Oops! The input is malformed!
Originally published 5 September 2007
Traditionally business intelligence (BI) has been used for historical analysis, helping the corporation with strategic decisions. Typically, information is held at such a level that there isn’t any operational data on which to base decisions. However, recently there has been more attention to operational data as a source and the need for proactive business intelligence.
In the following paragraphs, I’d like to focus on proactive BI and how it can be used within a manufacturing environment.
Reactive BI or strategic BI, as it’s commonly known, is exactly that – it’s consumed, analysed, digested and reacted to on a weekly, monthly or yearly basis. As we all know, it’s used for historical analysis, looking at trends and making decisions at senior management levels and higher. Reactive or strategic business intelligence can clearly indicate shortfalls in production of goods, poor sales targets and not achieving distribution deadlines. What it doesn’t help us with is how to avoid these situations. Can proactive BI help?
Figure 1:Typical Business Intelligence Architecture
Figure 1 shows a typical business intelligence architecture where data latency increases the closer you get toward the data warehouse. This is mirrored by the use of operational and strategic data.
Reactive business intelligence applications and technologies can help companies analyse the following:
The world of IT has been talking about operational BI for quite a while now, but the idea of near real-time intelligence at an operational level has not been implemented at many manufacturing enterprises. The main objective of using intelligence in a proactive way is to identify a potential problem that has a detrimental impact on business so that frontline operators and department managers can address the issues before they arrive at our customer’s door. Having a forward-looking perspective allows frontline workers to have information that empowers them to make production decisions at the right time, rather than in two, three or four day’s time.
As the name suggests, proactive BI is applied at the operational level. Somewhere in the business, there is a fuzzy area where strategic BI ends and operational BI starts. This generally sits just below department level.
Figure 2: Enterprise-wide BI Usage
As can be seen from Figure 2, the higher up the corporate ladder, the more strategic the business intelligence. The more strategic the BI, the less impact on operations in the short term.
Operational BI is about now! It is about what’s happening to your manufacturing process at the current time:
The above points may seem to be about process data rather than business intelligence, but if you combine operational BI with BAM (business activity monitoring), then you can build an event-based architecture that will inform the operational staff of anomalies during the manufacturing process. Take the following examples.
A production line manufactures widgets. At the end of the line, there is a counter that measures the speed of production. All this information has been passed into the data warehouse on a regular basis, and we now have dated information for the product. When the speed of production drops below a certain level, business activity monitoring triggers an event, reads a data source for the product and presents the information to the line manager or shift manager so that they can make an informed decision. It would be even better to have information on the line’s running configuration so it could be determined if there are any issues with it, thereby pinpointing the problems.
Another common use would be to monitor orders. If we have a customer that orders on a daily basis and the orders are always of a similar quantity, then we can adopt the combination of BI and BAM to alert managers of a dip in order quantity. This would avoid 300 being delivered instead of 3000. Let’s face it, typos still happen even with today’s technology! If it was a genuine drop by the customer, then it gives the sales team an opportunity to visit the customer and find out if there is a problem with service. I think the message here is that the benefits aren’t just for supply chain. There are extended benefits outside the immediate business area.
Although we are looking at an operational level, the strategic benefits are huge. If we can get the manufacturing process improved by using operational BI, then key performance indicators (KPIs) like “on-time, in-full” will improve, obviously affecting the customer satisfaction figures and reducing the need for reactive BI by the senior management.
Continuous improvement of the underlying business process by using quality business intelligence facilitates informed strategic decision making when revising the business objectives for the coming year(s)
Business intelligence tools are very useful when deployed on their own, but they can be complemented and significantly improved by other technologies to provide the business with proactive operational-level information. Following are some of the technologies (along with a brief description) that I feel are the most useful.
BAM is an easy concept to understand, and Figure 3 shows a typical BAM / BI architecture. You identify an activity in a process or an attribute within the activity and monitor it until it violates a business rule. For instance, as mentioned earlier, you could monitor orders. When it breaks a rule, notify someone.
Here are some examples of BAM usage:
Notification services such as Microsoft’s SQL Server Notification Service (SSNS) are used primarily to monitor databases for events or changes and then notify a person (by e-mail), a device or an application. Monitoring could take place on the data warehouse or an application database such as an error log where, for example, it checks whether there are any critical errors.
A business rules engine (BRE) is a very useful tool when it is complemented with BAM. BREs are used to store business rules and can usually be split into two distinct types. There are decisions and logic. BAM/BI integration is interested in decisions, such as: if order is greater than 1000, provide a discount of 10%, or if Plant A has error type 123, then do the following ……!
Most business rules are hard-coded in the existing systems, meaning that a rule can’t be accessed by multiple applications and that rules have to be continuously re-written. The rules in a BRE are abstracted into a database so that they can be reused by any system.
Figure 3: BAM / BI Architecture
Strategic or reactive BI still has its place within the business and will always be the domain of board-level executives and senior managers. However, operational or proactive BI has a more important role within the business as a tool to help provide our customers with the best service on a day-to-day basis. Unlike strategic BI, operational BI isn’t confined to the data warehouse, but extends itself to the complex world of operational applications and data stores.
Business can, and should, be complemented by other techniques and technologies, in particular BAM, BRE, workflow and notification services. These can leverage the business process to provide a powerful event-based architecture to proactively manage business-to-customer (B2C) service, ultimately improving the business process and reducing the need for a reaction to customer complaints.
Recent articles by Gary Holden