Making Work Visible Book Review and Q&A with Dominica DeGrandis

Dominica DeGrandis book “Making Work Visible: Exposing Time Theft to Optimize Work & Flow” takes the team as the starting point to improve workflow - both within a team, and between teams (identifying and visualizing dependencies). 

Time waste - especially in large organizations - can take the form of unknown dependencies, unplanned or neglected work, conflicting priorities and, above all, too much work-in-progress (WIP). DeGrandis uses the metaphor of “time thieves” to illustrate these root causes for delayed or low quality work, slow pace of delivery and, generally speaking, inefficient flow of changes and outcomes. The core tenet of the book is that visibility over these time thieves and their impact is a critical first step for reducing waste.

With visibility comes the ability to measure and identify the most damaging “thieves” at a given point in time. Flow metrics help monitor if the health of the workflow is improving as we decide and apply experiments to reduce waste. That is the focus of the last part of the book, which also includes a description of “beastly practices” - common practices such as Gantt charts or individually named work allocation (swimlanes in Kanban parlance) that can take a heavy toll on company-wide performance in favor of wrongly perceived benefits (for example, a manager can see at a glance what each individual is working on).

Although the book is strongly based on DeGrandis experience teaching Kanban, it goes beyond the specific framework practices by focusing on recurring symptoms and problems that teams face but are not always able to trace back to an origin - or they do not have the time and leeway required to address them. Most of the chapters include practical team exercises taken directly from DeGrandis workshops, allowing for immediate application of the improvement ideas presented in the book.

InfoQ interviewed DeGrandis to get some insights on her experience and common problems she find working with different organizations.

InfoQ: How did you come up with the "time thieves" metaphor?

Dominica DeGrandis: People grumble that there just isn’t enough hours in the day to get work done. But we do manage to get a lot of work done when we’re not interrupted. Interruptions tend to arrive from unplanned work and conflicting priorities and unknown dependencies - all issues that stem from too-much-WIP (work-in-progress). These troublesome (yet common) problems steal time away from us day after day after day. If we can protect our hours from being stolen, then we can get more stuff done - hence the time thieves metaphor.

InfoQ: In your experience, which of the time thieves are more prevalent?

DeGrandis: Depends on the context, but more often than not, thief too-much-WIP does more damage than all the other thieves. That’s why thief too-much-WIP is considered the ringleader. 

There is a relationship between the amount of WIP and how long things take to complete. 

The biggest deterrent to fast FLOW (smooth and predictable value pulled through a value stream) is too much WIP - because high WIP means that other work sits waiting for service longer. The more WIP in the pipeline, the longer things take.

One reason people take on too much WIP is because of thief conflicting priorities - when the top priority is unclear, we tend to say yes more to other work.

InfoQ: If you had to pick the top 3 factors which feed into those thieves, what would they be?

DeGrandis: Factor #1 - How people are measured. What we measure impacts people because people value what is measured. Errors or unplanned work may not show up in the data if they don’t feel safe to make their work visible. For example, quality metrics can lack integrity if people get burned by revealing problems. The validity of metrics becomes an issue if work is invisible. It’s hard to manage invisible work. 

Factor #2 - the practice of loading people to 100% capacity utilization. The higher the utilization, the longer the wait - especially in areas with high variability (ex: lots of unplanned work) like IT. The single most important factor that affects queue size is capacity utilization. Experts who are in high demand are often unavailable when you need them. When there is zero slack in the system (because people are allocated at 100%), thief unplanned work sneaks in and delays planned work.

Factor #3 - Dependencies and the practice of keeping teams small. Problems occur when dependencies span across different teams causing integrations to break. Small teams can optimize their own workflow, but it may be at the detriment to company wide performance. Cross team communication is hard. Every dependency increases the probability that you will be late. 

InfoQ: Changing people's ways of working and cultural norms is hard, but necessary for reducing waste. How do you go about introducing those changes in teams or organizations reluctant to change?

