Team Topologies
Skelton and Pais's case that most team structures are drawn to fit an org chart, not the actual flow of work, and a small set of team shapes and interaction rules for fixing that, built around a resource every leader habitually overspends: people's cognitive load.
Find this on the mind-map
Introduction
Skelton and Pais's starting claim is unsettling for anyone who's reorganized a team by instinct: most team structures aren't actually designed, they're inherited, shaped by history, by whoever asked for headcount last, or by an org chart drawn for reporting lines rather than for how work actually needs to flow. Their model gives leaders a deliberate vocabulary for team shapes and the rules for how those teams should talk to each other, built around the idea that a team's effectiveness is capped by how much any one group of people can actually hold in their heads at once.
Cognitive Load as the Real Constraint
Before naming team types, the book insists on a blunt diagnostic: how much does this team actually need to know and hold in working memory to do its job well? A team responsible for too many unrelated systems, tools, or domains will underperform regardless of talent, because the limiting resource isn't skill, it's attention. Most reorganizations that fail do so because they moved boxes on a chart without ever asking whether the resulting team's cognitive load had actually gone down.
Stream-Aligned Teams
The default, workhorse team type: organized around a continuous flow of work tied to a single business domain or user journey, able to deliver value end-to-end without needing to hand off to another team for most of what they do. Most teams in a healthy organization should be this type. The other three exist specifically to support stream-aligned teams, not to compete with them for primacy.
Platform Teams
Teams that build and run an internal platform, infrastructure, tooling, shared services, specifically to reduce the cognitive load stream-aligned teams would otherwise carry themselves. A platform team's product is really a service other teams consume with minimal friction; a platform that requires constant hand-holding to use has failed at its actual job, regardless of how technically impressive it is.
Enabling Teams
Teams of specialists whose job is to help other teams close a capability gap, temporarily and deliberately, not as a permanent dependency. An enabling team that's still doing the same hand-holding for the same team a year later has drifted from enabling into doing the work for them without saying so, which defeats the purpose.
Manager A's platform specialists get pulled into every stream team's sprint to fix problems directly, becoming an invisible bottleneck nobody budgeted for. Manager B sets up the same specialists as an enabling team with an explicit end date for each engagement, focused on transferring the skill, not just fixing the immediate problem. B's stream teams get a lot more self-sufficient over time; A's stay dependent indefinitely.
Complicated-Subsystem Teams
A narrower, more specialized team type for parts of a system that genuinely require deep, rare expertise, the kind that would overload a typical stream-aligned team's cognitive capacity if they tried to own it themselves. This type should be used sparingly; naming a team "complicated-subsystem" is sometimes just a polite way of describing an organizational silo nobody's gotten around to fixing.
Interaction Modes
Three deliberate ways teams work with each other: collaboration (close, hands-on joint work, usually temporary, for discovering something new); X-as-a-Service (a clean, well-defined handoff with minimal ongoing communication, for stable, well-understood work); and facilitating (one team actively helping another improve, similar in spirit to an enabling team's relationship). The authors are clear that mismatching the mode to the situation is a common failure. Defaulting to heavy collaboration for stable, well-understood work wastes everyone's time, and defaulting to X-as-a-Service for something genuinely novel starves it of the hands-on thinking it needs.
Conway's Law and the Inverse Conway Maneuver
The organizing principle behind the whole book: systems end up mirroring the communication structure of the organizations that build them, whether anyone intended that or not. The practical implication, the "Inverse Conway Maneuver," is that you can deliberately design team boundaries to produce the system architecture you actually want, instead of discovering after the fact that your team structure dictated an architecture nobody chose on purpose.
Putting It Together
Cognitive load is the resource being protected; the four team types are the shapes that protect it; the interaction modes are the rules for how those shapes talk to each other without just recreating the overload at the boundaries; and Conway's Law is the reminder that all of this shows up in the system itself, whether you design for it or not. The practical message for any leader, inside or outside software, is the same: your team structure is already an architecture decision. The only choice is whether you're making it on purpose.

Why It Matters for Leaders
Most reorgs fail for a boring reason: nobody checked whether the new team could actually hold its new job in their heads. The chart looks cleaner, the reporting lines make more sense in a slide, and six months later the same team is still missing deadlines, just with a new name on the org chart.
The habit this book teaches is asking "what do they have to know to do this well" before you add anything to a team's plate. If a team is already stretched across three systems and you hand them a fourth, the question isn't whether they're capable, it's whether you took anything away to make room. If the answer is no, you didn't lighten anyone's load, you just told yourself you did.
The same logic applies to any group of specialists you send in to help. If the same engineers are rescuing the same team's sprint a year after they were deployed to "help them get self-sufficient," that's not help anymore, that's a dependency with a nicer name. Put a real end date on the engagement and check whether the other team can actually stand on its own by then.
