Books on Strategy, Habit Building Products and Alignment
Scary, funny, informative. That’s a good description of most of the 30 or so books I’ve read in the first quarter of 2024.
So today, I want to share the ones I found to have the biggest impact on my thinking on the data science and AI space. I’ll try to highlight who should read each book because I don’t think there’s any book everyone should read.
Good Strategy/ Bad Strategy by R. Rumelt
This is one of the few dozen or so books I reread every other year or so; it is, in my opinion, the best business book written in this century.
But while it is an amazing business book, it is even more so important for data leaders. I think everyone close to either internal data strategy or (data) products should read this book.
The essence is simple: True strategy is about leverage. If you don’t have a lever, you don’t have a strategy. But if you have a lever, you still need a coherent set of actions to turn the lever into a full-fledged strategy.
This sounds simple, and all good ideas are simple. And yet, if you read through the book, you’ll realize how much of your strategy isn’t actually strategy and how much you truly need a good strategy.
Here’s my recommendation for reading it: Read the Introduction (5 pages), then skim through chapter 3 (bad strategy), then dive deep into chapters 5 (kernel of good strategy) & 6 (using leverage), and finally skim a bit through the rest of the book (part 2 is mainly a list of different types of leverage) - you’ll get most of the information within 30-50 pages of reading.
Storytelling with Data by C. Knaflic
I’ve picked this book up quite some time ago, and skimmed it, and read through most of it again this year, because it’s good! For every tip it has to offer in my mind, I say, “Yes, you’re right, I just messed that one up.” I wrote a longer summary of the book here.
The short version: Every data person should own and skim through this book. Every analyst should deeply study it, work through the sections, and adapt his old stories.
The essence of this book is this: How data is presented to us in 99% of the cases and how you should present it to tell a story are two very different things. This book is about storytelling and about helping people get to the point.
Read Write Own by C. Dixon
This book is a bit out of left-field. It’s only been published in January, and I stumbled over it because I have a fair interest in blockchain. But, this book isn’t really about blockchain. Sure it offers blockchain as a solution to a problem, but most of the book is about platforms and how our internet currently is dominated by closed platforms.
Since platforms are dominated and run on data science and machine learning, and we are all on those platforms, this is as close to a book for everyone as it gets.
The short version: Platforms like Google, YouTube, and Facebook follow an attract extract pattern. In the beginning, everything was open, happy, and free to attract lots of users. Until they have amassed enough to be the monopolistic market leader, that’s when they switch to an extract pattern, wringing us all for money. (Of course, the book argues that this can’t happen with blockchains, but that’s besides the point I want you to take away from this book)
How to read this book: You’ll need to read it face to face, but it’s a dense book, so instead, I’ll redirect you to a long video first.
Chip War by C. Miller
I like to look left and right with my interests; this book, however, is more like below my usual interests in data science. It’s about chips, the foundation of Moore’s law, and as such, about what drives all of the recent prosperity of the world and all of our jobs.
The essence of this book: The chip industry is a world of weird interdependence consisting of multiple bottlenecks. That’s a scary thought! This industry is so interdependent that a single blow to one place will bring down the whole world.
I like to keep track of the trends and forces that surround us, and this is a scary one. I recommend it to every leader in the data space.
How to read it? Read it in full. It’s written by a history professor with a taste for good writing, IMHO.
Hooked - How to Build Habit-Forming Products by Nir Eyal
I picked up this book because I reread a lot of traditional product management literature lately, but this one stood out. To me, data-powered products often lack emotions; they lack any kind of connection and certainly are not habit-forming in any way. So I think this is a great book for any data product manager or founder.
The essence of the book: If your product can form an addiction for the user, you are creating a competitive advantage. That might sound bad, but it’s the reality of TikTok, Facebook, and so many other products. Now this book provides a nice framework for creating such addictive behaviors (which can be good!)
We don’t have a phone addiction; we have a phone-based product addiction.
This book is to be read front-to-end, but you can skim quite a bit; it shouldn’t take more than an hour to pull out tons of ideas for your own product.
The Alignment Problem by B. Christian
While the alignment problem has caught on in popularity only in the past year, this book is actually already four years old. I find this book fascinating because it displays a good recent state of machine learning and the challenges that we have.
The essence of this book for me is that machine learning is harder than it looks. There are more approaches and failures hiding under the carpet than you know of.
I think leaders, PMs, and MLers who work on ML and AI products should read this book. It does teach a bit of humility and respect for the recent technologies we’ve been given.
How to read it? I think it’s fine to skim through this book. It has lots of small stories here and there, and I tend to focus on those.
Weapons of Math Destruction by C. O’Neil
My final recommendation is a controversial book, if not for the content then definitely for the author, one prominent figure of the Occupy movement and quite an outspoken and opinionated woman.
For me, the essence of this book isn’t about the bad that algorithms can cause but rather about the degree of penetration algorithms (be they simple data science or advanced ML) have in our society.
Once you read this book, you’ll be astounded by how much algorithms already control our lives and have done so for decades.
It’s a book for data leaders.
How to read it? Just like the alignment problem, this book has tons of stories, each of which you can skim through to get the point.
FWIW, I recommend all of those books, and I have read them all to the end and taken notes on all of them. I don’t write about books I won’t recommend, and I don’t read books to the end I don’t think are worthwhile.