When you mention metrics-driven coaching to an agile coach, you will often get an initial reaction that “Metrics are evil.” Often, a valid reason underlies this reaction. Many of these coaches have had experiences where metrics were used to evaluate, manipulate and control individuals and teams, rather than for self-managed growth and improvement programs.
Between “metrics are evil” and a reliance on anecdotal coaching lies a pragmatic and effective value proposition for metrics-driven coaching – an organization undertaking continuous performance improvement in a deliberate and demonstrable, yet human-centered way.
In this post, I will describe the underpinnings of such an improvement process.
What is a Performance Chain?
There is a field of practice focused on organization performance improvement1. One of the leaders in this field, Six Boxes, has codified an approach called the performance chain, that is useful for our discussion.
Figure 1 – Six Boxes Performance Chain
According to Six Boxes:
Human performance is behavior producing work outputs that are valuable because they contribute to business results. Work outputs are what the organization needs from its people to achieve its goals. They are the links between the behavior or activity of people and results.
By clarifying the relationship between activity and results, we align people with the mission and goals of the organization. For executives, managers, HR, training, and process improvement – anyone responsible for improving or changing performance — clarifying the Performance Chain increases the likelihood that their efforts will generate the highest returns on their investments.
Performance Chains in an Agile World
In the agile world, we have developed an understanding of the relationship between behavior and outputs. Business agility is an emerging frontier in the evolution of agile organizations, and tying the work performed by people to business results (outcomes) is a central theme of business agility. The chain helps us traverse the alignment of behavior to outputs to outcomes. One of the most common tools used to codify outcomes is called Outcomes & Key Results (OKRs). The addition of OKRs completes the means to define and assess the agile performance chain.
Inspired by Six Boxes performance chain, I have developed a visualization of the performance chain with agile behaviors, outputs, and outcomes. Note the outcomes are based on business results, not the agile approach.
Figure 2 – Agile Performance Chain
An agile coach is familiar with coaching teams and programs on behavior, the ceremonies and processes that agile development teams employ – the how. In addition, those coaches have models for the outputs produced by those behaviors; e.g. stories, features, capabilities, MVPs, specifications, test plans, defects – the what. One of the correlations between outputs and behaviors is to work with small increments of output enables an iterative approach to the behaviors enabling the adaptive characteristics of fast feedback and pivoting. Finally, outcomes are the whys that drive the behavior. We will talk about the correlation of outcome with behavior and outputs in the section below on the Software Development Productivity Index.
Metrics and the Agile Performance Chain
Over the past 10 years, adoption of metrics-driven coaching has been accelerating in the agile world. Perhaps the most significant exemplar of this trend is AgilityHealth’s measurement and growth platform, including the full range of Agility Health Radar assessment instruments. Note the categories of metrics illustrated in the following figure from AgilityHealth’s website:
Figure 3 Agility Health Metric Categories
The categories that AgilityHealth uses align with behavior, output, and outcomes in the Agile Performance Chain.
Figure 4 – Measuring the Agile Performance Chain
The Software Development Performance Index
In 2012, Larry Maccherone joined Rally Software to develop the Software Development Performance Index (SDPI)2. Rally is a cloud-based agile lifecycle management platform with a wealth of agile project data. Larry and his team anonymized and analyzed the data to uncover the relationships between behavior, outputs, and the outcomes of responsiveness, quality, predictability, and productivity. These data-driven correlations are intriguing and not always intuitive; for example, the Rally team discovered that mature teams did not task out stories yet still maintained a high level of predictability. The interpretation was that mature teams did not need the added information generated by tasking to know when their sprint backlog had the right number of story points to be delivered predictably.
The Rally team discovered that there are tradeoffs between performance measures that demand mindful optimization; for example, maximizing responsiveness can potentially sub-optimize quality of the product. A team needs to find the balance in the scorecard of outcomes and how to achieve that balance.
Figure 5 – SDPI Balanced Score Card of Outcomes3
The same data-rich environment that attracted Larry to Rally exists within AgilityHealth’s cloud-based measurement platform. By collecting data on the behaviors, outputs, and outcomes of teams across an organization, AgilityHealth helps uncover correlations between those behavior, outputs, and outcomes. Accenture | SolutionsIQ is leveraging AgilityHealth’s findings to inform large enterprises’ agile enablement journeys and transformations, helping them accelerate and maximize strategic business impacts from agility.
Figure 6 – Correlating the Agile Performance Chain
Continuous Improvement and the Measured Agile Performance Chain
Equipped with the insights gleaned from the metrics and their correlations, an organization undertaking an agile transformation can evolve their coaching approach from an ad-hoc and anecdotal approach to a more data-driven approach that can be evaluated and refined for greater effectiveness and impact. Agile coaching organizations and transformation leaders should possess data and data analytics skills to accelerate and maximize the benefits they deliver to their organizations.
On the other hand, coaching organizations do not need to wait until these correlations are encoded in analytic algorithms. Agile coaches can employ simple visual benchmarking to compare a team’s performance with other teams in the same organization, to stimulate conversations and improvement plans.
Figure 7 – AgilityHealth TeamHealth Radar.
These human-enacted analytics allows the coaching organization to draw the correlations between behaviors, outputs, and outcomes. For example, a team might ask, “What behaviors are other teams doing to achieve their predictability score that we might adopt to improve ours?” It is a question that a coach might bring to a coaching community of practice.
One of the foundational tenets of a learning organization is continuous improvement. The expectation is that the periodic assessments result in iterative, incremental growth plans developed to move the needle over the next product development period (sprint, program increment, quarter). The combination of metrics and growth plans helps close the loop of a continuous performance improvement cycle.
Figure 8 – Growth Plans and Metrics-Driven Coaching
I hope agile coaches will realize that metrics-driven coaching is not evil and that measured agile performance chains can be another powerful tool in their toolkit; a tool that can help their clients understand the way they work and continuously grow in their adoption of agile practices.