This is an episode transcript. Visit the podcast website for the full episode, show notes, and freebies.


Well hello again. Adam here back for the next episode of Small Batches. I’m going to switch gears away from the 12.1 factor app for a while. Rest assured, I will return to this thread a vengeance.

Also, I’ve recorded two different hour long interview episodes with Jason Swett on his podcast Rails with Jason. We talked about DevOps, Continuous Delivery, testing, and more. I’ll publish them as bonus episodes when they’re out. I think you’ll enjoy them.

Alright, today we’re covering the four types of work. Here we go.


Every team needs a taxonomy to categorize the work they do. Taxonomies make it easier to reason abstractly about work and its role in the business.

There are two types of work on the surface: features and non-features. Features require no further explanation. Non-features is a bigger group. Bugs, or defects, fall into this category. Defects and features do not account for all the other work software teams do though.

Features and defects are easy to grok because they correlate directly with business value. Features attract customers. Defects are broken features or quality regressions. Teams do more than developing features and squashing bugs. Engineers spike on new problems. Developers refactor code and pay down technical debt. There’s the routine maintenance of library updates and security features. These don’t fit squarely into features or bugs, so how are they classified?

These examples demonstrate the remaining two types: debts and risks. Activities like refactoring or updating libraries are debts because they improve the system. This work focus on the team’s ability to sustainability deliver the other three types of work. Risks address security, privacy, or compliance exposures.

The four types of work create a zero-sum model for choosing which support business objectives. If a team devotes one hundred percent capacity to features, then there is no capacity to fix bugs, pay down tech debt, or derisk the business. A start-up with fixed runway may choose to do this. Subsequently, they may devote more capacity to debts and bugs once they clear the runway. On the other hand, a business with strong regulatory requirements, like banking or health care, may allocate the majority of their capacity to risks since that's a core business priority.

The point is that a business' distribution of features, defects, risks, and debts will vary over time as business goals change. The four types of work enable different members of the organization to understand the business value of each and relative prioritization trade-offs.

Now, let's step back.

Consider the three ways of DevOps: flow, feedback, and learning. DevOps applies the three ways to the engineering value stream. Ultimately, one of these four types of work will land in an engineer's backlog. That's where DevOps kicks in. If we think more abstractly then we can break free of the engineering centric perspective.

The principle flow states that work should move from development to production as quickly as possible. Then the principle of feedback says that information about production should drive future development. The four types of work correlates business telemetry. Features provide value. Fixing bugs improve quality. Each of those makes customers happy. Paying down debts keeps costs down.  Mitigating risks keeps other factors stable. So, work—features, defects, debts, and risks—flow left to right across the value stream resulting in changes to value, quality, cost, and happiness.

This is underpinning of the Flow Framework. Mik Kersten introduced the Flow Framework in his book "Project to Product" published ITRevolution press. That’s the same publisher behind The DevOps Handbook, Accelerate, and The Phoenix Project to name a few.

Here’s the tip of the iceberg.

The Flow Framework applies the DevOps principles to the entire business. The Flow Framework models the business as networked value stream. The value streams turns "Flow Items" represented by the four types of work into business value. Business can track their value stream network with four metrics:

  1. Flow Velocity measures how many items were completed over a period of time
  2. Flow time measures how long it takes go form "work start" to "work complete" including wait times
  3. Flow efficiency is ratio of active vs wait time
  4. Flow load measures the number of items in progress on a current value stream

There’s more to Flow Framework than I can cover in a single episode of this podcast. I think this is a good jumping off point for your own research. I’ve already written a dedicated guide to Project to Product which you can get for free. Visit smallbatches.fm/9 for links to my free email course, an infographic on the flow framework (trust me it makes much more sense visually),  and other great material from Mik Kersten.

That’s all for this one folks. Thanks for listening and happy shipping.