February/2014 Process Intelligence: An Exciting New Frontier for Business Intelligence Claudia Imhoff, Ph.D. Sponsored by Altosoft, A Kofax Company
Table of Contents Introduction... 1 Use Cases... 2 Business Benefits... 3 Things to Consider... 4 Getting Started... 7 Summary... 8 1
Introduction Business intelligence (BI) has been used for decades to improve a company s relationship with its customers, build better products, open new markets in appropriate areas, and increase revenues while reducing overall costs. In many cases though, the analytics measured a point in time, often a key performance indicator or metric around an activity. However, companies don t perform just one activity everything they do is a process or a set of related activities (simple to very complex) performed to accomplish a goal or target. It may be to fill an order, resolve a customer s problem, create and execute a marketing campaign, monitor a call center, or ensure appropriate care in an emergency room. The next leap forward for BI then is to extend its capabilities to monitor, analyze, and improve an organization s critical operational processes. This extension is known as Process Intelligence. In a nutshell, it is the ability to access data about individual instances of a process for example, fulfilling a particular customer s order, tracking the care of each emergency room patient, or following a specific marketing campaign from creation through its execution. Each process instance s activity yields data that can be analyzed, then combined into the data set for that process instance, and finally aggregated to create the process analytics for the overall process itself. Traditional BI technology can help here but more analytical capability is needed. The data points must be further analyzed in the context of how they relate to each other to be truly useful and actionable by the organization. Creating this context is beyond the abilities of most BI implementers and does not exist in most BI technologies. It takes software developed explicitly for this purpose to layer the business logic over the analytics to provide needed context. But process intelligence must give you more than just process analytics. Intelligence without the ability to act on it is frustrating at best and useless at worst. Process intelligence must also include process compliance or information about the quality of each activity performed in the process instance and in the overall process. It should be obvious that BI is certainly needed but it is not sufficient to achieve these two capabilities. We must extend BI beyond data intelligence to enable it to fully support the new frontier of process intelligence. This paper will define process intelligence further through two use cases (process analytics and process compliance), describe the business 1
benefits garnered from it, give the reader a good idea of the challenges he or she will face and what technological support is available, and finish with a section on how to get started. Use Cases The concept of process intelligence can be broken into two use cases, each demonstrating a different aspect of the capability. Process Analytics As mentioned in the introduction, every activity performed within a process instance produces data. This data can be gathered, analyzed and used to answer myriad questions about the activity, the process instance and ultimately the overall process itself. These include: When did the activity start, when did it end, what was the next activity, how many instances used that activity, who or what system performed it, what was the average time for each instance, and so on. In addition, process analytics must also calculate the statistics (e.g., min. max, mean, mode) and discover patterns of process instance execution under certain conditions. BI can be used to create discrete analytics such as key performance indicators (KPIs) or operational metrics but process intelligence must go further. It must be able to analyze the different stages of a process, and then chain them together for comparison or other purposes. From these, a company can determine the time it takes to perform each activity in the process, the volume of instances being handled end-to-end and the productivity of the individual or system performing the activity/process. From a statistical study of this data, a company can begin to predict future bottlenecks or problems. This is certainly valuable information for determining leading indicators within each process. Understanding these real-time and leading indicators quickly means that the organization can take immediate correction actions to improve employee and customer satisfaction, avoid potential problems or mitigate real ones, and improve efficiencies by streamlining processes. Process Compliance The analytics generated from process instances is a big part of process intelligence. However, perhaps even more important but 2
less easy to obtain is information on how well the processes adhere to standard operating procedures (SOPs) documented by all organizations. These SOPs are the recommended or mandated processes a company develops to ensure that important processes are executed with all the critical activities performed in a timely fashion and in the proper order. Being able to analyze whether these expected behaviors were in fact done is difficult if not impossible to track without technological support. This technological support is not found in most BI technologies but is an important use case for process intelligence. Unfortunately metrics alone can only do so much. Process compliance detects when the flow of activities deviates from the recommended pathway, if activities were skipped, which ones were repeated or performed out of sequence, even when new activities were introduced unexpectedly. Without this intelligence, it may be difficult to truly improve an organization s processes. Business Benefits Understanding internal processes yields a number of benefits to any size organization. Specific benefits include: Improving the effectiveness of a process. An organization can see exactly where waste, inefficiencies and loss (in time, effort and resources) are occurring throughout the process. Actions can then be taken to eliminate or at least mitigate these negative impacts or bottlenecks. The organization also gets excellent information on orphaned processes that is, when an activity gets set aside for further processes later but is then forgotten about. Understanding when and where processes span multiple operational systems. Many times, it is at these technological interfaces that a business process or activity within the process often encounters a delay or increased probability of an error to occur. Knowing immediately when a transaction encounters an interruption means the organization can react much faster. If you know where these weaknesses are, you know what to do to fix them. Determining how often the recommended pathway is followed. Process intelligence determines how often the proper set of activities is performed, how many times exceptions to the process occur, when activities are skipped or repeated, and when new activities are introduced. This information can be used to highlight 3
