built some 500 screens into its products, although each customer utilizes no more than a handful of measurements, based on its unique needs. The tool gives users the data needed to pursue continuous improvement.

If one refers to the age old diagram of a supply chain's structure, one would find product flow as well as information flow (typically in the reverse direction). Even with such an explicit definition of the importance of information flows and awareness of the importance of data within a firm, my experience in Supply Chain consulting shows that data about even the simplest transactions within a firm are mangled in ways that might drive horror into Frankenstein's cold living heart. The one piece of data that is never mangled though is payroll. Ah! if only?
Robert alludes to that by writing:

Dashboards are only as good as the data they contain. The first step, then, is to set up a database that can act as the single point of storage for all relevant information generated by a company and its trading partners. In addition, the database must be scalable to accommodate rapid growth in sales, says Brad Fellows, senior partner of transportation, logistics and distribution with Teradata in Dayton, Ohio.

True enough, dashboards are only as good as the data contained therein. However, the very notion of a scalable database is database theory not business reality. Instead, I think a scalable database should be called businessable database meaning that business analysts must not only scale their databases as they see fit but must be measured on the quality of the data they maintain and be assessed on the contribution that such data has on decision making. The most common notion that I find when decision making from "terrible" data is concerned is the notion of directionally correct or good enough for government work. That notion betrays a certain innocence when it comes to the quality of government data and government work. Nevertheless, upon this beacon of data reliability (or unreliability as the case may be), an intelligence layer is grafted:

Just above the database layer is a piece of business intelligence software called Supply Chain Intelligence (SCI). It can conduct basic analyses as well as complex, predictive routines.

An important step has been forgotten. Its an engineering hangup. When it comes to important things, engineers don't trust themselves with their own design. Even with generous design factors applied generously across the design at hand, engineers feel the need to add one more layer into the mix - multiple redundancies - the ultimate "What if shit happens?" safeguard. Analogously, any attempt at creating a business intelligence layer ought to cull insights and reports from multiple data streams and compile results that can be independently verified from different angles. One of those angles is for the business layer to check against a set of normal expectations when it comes to data integrity.

Of course, merely having the data doesn't mean a company knows what to do with it. Managers must ensure that KPIs are monitored and measured on a routine basis, says Michael LaRoche, supply chain strategy practice leader with IBM Consulting Services.

You would not believe how true the above statement often is. (Multiple hat tips to Robert Bowman). Too often, data sits within multiple databases and excel spreadsheets with nary a medium to make it to the decision maker. Thus, often times, decision makers are only directionally correct and worse still some of them make it a matter of honor that they have taken important decisions under such duress.
An important point to note:

The campaign must also be extended to external supply chain partners, such as suppliers, carriers and distributors. But cooperation should never be assumed, says LaRoche. The company has to make the case for collaboration. Would-be partners must possess what he calls "the four Cs": compatibility, with each side deriving benefits from the relationship; commitment on the part of all senior managers; the capability of both sides to make use of the data provided; and control, with the role of each party in making key decisions clearly delineated.

Continuous improvement, whether they be of the Kaizen, Lean, Six Sigma or any other variety depends a great deal on assessing the current situation i.e. understanding reality. To that end, collecting, organizing and disseminating accurate data about the current state is critical. My opinion about the dismal state of data affairs within firms today is that it is insufficient to be directionally correct. Instead, create an appropriate sample of the dataset with parameters set and defined by the investigator himself and go about collecting the data if one cannot trust the data in the system.

Tags: Continuous Improvement, Lean, Business Intelligence, Supply Chain Management

posted by AJ @ 9:46 AM