Value Stream Mapping - Extending for Automation

In any digital transformation conversation the topic of automation is likely to be discussed. One approach to identifying and prioritising transformation opportunities is to apply Value Stream Mapping. But how do you measure improvement over time? More specifically, how do you measure the impact of automation over time?

During my time applying value streams to continuous improvement, I've found the value stream mapping convention was lacking a succinct way to represent the impacts of automation within a value stream. This blog aims to introduce the concept of automation cogs to bridge this gap.

What are Value Stream Maps

If you're new to Value Stream Maps (abbreviated to VSM from herein), the high-level summary is that they represent an end-to-end (E2E) sequence of high-level activities to deliver value. Lucidchart provide a good overview here. For those looking for detailed information, I highly recommend reading the Value Stream Mapping book by Karen Martin and Mike Osterling.

Used in conjunction with Porters Value Chain, value stream mapping is a powerful mechanism to understand the high-level flow of your organisation, without the need to dive into detailed business process and/or technical modelling (for example BPMN 2.0 and UML).

Introducing the Automation Cog

As discussed above, the aim of this blog is to propose the automation cog as a standard means to represent the impact of automation within a value stream. The diagram below illustrates the concept, accompanied by some lessons learned and its potential usage.

Initial Concept: Percentage Automated

The original intention of the automation cogs was to represent the percentage of a value stream process that was automated. It didn't take long to realise this wasn't the best approach as:

Automation coverage (%) is difficult to measure due to the variations a process will go through in practice (the underlying decision gates and flow of information). For example, an automated approval may only be automated for certain teams or use cases.
A target of 100% automation is not realistic, many processes rely on human interaction, automation cannot always replace the power of human skills and knowledge.
Introducing automation takes time, it requires investment design, development / testing, and organisational change to facilitate adoption. In this context, automation coverage is a number that will change slowly, negating the benefit of representing within value stream processes.

Improved Concept: Automation Impact

Taking the lessons learned from the initial concept, I found a much better approach is for the automation cogs to represent the percentage impact of automation. This is a more natural fit, as we have existing metrics such as lead-time (LT), processing-time (PT) and percentage complete and accurate (%C&A) that will change over time.

By introducing the impact of automation, it provides further context, to be used in conjunction with the existing metrics. Shown in the diagram below, the measure can be represented in a few different ways:

At the process level (left), the impact of automation is measured as an overall impact to a value stream process. This is valuable when trying to communicate the message in its simplest form. It is recommended this metric represent processing time (PT), as it is difficult to "roll-up" multiple metrics into a single aggregate percentage.
At the metric level (right), the impact of automation is more granular. This is valuable for those trying to identify more specific impacts within a process. By pinning the automation cogs to existing metrics, the transition to representing automation impact is relatively easy.

Some Example Scenarios

To help illustrate the real-world application of measuring the impact automation in conjunction with existing metrics. Some resulting impact of automation conversations could include:

Automated routing of incidents has improved LT by 20%.
Automated log collection has improved PT by 30%.
Automated validation of user details has improved package delivery %C&A by 5%.

Final Note on Automation

It's important to note the introduction of automation can produce both positive and negative impacts. You should not assume introducing automation will always provide a positive impact.

By measuring the impact of automation within a value stream, you can enhance the insight provided by the existing metrics (i.e., LT, PT, %C&A).

The following quote from Bill Gates sums up the efficiency impact of automation nicely:

The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.