unusual patterns of behavior as well, which could indicate a significant ongoing problem like internal fraud. Creating fully documented workflows. Many organizations have a difficult time capturing all the activities performed in large, complex processes. Using process intelligence, allows the company to get a fully documented workflow that can be used to determine where they have the most risk, exposure to unwanted behaviors, exceptions they should investigate, and even activities that were not documented, are unknown, or were unexpected. Full auditing the process from end to end. Not only do organizations need critical processes to be fully documented but they must also understand who specifically performed the activity, when they performed it, what data they accessed during the process, even how often they reopened a closed transaction or case. For compliance, security and privacy purposes, this detailed audit trail is mandatory. It allows the organization to determine if this behavior was potentially breaching privacy and security procedures or was a legitimate reassessment of a situation. The faster an organization can detect and determine the purposes of these behaviors, the less risk they encounter. The bottom line is that organizations with process intelligence perform better, react to problems faster, make better decisions, and do more with less by reducing waste and eliminating errors. Things to Consider There are two areas of consideration: the challenges to achieving process intelligence and the technological requirements to support it. Challenges The benefits of process intelligence are obvious, yet many organizations have difficulty implementing it. Here are a few of the challenges that may be encountered: 1. Perhaps the biggest challenge for most organizations is a lack of understanding of the internal workflows especially when processes traverse technological barriers (e.g., when multiple systems are involved) and/or organizational ones (e.g., when there are handoffs between multiple departments). The understanding may be limited to the fractured operational knowledge as understood by each 4
individual within a department performing part of the overall process. Piecing the various sets of disjointed activities together can be quite difficult. 2. Even if the overall process is known, the underlying systems may not be set up to provide a means of determining the state or status of each activity or supply the data needed to create the performance metrics. Furthermore, processes change over time so the process intelligence capability must be able to embrace these changes quickly and efficiently. 3. Many processes are complex with multiple recursions and branching. It is rare to find a critical process that is straightforward in a linear sense. Process intelligence technology must gather statistics about the various twists and turns that occur in the end-to-end set of activities. 4. Organizations all have BI tools at this time. These tools can do a portion of process intelligence but not all. Unfortunately, many consider their BI capabilities good enough for process intelligence not realizing that, while the technology is needed, it is not sufficient. Organizations must have the process compliance capabilities to obtain the full promise from process intelligence. Technological Considerations Many of the challenges to achieving process intelligence can be overcome with the proper technological capabilities in place. BI can give you some of these analytics but it is not sufficient. Process intelligence must also include the following: 1. Ability to discover operational processes. The technology must be able to detect processes even when they cross multiple system platforms, are unknown or poorly and/or incorrectly defined, are quite complex with loops and branching, or have no system of record for central administration of the activities. The technology must also allow business users to manually define processes and update them as they change. 2. Provide continuous access, analysis and reporting on activity data. This must be in real time on real data not simulated data to generate the appropriate and timely metrics on the integrity of each process instance. The easy creation of 5
appropriate metrics and analytics is a mandatory part of this consideration. 3. Drill down from aggregated metrics. It should be obvious that to act on process intelligence, business users must be able to drill down from the aggregated process measurements into each individual process instance s metrics. The ability to act on this intelligence is dependent on knowing exactly where exceptions are occurring, what transactions are affected, which customers were impacted, and what actions should be taken to resolve the issue. Individual metrics on the process instance activities will guide the decisions made. 4. Be able to handle massive volumes of process instances. Organizations with high transaction volumes are challenging for process intelligence technology. Retail, financial, telecommunications, and health care organizations are high volume transaction environments. The technology must be proven to work seamlessly in just such companies. 5. Display of process intelligence through dynamic dashboards. Clear visualization and graphical mechanisms in dashboard designs enable the business user to quickly understand the situations as they are unfolding or in a timely manner to take action. Furthermore, the historical statistics must be presented so they can be easily consumed and understood. For example, a Swim Lane type of visualization provides an immediate comprehension of the actual workflows for process instances as well as the frequency of all variations from the expected paths. 6. Creation of alerts or alarms. Business users must be immediately notified when activities or process instances have breached preset thresholds. For example, in telecommunications, if a service level agreement (SLA) is about to be breached regarding levels of service, the user should receive an alert to that effect. He or she can then react to correct the problem before the SLA is broken and before the customer even realizes there was a problem. 7. Use of non-invasive technology to collect the entire end-toend process intelligence. We have had technology for many years touting to model and analyze an organization s processes. Unfortunately, many of these did not incorporate process analytics as part of a BI platform nor did they support 6