DeGrandis: I start by asking teams what prevents them from getting work done? What causes team and/or organizational pain? Once constraints are identified, then I help teams expose the pain by making pain points visible. Making work visible provokes the necessary conversations for change. When people realize that the goal is to surface problems so pain points will be addressed, they usually jump on board. 

Good visuals are key to showing the problems. Good visuals can open the door to new understanding. Diagrams are less threatening for people to view and helps everyone see and understand and relate better – “Oh, I see where you’re coming from….” 

InfoQ: Does the way teams are organized also play a role in the effectiveness of those teams? In other words, are there occasions where the negative impact of dependencies between teams cannot be overcome without re-structuring teams?

DeGrandis: Organization plays a big role in performance - especially in large organizations. At a certain point, large orgs grow to a point where it’s impractical for lots and lots of people to work on lots of different projects - how can team members be aware of every decision (such as architecture changes) that might impact them? Thief unknown dependencies loves tightly coupled architecture.

This is why effective orgs are moving toward organizing by product instead of project. It reduces the problems associated with projects and allows the people who developed, tested, and delivered the functionality to stay in their area of expertise. There is no need for complex, dependency driven handovers. Instead, organizing teams by product decreases dependencies during handoff to ongoing operational support.

InfoQ: The team exercises in the book recommend paying close attention to the visual aspect of the work management artifacts. Why is that so important, and why aren't annotated post-its enough?

DeGrandis: We are visual learners. This is the power of visualizing work – the eye is drawn in automatically to try and figure out what colors, shapes & patterns mean. It’s like a visual language that is easily understood. We need very little education to get the message that demand is much bigger than capacity when we can see a million incoming requests enter the system. 

Think about when you visit a badly designed website and how little you trust the information on the website. Information and visualization go hand in hand. Annotated post-its might be good enough for some teams - depends on the context. More and more, leadership wants to see actual data in dashboards to provide them with the necessary knowledge and trends. It’s helpful and useful to show information in ways that people enjoy and understand. The combination of pictures and text responds to our need for a unified agile language.

InfoQ: In general, how does the book fit in today's DevOps trend where organizations try to accelerate their delivery while increasing quality and auditability? Is there a risk that optimizing "locally" inside teams does not actually move the needle for the organization as a whole if there are other bottlenecks/challenges upstream or in the overall culture?

DeGrandis: Local optimization can occur to the detriment of company-wide performance, making it hard to meet CIO objectives. It could be that the most valuable thing for the company as a whole is for John from team A to go help Brian from Team B finish something. The organization can only move as fast as the slowest moving part of the system or value stream. This is where the concept of systems thinking is so vital to DevOps.

InfoQ: Queueing theory tells us that wait times due to work being queued up negatively impacts lead/cycle time more than the actual execution time, thus the importance of reducing work in progress and focusing on finishing on-going work rather than starting new work. How do you explain this in a way that both the business and tech side of the organization can relate to?

DeGrandis: In several different ways. Most people can relate to a freeway loaded at 100% capacity utilization that clogs up because cars slow way down. Everyone knows that the minute they get on a crowded freeway, their commute home will take longer. 

Same thing at the grocery store - when there are fewer checkout stands, the lines lengthen and it takes you longer to flow through the checkout line.

InfoQ: Should we look at improving our workflow as a continuous learning process? 

DeGrandis: Absolutely. Continual learning is a must in our complicated and complex work domains where best practices don’t yet exist. If we don’t know cause and effect, then experimentation is the way to go. 

About the Book Author

Dominica DeGrandis, author of "Making Work Visible: Exposing Time Theft to Optimize Work & Flow", teaches technology and business organizations how to design systems that optimize workflow across value stream networks. As Director of Digital Transformation at Tasktop, she helps teams reveal mutually critical information using Lean, Kanban & Flow methods. Dominica blogs at ddegrandis.com and tweets at @dominicad.