There are many business process management (BPM) solutions that can help create and manage workflows for all business activities, but only some of them are going one step further to offer BPM simulation functionality. Such solutions can be used to simulate different scenarios for business processes changes and determine the potential effects they might have on a company's activities.
Lanner is one of the vendors providing this type of business process simulation software. I had a chance recently to speak with Geoff Hook, BPM solutions director at Lanner, and better understand Lanner's offering for business process management simulation.
What Is Lanner and What Does It Offer?
Formed in 1996, Lanner Group was originally named BL Systems, and was a spinoff (consisting of merged IT departments) of the British Leyland car manufacturer. BL Systems created the world’s first commercially available simulation package in 1978, which was officially released under the name of SEEWHY in 1980. Further development led to the release of WITNESS, launched for IBM PCs in 1986.
In the last decade, Lanner Group has released simulation packages for niche industries such as police and criminal justice (PRISM), pharmaceuticals (PX-Sim), and a version of WITNESS for service industries.
A Java-based simulation solution was released in 2006 under the name of L-SIM. A partnership with IDS Scheer (now owned by Software AG) was established to integrate L-SIM and IDS Scheer's ARIS (a product that is white-labeled by Oracle and SAP).
Similar integration partnerships for L-SIM and WITNESS exist with vendors such as PNMSOFT, IBM, and Trisotech.
Lanner's more recent offerings include apps for business managers who do not have extended expertise in simulation.
Lanner also provides consulting services that aim to help customers benefit from external expertise, an extensive library of simulation templates, industry best practices for BPM simulations, and assistance in creating simulation models for customers' specific needs and requirements. Lanner is also involved in research initiatives through its Academic Partnership Programme and is a founding member of the Workflow Management Coalition, which aims to create a standard for simulation data.
The Lanner approach to business process simulation is based on the idea formulated by Michael Hammer in the book Reengineering the Corporation: A Manifesto for Business Revolution, where he states that the goal of any business process improvement initiative should be to radically improve the activities in the company or even revolutionize them. He also says that business process improvement should be a constant activity in any company, in much the same way that market change is constant.
What Is BPM Simulation and How Does Lanner Do It?
Naturally, it's wise to test business processes before implementing them, for several reasons. First of all, new processes will cause disruptions in company activities; decision makers cannot afford to risk these disruptions unless they’re pretty sure that the change will benefit the company. Also, any business process improvements performed after implementation can cause further disruptions, even putting the entire initiative at risk of failure. Finally, there is always a difference between the estimated impact and the real impact of changes made to a company's business processes, but simulating the new processes can provide a closer approximation of the impact than theoretical analysis.
When performing business process simulation, companies have two options:
- real-life scenarios, where the new business processes are implemented in selected areas of the company and work in parallel with existing processes, or
- software simulation, which allows them to create a virtual implementation and track the results.
Both options have advantages and disadvantages, as summarized below:
Besides the employees, customers and business partners are also involved in the simulation.
Unpredictable changes in human behavior (which are hard to reproduce via software) are factored into the simulation.
Sophisticated algorithms can analyze high volumes of activity without involving many people.
This approach is less disruptive, since the processes being analyzed don’t need to be used by people in parallel with daily tasks.
This approach creates disruption for employees, partners, and customers.
This approach has a limited reach, since not everyone participates all the time (only a limited number of employees are involved).
This approach is limited in terms of human behavior simulation, due to behavioral unpredictability.
The required tools (for simulation as well as data analysis) can be expensive.
Generally speaking, software simulation is more useful for medium and large companies that already have (or can afford to invest in) the technology required. This approach also offers more flexibility when it comes to creating a wide variety of scenarios, and also allows for easier data gathering and analysis.
Lanner specializes in discrete event simulation (DES), which simulates series of activities based on predefined criteria (e.g., time of the activity, time between activities, frequency) and scenarios. In contrast to real-time computer simulations (where an activity that takes 10 minutes in reality will also take 10 minutes to simulate), DES skips to the next event following the pre-defined order. Another important characteristic of DES is the fact that it can generate pseudorandom numbers, which allows users to rerun simulations with identical sets of random numbers. This permits an apples-to-apples comparison of different scenarios by changing only certain variables, while keeping other parameters constant.
For instance, a simulation can be created for a certain number of users performing tasks that take a certain amount of time; for an effective comparison of simulation scenarios, the number of users and the amount of time for each operation should remain constant, while other variables can be changed (such as the time between each task or the efficiency of the user).
Lanner defines four classes for simulation models, depending on the maturity of business processes, the level of involvement of the users, and the reach of the processes within the company. Class one is used for specific cases and cannot be reused for others; class two simulations can be reused, but versioning is hard to track; class three requires software solutions to manage the process and the data; and class four is basically part of the enterprise DNA, simulation being applied all across the company and most employees being involved.
When simulating business processes, some extremely important factors can only be managed using software solutions. Such factors include variability (probability of change over time), disruptions (exceptions to normal behavior), and interaction complexity (processes running in parallel or separately, which can be interconnected or not).
Traditional spreadsheet-based simulations offer limited flexibility to include exceptions, thus forcing users to use averages for the variables defining the business process being simulated. The problem with averages in simulation is that they rarely offer a good estimation of the reality, since they do not let you see what happens when exceptions occur. An example of a pharmaceutical company trying to determine the overall efficiency of an equipment by simulating breakdown and repair times shows that failure to include variability and disruptions can generate an overestimation of equipment efficiency by 73 percent. This is because the "safest" scenario (i.e., where time to breakdown is regular and time to repair is fixed) is also less probable, since machines break down at different times, and the length of time it takes to repair them is never constant.
Lanner also provides its Optimizer technology, which can be used in simulations to find the best solution to improve business processes, not only based on the criteria used for simulation, but also on previous simulations. The Lanner Optimizer actually uses advanced search algorithms to learn from previous simulations, and uses that information to determine optimal parameters for new simulations (for instance, based on a previous calculation showing that a customer needs at least one minute to enter an online order, the system will automatically set this parameter for one minute or more in future simulations). An intelligent simulation system should also be able to "understand" the objectives of the users. Ideally, all simulations aim to find the best answer to all problems, but the reality is often far from ideal. As Geoff told me during our briefing, sometimes "the very best answer is the time it took to get an answer." In other words, you need to find a balance between your objectives and the effort it takes to obtain them.
No matter what industry you're in, the more complex your business gets, the more you’ll need business process simulation. Medium and large companies already using software for manufacturing or business intelligence will benefit the most from simulation technology because they already have a lot of data that can be used, and have at least some of the IT infrastructure required.
The Lanner family of products for business process simulation can address the needs of various types of users: WITNESS for simulation professionals, L-SIM for embedded simulation within BPM solutions offered by partners, and apps for business managers. Lanner's various products have also been adapted to the specific needs of a wide range of industries, including automotive, aviation, criminal justice, aerospace and defense, financial services, food and beverage, health care and pharmaceuticals, logistics and supply chain, manufacturing, and energy (oil and gas, and nuclear).