I recently chatted with Ron Webb, executive director of open standards research at APQC about the intersection of Big Data and business process management (BPM). During our conversation Ron shared his perspective on the role of Big Data in process management, how to effectively match measures to purpose, and the role of people in Big Data and BPM.
APQC: What role do you see Big Data playing in process management? What are its current limitations?
RON: Big data has the potential to ramp up our understanding of how processes perform and provide data on the impact of process entities like employees, suppliers, and machinery on process efficiency. In other words Big Data has the potential to help organization’s understand how employees influence the process and if high performers are what create high performance processes.
Unfortunately this type of information is currently limited to transactional data—tracked in an ERP or other automated system—and self-reported information on these entities. Though you can run analysis on self-reported information it’s difficult to track all the processes an individual formally and even informally touches. In centralized organizations it’s easier to pinpoint these things and hold conversations to gather this information, but in decentralized organizations it’s much more difficult to find out who touches what process.
APQC: What do companies need to do to prepare for the integration of Big Data and process management?
RON: One common mistake that most organizations make is the assumption that they need a good data scientist to get started, in my opinion what they really need is a good data plumber. In other words to get started integrating Big Data with process management, organizations need to “unclog” access to the information they need in key parts of the business. Because Big Data systems push out tons of data, it can often be difficult to identify what information is worth tracking and what information is just noise.
To pick the right information the organization needs to start thinking through what’s the goal of its process management effort. Is the main purpose of its process management efforts to improve speed, cut costs, or improve throughput? Defining the purpose of the organization’s efforts will help it start to understand what to measure. This is where a data plumber comes in. A data scientist typically has in-depth statistical expertise. While a good data plumber combines statistical knowledge with the business acumen necessary to understand what to measure and test.
Take the broad topic of customer engagement for example. The first question to ask is why your organization wants to understand what drives customer engagement. Is it to improve retention rates, create a lead scoring system, or improve customer satisfaction? Asking this question helps define what the right measures are. If the purpose of understanding engagement is to improve customer retention then you can safely determine that customer behavior measures are the right thing to track—things like content downloads, purchases, and interaction with marketing touch points.
APQC: Based on a recent survey, one of the top process challenges in 2015 is “aligning continuous improvement efforts to avoid divergence across the business.” How do you see Big Data alleviating that challenge?
RON: Big Data by itself cannot help alleviate this challenge. Organizations have to start by adopting a cross-functional process perspective. Without a cross-functional process, that includes common goals, optimization in one part of the process may come at the expense of another. Additionally, it’s important to define the process goals as something your customer cares about, if you only focus on what you care about then your process improvement efforts are likely to put you out of business.
However, Big Data helps the organization look at the processes closest to the customer first. For example, let’s take a look at understanding the process for customer purchase. The first step is to use Big Data to identify what influences a customer’s likelihood to purchase. Is it a particular type of interaction like an event? If it’s an event, then pull the data and run analysis on what makes people go to those events. Is it a specific type of email or how long they’ve been a customer? Improving those processes will drive the most value for both the customer and the organization. In other words it’s not as big a pay off if you tackle the back office stage of the process first. Instead start with what the customer values and move backwards from there.
APQC: What is Big Data’s role in process risk identification and assessment?
RON: At high level this whole approach is about mitigating risk by moving to data supported decision making, rather than making decisions based on experience or gut. The best way to avoid risk is by having experienced people use data to inform and support their decisions. Risk mitigation is also related to your understanding of which processes and measures have the highest degree of influence on your desired outcome.
APQC: Is gathering behavioral data important for truly effective BPM? If so how can organizations ethically gather behavioral data for process management?
RON: In my opinion, capturing behavioral data is the biggest goldmine for BPM and process improvement. We tend to look at processes without really facing the fact that it’s people who conduct the work. If we can connect these two things then it’s better. Furthermore it’s becoming an accepted part of the employee contract that organizations will track behavioral data for productivity measures. For example there’s an expectation that organizations will track digital logins, internet activity, and in some cases even key strokes. Anonymity, by using employee IDs and aggregating the data help ensure that organizations can ethically capture behavioral data. For example organizations can use IDs to collect behavioral data on centers of influence and the cross-pollination of ideas.
APQC: What are the advantages and limitation of integrating Big Data with process frameworks?
There are two predominant advantages of a process framework: developing common language and categorizing process into related groups. A process framework provides context about what should be measured, how the measurements are related, and provides an attribute to attach data to. Hence you can see where this person connects what parts of the process. This gives you the ability to collect data and analysis, drill in for understanding root causes and aggregating performance information up to the process, business, or organizational level.
However using a process framework limits you because you are tied to that framework’s view. For example if the organization choses to use APQC’s PCF, you get that view of your processes. It is hard, then to analyze the data against another framework, such as SCOR, ITIL, or another framework. . So the limitations are those tradeoffs that come with the framework you choose.
APQC: One critique of Big Data is that it tends to provide correlation without causation. How can organizations overcome this challenge for process management?
RON: I don’t think you are ever going to get to causation out of Big Data when dealing with large processes, but you don’t have to with process. Causation is necessary for discrete behaviors or massive cases. There are very few processes that will need to get to causation, especially when factoring in business context. Correlation analysis helps identify the potential issues, which are then root caused and addressed by process managers and the employees conducting the processes.
Want more business excellence insights, you can find Ron on twitter and LinkedIn.
To better understand and address challenges on BPM APQC is conducting a survey on putting its Process Classification Framework SM (PCF) in action. Please take a few minutes to share your insights by completing a 10 minute survey on how your organization is using the PCF, your adoption and implementation practices, and the challenges you face using it.