VSM, DevOps, and Agile

VSM and ITSM

A value stream doesn’t care whether you are a developer or an IT operations person. It doesn’t care if you practice Extreme Programming, scrum, or ITSM. All a value stream cares about is whether you are optimizing flow and understanding your customers’ experience.

Traditional processes can carry a lot of weight, in the form of controls, perhaps bureaucracy, that slow work down to the detriment of customer experience. DevOps is partly about bringing agility to IT Operations and then, by its very definition, to ITSM. This manifests in many ways:

• Non-functional requirements and tests are part of a single product (value stream) backlog

• The work on checklists (e.g. for release or security) is automated for speed and governance

• Change Advisory/Approval Boards (CABs) are replaced by lightweight, peer-reviewed processes

• Environment provisioning is automated and self-service using Infrastructure as Code (IaC) and cloud

• DevOps toolchains provide the foundations and capabilities to automate releases and deployments 

• Incident management is dealt with using intelligent swarming and augmented with AIOps

VSM and SAFe

The leading agile at scale framework, SAFe, uses value streams as a core organizational model. It’s fundamentally built on the concept of value streams and is a flow-based system. Value streams and ARTs are described as “the organizational backbone of a SAFe implementation.” It seeks to create a new organizational model that is primarily concerned with the flow of value to the customer - just like DevOps and Value Stream Management. SAFe identifies two types of value streams: operational and development, the latter being the teams that provide the solutions the business uses to deliver value.

There is a difficulty in DevOps and VSM in treating digital value streams in these two types: DevOps seeks to diminish to the point of exclusion barriers and handoffs between the two disciplines. Defining value streams as operational or development only perpetuates the separation.

Using the two variations of core and supporting is more effective. In a world that has been eaten by software, digital value streams are core. They are the value streams that differentiate against the competition and on which survival is bet. Other value streams exist to support them - whether they are sales and marketing, procurement, finance, or GRC, their role is to ensure the flow and value realization of their associated digital value stream(s) are optimized.

Enterprise Value Stream Management

It’s more commonly asked whether VSM can be applied to small, startup businesses than its relevance to large, traditional enterprises. Enterprises are inherently aware that they are constructed of value streams – the work they need to do is identify, name them, arrange around them and work to their tune, daily. Where enterprises come unstuck, and they are often painfully aware of their situation, is that their value streams exist in a multitude, crossing over one another, connected, coupled, and seemingly unravelable in their complexities.

It is a truism of DevOps that dependencies must be broken, not managed. But reaching a fully autonomous, loosely coupled state is not achievable overnight. It’s a deeply intertwined system of organizational modeling, system architectures, and cultural and technical debt that needs to be steadily chipped away at.

VSM and the DevOps Toolchain

VSM is about people and process and tools. DevOps roared into life at the backend of 2009 and is a ways of working movement driven by demand to balance and improve throughput and stability. It’s about culture as much as it’s about automation but it’s what we’ve learned about building DevOps toolchains that really changes the VSM game.

We started with version control and evolved to continuous integration which allowed us to continually deliver. CICD pipelines became commonplace. But they didn’t handle the end-to-end software development and delivery processes. For that, we needed DevOps toolchains that incorporated the definitions of work in the form of portfolio management and product backlogs. And also the tools that manage software post-release; the service desk tools and the monitoring and observability that provided the feedback to keep driving improvements and enhancements around the cycle. Now, we have data points at every step in the digital value stream.

Challenges remained though:

1. How to get at the data and make it useful

2. How to integrate the heterogeneous tools to be able to observe flow

Enter VSM. Experiments with DevOps dashboards were limited in usefulness since they effectively scraped and presented the data points without offering traceability through the end-to-end toolchain. They also weren’t capable of augmenting human analysis with insights. Data is not useful until an actionable insight has had a learning outcome.

VSM platforms like Plutora do things differently. They:

• Integrate all the tools in the end-to-end software development and delivery lifecycle (the digital value stream) making it possible to trace work items from their inception to realization

• Can do this regardless of the heterogeneous tools in a pipeline and variations across multiple DevOps toolchains providing enterprise-level visibility

• Automate the value stream map and the key data we’d expect like processing, wait and cycle times but also extract insights into delays, waste, and flow acceleration opportunities