Self-Organizing Teams

This piece belongs to the ongoing Technical Manager Guide. I’m writing it for leads who suddenly find themselves in charge of other humans and wonder how to scale the mess without turning into Dilbert’s manager.

---

Three-ish years ago our tiny digital agency was just me and another developer arguing about indentation. These days we’re hovering around 50 people spread across design, comms, engineering, QA — the whole buffet. We experienced significant growth over the years, which sounds great in board decks but introduces a never-ending stream of “who decides this?” conversations.

The growth spurt exposed every classic agency pain: clashing priorities, thinly stretched seniors, turnover that kept HR busy, you name it. A product company at least rallies around one backlog; an agency juggles half a dozen roadmaps, each with a different impatient client on the other end. Herding cats feels trivial in comparison.

Our first instinct was the textbook setup: product owner talks to client → writes stories → hands them to the team. The conveyor belt worked on paper, but we kept tripping over bottlenecks, and the “why are we even building this?” questions never reached the people who could answer them.

So we changed the script. Rather than piling more process on top, we put teams in the driver’s seat (well, more like tossed them the keys and promised roadside assistance). No status ladders, no command chains yelling downwards — just small, cross-functional squads owning the outcome. Everyone outside the squad became support staff whose only job is unblocking them.

From Bottlenecks to Team Ownership

Step one was killing the sentence “go ask X.” Every time that phrase surfaced, we mapped the decision back to a role and moved it into the team. (Confession: I still get pinged as X occasionally. Progress, not perfection.)

Responsibility sits with the squad as a whole. If something ships broken, fingers point inward, not upward. That shared burden makes late-night fixes hurt less and stops one unlucky engineer from becoming Atlas.

The other pillar is treating adults like adults. We don’t track 8-hour blocks or police lunch breaks. In a self-selected group that’s fired up about the goal, someone coasting shows up fast — peer pressure is a stronger force than any timesheet.

A typical mindnow squad is seven-ish people mixing design, dev, comms. We can drop them into a new client repo on Monday, and by Wednesday they’re merging PRs. They talk directly to end users whenever possible — a developer scrapped an entire reporting feature after a five-minute call revealed the client only needed a CSV export. That kind of pivot never happens when questions have to crawl up and down a hierarchy.

We rarely do fixed-price projects. Instead we commit to the slice we can deliver in the next sprint and renegotiate afterward. Keeps things honest, keeps the scope small enough that a squad can own it end-to-end.

(Quick aside: removing explicit hierarchy can spawn an invisible one. The loudest voice or most senior tenure still shapes decisions. We surface these dynamics with weekly retrospective cards — anonymous ish — and a rotating facilitator who isn’t the formal lead. Not perfect, but it drags the shadow hierarchy into daylight often enough.)

Scaling this without fragmentation needed extra glue. We run a lightweight architecture guild every Friday — anyone can join, share a design doc, get roasted kindly, and walk away with patterns the other squads already solved. There’s also a tech-sync Slack channel where questions about, say, our shared Terraform modules get answered before people invent a second standard.

Reaping the benefits

The upsides showed up fast. Emotional buy-in skyrocketed once folks realised they could shape the backlog instead of just processing it. There's a commonly cited sentiment among contractors who used to punch Jira tickets that they’ve rediscovered that “builder” feeling.

Management’s life improved too. My calendar shifted from “approve everything” to “curate the right mix of skills and guardrails.” I spend time aligning on goals, untangling crises, and making sure there’s funding for whatever moonshot a squad wants to attempt next quarter. (Some decisions — budgets, overall portfolio, legal exposure — still stay at the exec layer. Total anarchy is fun until the invoice hits.)

Known pitfalls, because nothing’s free: squads can drift in different directions, duplicate work, or simply move slower when consensus stalls. We hit all three. The fix was brutal transparency — shared dashboards on velocity, a monthly “show the mess” demo, and, when needed, a temporary tech lead parachuting in to re-align architecture. Momentum returned within a couple of weeks each time, but those bumps reminded me self-organisation isn’t a set-and-forget toggle.

Breaking the old habits

People fresh out of “here’s your ticket, don’t ask questions” need time. The first couple of sprints are bumpy — decision paralysis, nervous slack pings, the works. Pairing them with long-timers who model the behavior (and occasionally nudge them with a gentle “pick something and iterate”) speeds that up.

Other habits we keep an eye on:

1. No micromanagement. Product owners stand next to, not above, the squad. Their weapons: a clear vision, an unblocked backlog, and direct user feedback.
2. Meetings are the last resort. We default to async threads or a quick Loom recording. Ten people stuffed into Zoom to listen to status updates is a crime against payroll.
3. Mentorship is peer-to-peer. Juniors shadow whoever’s nearest, regardless of org chart. Knowledge spreads faster that way and avoids the accidental hero role.

Next steps

Holacracy keeps popping up in my reading list. I’m cautiously poking at it — mainly to steal the bits that make sense and skip the religious parts. If we do attempt a bigger rollout, I’ll document the bruises.

We’re also standardising our stack so any engineer can jump squads with minimal friction: shared ESLint configs, one flavour of React, one CI template. Monoculture has risks, but context-switch tax is worse for us right now.

Long-term we want everyone sharing in the upside. A lightweight profit-share pool is on the roadmap for 2022 — the Excel sheet exists, which is 80 % of the battle.

If the growth curve keeps behaving, we’ll need even more brains who enjoy this style of organised chaos. Fingers crossed.

---

Thanks for reading. You might also enjoy:

• 🛠️ How to be a better technical manager
• 🚀 Founders Guide 2022
• 📚 Q&A for Startup Founders
• 📰 Previous Newsletter Issues