Author Profile - Mike Darrish is a certified Lean Six Sigma Black Belt and Industry Specialist at OpenConnect Systems, where he works on teams delivering process improvement in large enterprises. Prior to OpenConnect, Mike worked for over 25 years in large enterprises and startups providing IP connectivity, network management, IT performance improvement and financial services. His roles have included programmer, systems administrator, network analyst, instructor, technical and management consultant, and a variety of sales support and sales management positions. Mike lives in the Atlanta, GA area.
World class methodologies such as Lean and Six Sigma are well known, proven approaches to process improvement in the manufacturing sector of the US economy. As a result, according to the US government's Bureau of Labor Statistics, manufacturing productivity doubled between 1987 and 2007. Since these methodologies were adopted much later in the services sector, productivity there increased only 22% in the same time period. See Figure 1.
One area of process improvement, for knowledge workers, used here to mean people using networked computers and other digital systems, performing financial services activities, customer support, IT services and the like, has been a particularly intractable problem. In the past, a common approach to improving these processes has been to send consultants out with clipboards and stopwatches, stand behind the user and record activities, timing and process data, and then, back in the office, to draw a process map, perhaps capturing and printing out a few screen shots for process documentation. There are a number of problems with this approach. First, users at a work station behave differently when they are actively being watched. While watched, these workers tend to be more productive and typically don't engage in personal activities like surfing the web or chatting with co-workers like they might otherwise do. Then, when the observation period is over, they generally lapse back to normal behavior. This phenomenon, which presents an inaccurate view of the normal process, is called the Hawthorne effect, after a landmark productivity study conducted at the Cicero, Illinois Western Electric Hawthorne plant in the 1920's. Second, and more importantly, the amount of data observers can gather is a tiny fraction of the total computer and network-based activity that goes on in a large enterprise, leading to an inaccurate process model, lacking the details necessary for process improvement with the well known approaches mentioned above.
In recent years, several vendors in the Business Process Management space have used software to collect data from applications and networks as a more robust input to process mapping. Inputs have typically been extracts of database or other enterprise application logs. This approach is more accurate and more scalable than human based observation, but is dependent on the quality of the logs, which are generally not very granular and have the disadvantage of not having user screen captures for detailed analysis of user behavior. Additionally, each application has its own log format and content, making for a difficult data collection and analysis task.
Achieving breakthrough levels of process improvement requires both very granular and scalable data collection and storage as well as world class process improvement methodologies. While collecting process data, one must also avoid materially changing the performance characteristics of the process under study and the enterprise infrastructure. Typically it is not feasible to ask an IT department to retrofit applications with performance metrics and reporting. Ideally, then, one would use passive network taps and/or switch or router based constructs like Cisco VACL Capture or mirror ports. These collection points are realtively inexpensive, non-invasive and, for a single process under study, even in the largest enterprises, can generally can be collocated near the server farm in a small number data centers. Given government regulations such as HIPAA security, and the need to protect financial transactions, the system must encrypt the collected data in place and in real time. In large enterprises, such as healthcare insurance companies, which process upwards of 100 million claims per year, there are several gigabytes per second of data to manage, requiring powerful servers and fast storage. There is also one broad case where traffic across the network, unlike web oriented applications or the still-common dumb terminal emulation, does not accurately reflect the user's behavior. That is the fat client, where an application on a PC does significant work. For fat client applications, a distributed agent with a small footprint and a low reporting traffic rate can generally capture screen shots and field changes, forwarding the information to a central server for storage and later analysis. This approach has proven economical even in VPN-based home office workers, thanks to the widespread availability of broadband data services in most homes.
As we know from Lean and Six Sigma, in order to improve the process, one must measure its key performance indicators (KPIs), including items like user think time, system time, key data inputs and granular transformation procedures. The data that is collected as described above is the key input to a business process discovery. Once we discover and analyze the process sufficiently, we generally have sufficient information to do root cause analysis of problems, hypothesize the changes needed to improve the process, do experiments to better understand the main factors and their interactions and to measure changes to the process as we make them. Since we now have powerful computers at our disposal, it is possible to automatically map the collected data into a business process along with timestamps, userids and other user interactions, as shown in Figure 2. Having large volumes (empirical) data available, rather than the usual anecdotal data and estimates regarding the process, gives us the ability to analyze it scientifically, to graph metrics like end user response time, identify paths users took through the process, analyze the status of key indicators at various times and places in the process, whether straight-through, as we we'd like, or rather, down exception paths. We can also analyze how often and why the long tails of exceptions take place, then relate user and process behavior back to business rules for improvements.
Process metrics don't just materialize and processes have no pre-existing context, unless they were originally imbued with BPMN or similar metadata. In 2010, the vast majority of processes in use were built before BPM standards gained traction, so it is necessary for business subject matter experts to label the discovered activities (applications) with names that reflect what they are called in everyday usage. The discovery engine can then generate meaningful event-based intelligence from transitions occurring in the monitored business process.
One approach to generating business events is commonly called screen scraping, though in reality it is a fairly sophisticated form of data analysis in its own right. One must be able to analyze the data moving between client and server, whether that data is from dumb terminals, web servers or whatever traverses the network. Then one must render the data in the same way that the target machines do, in order to recreate what the user saw and how he responded. See Figure 3 for an example of this kind of data analysis, recreating the user's experience for the business analyst and providing information needed for the second use of the original data, an analysis of user and system behavior. An additional requirement for breakthrough improvement is the need to analyze and report on process data without the restrictions of traditional Business Intelligence systems with pre-defined schema, event summarization and the subsequent restrictions on analysis and reporting. In the real world, these concepts and methods have been embodied in commercial, off the shelf software and put to use in healthcare insurance companies, improving processes like claims operations, as well as in banks to improve call center operations. By discovering and analyzing process intelligence, workforce intelligence and customer intelligence, enterprises save millions of dollars annually through identification and elimination of defects and waste, especially work in progress. Often, these enterprises write robots to automate some or all of the work previously done by humans, speeding up the process and freeing people up to do more sophisticated tasks and improving customer satisfaction.
About Comprehend and OpenConnect Systems
For more information about product improvement for knowledge workers, the reader is invited to visit the web site of the author's employer, OpenConnect Systems, to read about Comprehend, a product suite which implements the concepts described above and from which the screen shots are excerpted. OpenConnect Systems, based in Dallas, TX, delivers software and service based solutions focused on improving knowledge workers' business processes.








Recent Comments