One of the most important steps -- and biggest challenges -- when optimizing a business process prior to automating it lies in figuring out how the process currently works.
A variety of business process modeling (BPM) techniques, such as process mapping, have been developed over the years to document business process flow, but these often require painstaking manual inventories. Process mining, powered by specialized algorithms, emerged out of the BPM community to provide a set of tools to automate this work.
More recently, value stream management -- a set of techniques and tools for automating process improvement -- has emerged out of the Lean Six Sigma movement.
Both process mining and value stream management techniques capture data from event logs to build a more comprehensive picture of a particular process. But value stream management (vs. process mapping) is also able to capture more context from the apps used by employees in a business process and see how -- and if -- usage correlates to business value.
"The biggest difference is that value stream mapping is focused on driving process change, while process mapping, mining and discovery are well-suited to making existing processes more efficient," said Sylvie Thompson, associate partner of the digital supply chain practice at Infosys Consulting.
Value stream mapping vs. process mapping for RPA
The debate over when and how to use these respective approaches to improve business processes has come to the fore as more organizations deploy robotic process automation (RPA) software to automate mainly rules-based, repetitive tasks. Many organizations use the information obtained during process mining to look at metrics such as frequency of use to select the steps that will deliver the biggest value if RPA is introduced.
However, "These initiatives rarely look at the opportunity to execute tangible process improvement," Thompson said.
Andrey MihailenkoCEO and co-founder, Targetprocess
Andrey Mihailenko, CEO and co-founder of Targetprocess Inc., which makes Agile project management tools for the enterprise, agreed that value stream management tools can help organizations look beyond using automation tools to simply make a current process more efficient.
"Value stream mapping [vs. process mapping] is not about a specific process but is centered on analyzing the holistic big picture view of how value flows to a specific customer," he said.
Value stream mapping's goal is to provide greater strategic insight into how to deliver value to a customer, Mihailenko said, rather than giving detailed specifics on tactical implementation and improvement for a specific business process. As such, he argued that value stream management is good at capturing and analyzing various indicators relevant to building digital products.
For example, value stream management tools often plug into software development and deployment tools in a way that makes it easier to correlate the total time it takes to create a new feature. The tools can also capture information from customer support tools to track the kinds of issues these features raise for users. This data can then be correlated with the business impact of new features, like improved usage.
Process mapping basics
Process mapping is simply the visualization of a workflow, including the materials, services and information involved. It documents the various steps from start to finish with no indication of the significance of each step. For example, step two is no different from step three regardless of the time, level of effort or value added during that step.
Process mapping can occur at various levels of a workflow.
"Only at the lower levels of process mapping can you begin to identify areas of improvement based on flow," Thompson said. For example, redundant work within the process can be reviewed, modified and potentially eliminated.
Various automated process mapping solutions have emerged that use large amounts of transactional data to create process flow maps with key statistics for each step in the process flow, Thompson said. These solutions allow for rapid process mapping.
However, the solutions are limited in that they can only map those steps that have set data transactions, and not all the steps within a process flow are tied directly to a set system transaction, Thompson said. Nor do these tools consider the non-system efforts that make up a business process.
For example, a process such as approve purchase order may only take a few seconds from the time a user opens a PO to when he clicks the approved button. However, prior to that transaction, other business process events likely took place that gave the user the confidence to approve it.
"It is those efforts that are missing from automated process mapping solutions," Thompson said, stressing that process mapping and mining tools provide an "outstanding starting point for a process mapping exercise, [but they] do not fully replace the efforts needed to effectively process map."
Value stream mapping vs. process mapping for process improvement
Value stream mapping (vs. process mapping) aims to provide insights into each step of the process. It is used to clearly determine which steps can be eliminated, refined, consolidated and so forth, while at the same time identifying which steps call for investments to further enhance the value they bring to the end user or consumer.
By its very nature, value stream mapping (VSM) puts a value on each step and recognizes that not all the steps are equal, Thompson said. Further, VSM looks at lead time, which is generally defined as the time between the initiation of any process and the completion of that process -- a metric that is extremely helpful when making cycle time improvements.
The automating tools used in process mapping -- process mining and process discovery -- also complement value stream mapping, providing tangible metrics that can improve the value stream map.
"Process discovery has the added benefit of uncovering actual process flow patterns, which may differ from a value stream map, allowing you to investigate the variance intelligently," Thompson said.
What to measure
The primary metric used to determine the efficiency of value streams is end-to-end cycle time, said Victor Wu, product manager at GitLab BV. Time to value is money. AI and other analytics techniques should be used to accurately measure that end-to-end cycle consistently and correctly at a minimum. AI can also be useful to find hidden patterns and detect anomalies, especially early warning signs that a project may end badly, he said.
Additionally, measuring the time it takes for each stage is important to help optimize the process, as process improvements can be made at individual stages, which are, in turn, often controlled by individual teams in an organization.
GitLab, for example, focuses on measuring end-to-end cycle time as part of its value management process because it simplifies product iteration and customer feedback.
"There's a major risk when organizations spend a long time designing and implementing a supposedly great feature because, during that time, customers are not experiencing your most updated business ideas -- and not validating them," Wu said.
The faster you can experiment with new features, the faster you can find out if a business hypothesis is accurate.
Change management required
Value stream mapping, whether complemented by process mining or not, is a fruitless exercise unless management is committed to the diagnostic and treatment phases.
Targetprocess' Mihailenko said that, unfortunately, in many cases, management provides the budget for the value stream mapping exercise, but does not get behind the implementation of the required changes. The whole process grinds to a halt after the initial work uncovers some areas for improvement, but the changes aren't made.
One big challenge lies in breaking down silos between business units. There are often long-standing cultural or time-zone barriers in large enterprises -- e.g., marketing is in the U.S., support is in India, product dev is in Europe -- that are difficult to overcome. Each group needs to understand that the overall goal is to optimize the entire value stream -- not just one activity in the value stream, Mihailenko said.
Real change requires some manual work, as well. Not many enterprises are up to this task. Infosys' Thompson said, "Unfortunately, few organizations take the time to build out effective value streams and tend to jump directly into these new analytic solutions."
But without value stream management, the value of the information provided by automated process mapping tools is significantly reduced, he said.