I hope everyone had a great Thanksgiving! Not that I am very productive in the kitchen, but I spent a good portion of yesterday with flour on my hands. It's a strange type of relaxing to stand around and make food all day. I'm thankful that I'm in a situation where slightly burning the herbs on top of our turkey was the lowlight of my day.
A number of people, who used to work for me, have slowly moved into larger leadership positions. I've had some reach out recently, talking about the challenges they have as their group grows. I remember one product manager complaining years ago that his startup had been “much more productive” than teams at Amazon. Now that he has teams of his own, he admits that it's significantly harder to be productive at scale than he had realized. It's not that big companies are inefficient because they're lazy or missing some magical startup energy. It's just that it's easier to be efficient when you work on smaller systems with fewer people.
I love sharing what I and others have learned over the years through trial and error. I'm glad you're on this journey with me! If you'd like to dive deeper, consider my private hourly coaching or email coaching programs. Have a great rest of your holidays!
When working for Amazon, I spent a good amount of time thinking about large vs small companies. Amazon could throw a billion dollars at a problem if it wanted to. In many situations, small companies couldn't hope to compete if Amazon was interested in beating them.
Yet I was also aware that in any particular area, a little startup with a small advantage could easily surprise everyone. The Innovator's Dilemma (great book by the way, associates link — because why not) covers this topic in detail.
Amazon is a collection of many startups under one company umbrella.
Some of these internal “startups” are large, like the AWS EC2 service. Some are little 10-person groups, in hopes of becoming a large part of Amazon's business someday.
Efficiency vs Size
As groups grow within Amazon, they become less efficient (per headcount) the larger they get.
Imagine you have a group of 20 people, launching features. As you grow to 40 and then 80 people, you won't see a corresponding 2x productivity increase. While you can't compare the complexity and cost of building a feature in a vacuum (they're never apples to apples comparisons), you can tell that groups get slower as they get larger.
This matches what you see with external startups as they grow as well. You'll see a fantastic little 10 person startup, throwing feature after feature into their product. As they grow their number of employees, their launch pace does not increase to match their personnel growth. Often you'll see a decrease in new features as time goes on.
There are many reasons that efficiency decreases.
Operational costs — Your servers need replacing, your customers need a refund, and there's a strange bug you just can't find. Running a live system costs time and money. As you build new things, you constantly add to your costs.
Production systems are harder to change as they get customers — If you have 7 customers, it's simple to make major changes. If you have 7,000 customers, it's harder. If you have 7 million customers, it becomes exceedingly difficult to do anything. I remember an anecdote about testing some exceedingly minor color change on Amazon's navigation would impact revenue by millions of dollars. This meant that even the tiniest change requires extreme due diligence. As companies become successful, simply having customers slows them down.