critical operational (real-time) analytics. These technologies had to pass their data and model definitions to yet another set of tools to execute the full set of analytics. Today, we have fully integrated capabilities that offer one place to get all the answers for process intelligence without ripping and replacing software components or passing information to other technologies for processing. Altosoft (the sponsor of this paper) is one of the few companies offering this type of noninvasive and self-contained technology. Getting Started Many organizations believe that they must fully document their entire set of processes before starting a process intelligence project. This is a fallacy. Most organizations can get started by making intelligent guesses or roughly documenting their workflows. While the organization must have a good understanding of what a process is and how activities fit together to form the process, the technology will highlight the real flow of activities and even find those that were not documented. The second step is to begin collecting the data from each activity. Use technology that makes the collection and processing of this data easy, including process discovery and activity monitoring. These sets of data will feed into the strategic and individual KPI calculations. The third step is to analyze, explore and interpret the initial set of data. This serves as a baseline of the as-is processes. Without this baseline, it can be very difficult to show measurable progress. Without measuring where you are, how do you figure out where you want to go? The fourth step is to begin correcting specific activities within a process and repeating these steps to collect new metrics and compliance information. Corrections may include retraining of key personnel, establishing new or updated SOPs, creating new metrics, changing operational systems, updating security procedures, or other actions. Each time the data is analyzed it should be compared back to the previous measurements and to baseline analytics. Over time, enough data will be collected and made accessible that it will become possible to predict when and where processes may break. As volumes grow or systems age, the organization can make timely changes to avoid bottlenecks, inefficiencies, and wasted efforts. In addition, as process intelligence grows within the organization, it may prove beneficial to establish a center of competency or excellence to 7
manage the overall process intelligence efforts. Finally, process intelligence is not a specialty, limited to only a handful of dedicated business users. It is for everyone in the organization who performs or monitors any of the activities in the critical processes. However, for adoption to grow, the organization must invest in educating its workforce and evangelizing the benefits from process intelligence. Since it is a relatively new area for many organizations, this education will be mandatory and should occur at regular intervals as new processes roll into the process intelligence domain. Summary Organizations everywhere are struggling to become more efficient, enhance customer and employee satisfaction, and reduce significant costs from fraudulent behaviors, rework and waste. Now organizations have technological support to help them understand and optimize all aspects of their performance. Process intelligence gives organizations timely metrics from which to improve operations by determining the root causes of problems. It can even predict problems before they impact the organization or its customers, ensuring better compliance with SLAs and regulatory agencies. Process intelligence uses the power of BI technology for metrics measurement and then extends it to understand how processes and workflows really work, where bottlenecks occur, and what exceptions or behavioral trends may be putting compliance and risk management in jeopardy. In today s business environment an organization simply cannot afford to not have this important and now easily attained intelligence at its fingertips. 8
A word from our sponsor Altosoft is driven by the belief that the successful use of business intelligence is critical to helping organizations manage today s complex operational challenges. As described in this white paper, Altosoft also agrees that process intelligence is the next evolutionary step for the advancement of BI which can provide new tools for addressing these challenges by extending BI to help accurately understand the business processes that drive every organization. By clearly linking business data and information to its related business processes, process intelligence provides the insight necessary to understand how processes and, more importantly, the operations they represent are working, where bottlenecks may exist and what types of exceptions could be putting compliance and risk management to the test. In addition, Altosoft is committed to its mission to fundamentally change the economics of delivering BI solutions by making BI more accessible to a broader range of users and dramatically shortening implementation cycles. To achieve these goals the Altosoft platform has been designed to support the development of enterprise-class solutions with zero coding required while not compromising on power or scalability. To learn more about Altosoft s products including its process intelligence capabilities please visit: www.altosoft.com. The comments reflected on this page are those of the sponsor organization and are not intended to be an endorsement by the author or Intelligent Solutions, Inc. 9