An Elegant Puzzle deep dives into pretty much every aspect of being a tech leader at modern software companies. The author, Will Larson, has worked at Yahoo and Digg, moved on to engineering manager at Uber, then joined Stripe, and is now CTO at Calm. Quite the CV, and it shows.

Impressions

I’ve nicknamed the book “The Manual”. I keep coming back to it whenever I need guidance or just a reminder that the struggles of leading teams are universal. If you’re even slightly tempted, buy the physical book and not the Kindle version. Much easier to highlight, reference, and hand around when you want to convince someone else on your team to take a look.

One of my favourite aspects is how grounded the book is in reality. Larson’s been at the coal face for years. Whether it’s managing tech debt, negotiating hiring headcount, or knowing when to say no, you get the sense these are lessons learned by running into the wall himself, not just observing from afar.

A few standouts for me:

  • What to do when your systems are slowing you down, but migrations are stacked up higher than you can count.
  • How to say no, when there’s endless work and not enough people.
  • Good, actionable ways of paying down tech debt (not just hand-wavy “prioritise it” advice).

Structure

The book is split into six chapters, all independent, and you can dip in wherever you like:

  1. Organisation: Team structure, org design, and the subtle ways design decisions play out over time.
  2. Tools: Systems thinking, and practical tools to manage change at org, team, and individual level.
  3. Approaches: Challenging situations engineering managers run into, with options on how to navigate them.
  4. Culture: Small, persistent actions that shift culture (or entrench it).
  5. Careers: The messy world of hiring, performance, promotions, and retention.
  6. Appendix: Useful templates, checklists, and reading lists and real tools you’ll actually use.

Highlights

If you’ve just started a new job, create a 90-day plan what you want to learn, ship, and understand in the first three months. Larson lays out why this matters and how you can use it to spot gaps early or steer yourself back if you drift. (If you’ve read A Manager’s Path, this will sound familiar, but Larson brings more concrete guidance.)

Another bit I keep returning to is the simple framework for presenting to senior management (Chapter 3). It’s not just about “sounding smart”, it’s about building trust, showing that you understand the trade-offs, and making it easy for others to support your decisions. I’ll probably do a deeper dive on this alone.

For engineers, build a real relationship with your manager. Larson lists the things your manager should know about you:

  • What problems you’re trying to solve
  • What you like to work on
  • How you believe you’re being measured
  • How busy you are
  • Your goals and growth areas

That list alone is worth checking in with your manager next 1:1. If they can’t answer those about you, odds are you’re not getting what you need, and neither are they.

One idea I’ve stolen for my own toolkit is Larson’s “four states of an engineering team”:

  • Falling behind: Backlog growing weekly, always behind.
  • Treading water: Critical work gets done, but no space for tech debt.
  • Repaying debt: Starting to pay it back, making room for progress.
  • Innovating: Low tech debt, high morale, new features flying out.

The beauty is that Larson doesn’t just name these, he maps out steps to get unstuck from each state.

What makes An Elegant Puzzle different from most leadership books is its honesty about trade-offs. It’s not about chasing perfection, or pretending every problem is solvable if you just “empower your team”. It’s about wrestling with constraints, making tough calls, and picking your battles.

Larson reminds you that organisational design is never finished. Teams, structure, and incentives quietly shape every outcome. You can’t out-process a bad structure, and you can’t out-hire a broken culture. The work is ongoing.

He’s also refreshingly pragmatic about topics like tech debt. Not all debt is created equal, and sometimes the real work is surfacing which kind matters, and which you can safely ignore. There’s a discipline to naming the pain, not just labelling everything as “debt”.

The book is packed with frameworks and checklists, but the real value is in the stories and context. You’ll find yourself coming back to it, especially after a rough week. It’s the kind of book that gets better the more real-world experience you bring to it.

If you’re leading engineers (or just want to), An Elegant Puzzle is the book I keep reaching for when things get complicated. It won’t solve your problems for you, but it will help you see them more clearly, and that’s sometimes all you need to find a way through. You don’t need to agree with everything, but you’ll find yourself referring back to it when you hit a wall.


Published on .