It also reframes how you read your own org chart. The structure you have isn't neutral, it's already producing the system (or the process, or the handoffs) you're currently living with, whether you designed it on purpose or not. If you don't like the architecture, the team boundaries are usually the place to look first, not the technology.
Picture a manager named Noah who keeps hearing that releases are slow because "the platform team is a bottleneck." Instead of adding more platform engineers, he asks the stream teams directly what using the platform actually requires of them day to day. It turns out every release needs a platform engineer to manually approve a config change, because the self-service tooling was never finished. Noah reprioritizes finishing that tooling instead of headcount, and releases speed up without adding a single person.
What to do next
Reinforce your understanding of Team Topologies
Books related to Team Topologies
Coaching for Performance
Sir John Whitmore's foundational case for coaching-style management over command-and-control, built on a single idea: performance is limited less by skill than by interference, and a manager's job is to remove the interference, not supply the answer.
Elevate
Robert Glazer's case, from his book "Elevate: Push Beyond Your Limits and Unlock Success in Yourself and Others," that performance isn't a willpower problem. It's a capacity problem across four specific domains, built like a muscle, not willed into existence overnight.
First, Break All The Rules
A large-scale look at what the best managers actually do differently, and the uncomfortable finding that most of it runs against conventional management wisdom, starting with the idea that you shouldn't try to fix people's weaknesses.
Multipliers
Liz Wiseman's case that the biggest lever on a team's output isn't headcount or talent density, it's whether the leader in the room makes people smarter or slowly makes them smaller.
The Coaching Habit
Michael Bungay Stanier's pitch that most managers have an "advice monster" problem, and a short list of deceptively simple questions designed to starve it, a little at a time, in conversations that already happen every day.
The First-Time Manager
The unglamorous, practical handbook for the specific week your job title changed and nothing else came with instructions. Less a framework, more a set of answers to the questions new managers are usually too embarrassed to ask out loud.
The Five Dysfunctions of a Team
Patrick Lencioni's pyramid of the five behaviors that sink teams without much notice, stacked so that each one is actually caused by the dysfunction beneath it, which means fixing results or accountability first almost never works.
The Making of a Manager
Julie Zhuo's field notes from becoming a manager in her twenties with no idea what the job actually was. A clear-eyed answer to the question most new managers are too embarrassed to ask out loud: what am I actually supposed to be doing all day?
The Manager's Path
Camille Fournier's map of the actual career ladder from doing the work to leading an organization of leaders, written for engineering managers, but mostly a clear-eyed account of how the job itself changes at every rung, which almost nobody explains to you before you climb it.
Turn the Ship Around!
A submarine captain's account of giving up the one thing every leader is taught to protect, being the person who gives the orders, and turning the worst-performing submarine in the fleet into one of the best by making every sailor think like a captain instead.
Skills related to Team Topologies
Frameworks related to Team Topologies
Dunning-Kruger Effect
Knowing little makes you feel unstoppable; knowing more makes you hesitate — the loudest voice is rarely the most reliable one.
Goldilocks Principle
Support runs on a dial between micromanagement and abdication — enough direction for clarity, enough space for ownership, always moving.
Hawthorne Effect
People change how they behave when they know they're watched — your attention is an intervention that changes what you see.
Maslow's Hierarchy of Needs
Abraham Maslow's five-level model of motivation, applied to leadership as a stack — mastery, trust, belonging, up through recognition and vision.
McKinsey 9-Box Model
A 3x3 grid that plots each person on performance against potential, giving nine distinct calls instead of one blurry verdict.
Multipliers and Diminishers
Liz Wiseman's split between leaders who bring out best thinking in people around them, and those who shut it down.
Situational Leadership
There's no right way to lead — only the right way for this person right now: direct, coach, support, or delegate.
The 6 Human Needs
Every behavior at work is really an unmet need in disguise — six needs, popularized by Tony Robbins, worth checking first.
The Architecture of High-Performing Teams
Six elements, built like a temple — culture as the base, four pillars, vision as the roof — name the missing piece.
The Bannister Effect
One person breaking a barrier makes it stop being fixed for everyone else — break the unspoken belief in public first.
The Five Dysfunctions of a Team
Lencioni's pyramid of five team-killing behaviors — a lack of trust at the base drags everything above it down.
The Inner Drive
A leadership take on Daniel Pink's *Drive* that turns three conditions for motivation — autonomy, mastery, purpose — into habits for 1:1s.
The Pygmalion Effect
What you expect from someone leaks into how you treat them, which shapes what they become — for better or worse.
The Thomas-Kilmann (TKI) Model
Five conflict styles, two dials — how hard you push, how much room you give — chosen on purpose, not by default.
Tuckman's Stages of Team Development
Every team moves through five predictable stages — forming, storming, norming, performing, adjourning — each demanding a different move from you.