Trask has been implementing BPM (Business Process Management) and business process automation using software technologies for a long time. With this experience we can confirm the statement is correct: Yes, the work is not done after a processing system is just implemented and introduced to the company. On the contrary, the running processes must be monitored and measured, partly in order to see if the processing machine is working as intended and is producing the expected results, but also in order to find out where we can improve and accelerate it. This is how the BPM lifecycle is completed: process design - process operation - measuring and monitoring - improvement proposal.
A gauge for processes
So what do we measure processes with? It is simple - we use a gauge or, as we say in the BPM field, a business indicator. Using our business indicators we can measure almost anything. We would certainly like to see the company economic results. Also we want to know everything about the state the processing machine is in. That is why we watch the number of currently processed tasks and exceptional situations that are being solved. Measuring performance using KPI is a whole different chapter. Key performance indicators, besides doing the measurements, help us compare the current performance to the selected target value we try to reach - for example, you could see it must take maximum two days on average to process a mortgage request and that everything is fine as it now takes 1.3 days.
Business intelligence vs. business activity monitoring
Until recently, process measurement was only the privilege of DWH (data warehouse). Applications send data extracts and logs to DWH who calculate different indicators and display them in various reports as a part of business intelligence. However, these reports are only based on historic data. Therefore, if an exceptional situation occurs, it would be visible in the next day's report at best.
No more outdated reports, I need operational control!
A new approach towards business indicator measurements - business activity monitoring or BAM - was created as a reaction to such needs. Unlike BI where the analysis is based on historic data, BAM works with real-time events. Once a situation occurs that is relevant to a given business indicator, the system sends a so-called event with information about the situation. This is received by a monitoring tool that processes the event immediately. For example, we can learn that time to process an important task is just over and we can come with an instant reaction. If there are pros, there must be some cons and BAM is not an exception. For BAM to work, the monitoring tool must know whether and how to use an event at the moment it receives it. It must know the context the event came from and the business indicator it relates to, which is an extremely difficult task due to the number of generated events. The monitoring model must be a partial copy of the processing model. Therefore, whenever we perform a change in one of the processes, we must take into account the impact of this modification on monitoring. Lucky for us, most of the monitoring tools offer functions for easy synchronization of both model to remain consistent.
BAM is a complement to BI reporting and operational management inputs
However it may seem, BAM is in no way a successor to BI - both the approaches are complementary. BI tools are priceless when you need to analyze complex indicator patterns in order to get a clear view of the economic status, irrespective of the historic data. On the other hand, when it is necessary to analyze and react on the situation immediately, BAM tools have to be introduced.