The Polaris Flow Dispatch
The Polaris Advisor Program Blog
Hello, I’m Dr. Krishna Kumar, Founder and CTO of The Polaris Advisor Program: a technical advisory firm helping software companies build rigorously crafted measurement systems to manage their product development and engineering operations.
I’m a software engineer and entrepreneur with a PhD in computer science, and before starting my technical advisory practice in 2011, I worked at all levels the software industry in technical and managerial roles for 20 years, after a brief teaching stint in academia.
My practice is focused on developing rigorous measurement techniques to improve the efficacy of measurement tools in software operations.
Our philosophy
Our solutions start from proven mathematical foundations and concepts vetted over several decades in operations research, and we adapt them for applications to the specific problems we face in software operations management.
This is an approach that has proven very fruitful in other sectors such as manufacturing and supply chain management and we aim to bring similar rigor and discipline to operational management in the software industry,
To begin, this requires facing the reality that processes in software development operate within a complex adaptive system. In crafting measurement and management solutions for this domain you can’t copy ideas from manufacturing wholesale and treat them as though they were production processes.
You can adapt some of them if you pay careful attention how you interpret and adapt the underlying mathematics. Many other problems require wholesale rethinking of fundamental concepts of measurement and bringing in new abstractions.
What I’ve discovered though, is that much of the underlying theory of stochastic processes can still be applied to real-world measurement problems in complex adaptive systems, and yield solutions that are significantly more robust than the current state-of-the-art of operational measurement.
This is true even when we accept that what we measure is driven by the collaborations between people doing deep knowledge work. My research is about making those adaptations precise starting from proven mathematical foundations.
The Presence Calculus Project is our flagship research effort. After nearly seven years of research and development, building on several prior years of exploration, implementation and experimentation with a now-retired commercial product (Polaris), we now have a defensible theory and viable new approaches to process measurement and tooling that start with provable fundamentals and are rich enough to tackle the unique challenges of the software domain.
This substack is focused on jargon-free exposition of the key ideas behind our research projects with the aim of making them accessible to a wider audience of practitioners.
The practical tools we develop in our program are open-core or open-source, and we believe the knowledge required to build them should also be free. So, we share the theory freely, and the applications and the open-source implementations can be found on this substack and on GitHub.
Our motivating thesis that the dominant theory-free, vendor and influencer driven, approach the commercial software industry takes to operational measurement is deeply broken and fixing it requires a fundamentally different approach to both theory and tools.
If, like me, the conventional wisdom and theory-free debates around “metrics” in the software industry has left you frustrated, subscribe to this substack, follow our research on GitHub, or follow me on LinkedIn and see if what I am saying strikes a chord.
The Polaris Advisor Program
If our ideas resonate, and you want to work with us to help apply these ideas in your company, reach out and we can explore whether there’s mutual value in collaborating.
The Polaris Advisor Program offers something unique: context-aware, deeply researched and cutting-edge measurement products, paired with a consultative process that helps you design the bespoke systems you need to measure what matters in your company, to achieve the outcomes you want.
We can help you build these systems.
What we talk about here
Everything related to operations management in software companies and how we can use data and measurement to rigorously improve how we can create and deliver the most valuable software, even in the messiest, most complex environments we all live and work in.
Our theoretical focus include stochastic process theory, queueing theory, value networks, complexity science and a whole lot more. Application areas include DevOps, Value Stream Management, Lean continuous improvement, product and engineering metrics and more.
If it is relevant to understanding how to measure and improve work and value creation in software development, it’s fair game.
Software development, as an industry, remains notoriously immature when it comes to using data to improve the way we work. We much prefer to follow frameworks, practices, and anecdotal evidence to do this. Most industry leading approaches are completely unmoored from any defensible theories.
We prefer analog techniques to digital tools and the comfort of imprecision and folklore to the hard work of collecting and analyzing high-quality data and using rigorous theories to make decisions.
Many aspects of software development are hard to measure, but we argue that’s because we have not built the right theoretical foundation to do so.
This is what we will dive into each week here in this publication.
Why subscribe?
We will bring theory and practice together and explore what the state of the art in this space looks like and what it will take to drag our industry kicking and screaming into the 21st century.
I hope you’ll join me in this exploration.
Subscribe to get full access to the newsletter and website. It’s free, and you’ll never miss an update.
