My process vision

The Mumbai Dabbawalas deliver roughly 200,000 lunches every day across the city—from homes to offices and back—with an error rate so low that Forbes gave them a Six Sigma rating. They do this without computers, without mobile phones, and with a workforce that is largely illiterate. Their system works not because it is complex, but because it is beautifully simple: a coding scheme painted on each tin, a set of well-understood handoffs, and a culture where everyone knows their role and executes it consistently. Watching them is a lesson in what process is actually for.

Working toward CMMI certification at an ISO-certified firm, I found myself thinking about this constantly. The organizations that struggle with process improvement are rarely failing because they lack the right frameworks. They are failing because they have confused the map with the territory. They have built processes that describe the work in impressive detail without actually helping anyone do it better.

My vision for what good process looks like starts from a different premise. A process should create guidelines, tools, frameworks, and best practices that address real problems the team faces. It should develop models for specific business areas and problem categories—not generic templates applied indiscriminately. It should be treated as a serious organizational project, one that deserves the same rigor and investment as any customer-facing work. And crucially, it should allow freedom of creativity within its structure, not replace that creativity with bureaucratic compliance.

The resistance to new processes is almost always a communication problem. When people understand why a process exists—what problem it solves, what it prevents, what it makes possible—adoption is far more natural. When they experience it as rules handed down from above without explanation, they follow it reluctantly at best and find ways around it at worst. The first question for anyone defining a process should be: can we explain the purpose simply enough that someone joining next week would immediately understand it?

Good processes have clear characteristics. They are simple enough to be understood quickly. They have well-defined entry and exit criteria, so people know when the process applies and when it is complete. They specify their inputs and outputs explicitly. They are technology-independent, because the underlying principles should not change every time the stack does. And they are grounded in the organization’s actual values and principles, not imported wholesale from a certification body’s checklist.

The Dabbawalas are successful not because they have a process manual, but because every person in the network has genuinely internalized how the system works. That internalization is the goal. A process that lives in a document and dies in practice is not a process—it is a formality. The aim is a process that becomes habit, that people reach for naturally because it genuinely helps them do their work. That outcome is harder to achieve than it sounds, and more worth pursuing than any certification.