Interesting line of thought. 37:35 James Lewis asks the question of why Amazon's growth is accelerating with time whereas most other companies' growth decelerates with time. That's the difference between marketplaces where the winner takes all most of the time, and product/services businesses.
And flywheel effects. Monetizing costly achieved data, customer relations, marketing / word of mouth etc etc. Marketplaces are just an overarching echelon of those principles.
@@dinoscheidt Right. In addition to its marketplace business, Amazon also has AWS which is a platform business, and platforms also show flywheel effects. If you have petabytes of data stored on AWS and your software is dependent on AWS services, you are unlikely to move onto another cloud. Also you are more likely to find hires who are skilled in AWS vs other clouds. In addition, Amazon and AWS have received a lot of VC money, which helps when it comes to marketing and execution capability.
Also at 11:45 he wondered why each new employee was bringing in more money. I think an even easier way of explaining this is just 'economies of scale'. Yes the interactions between the teams are vital to enable this over the long term but I think the general behaviour is well understood and not that surprising
I'm sure the software architecture and organisation structure is a big part of it, it seems like some of the super-linear growth of Amazon could be explained by it's monopolistic & monopsonistic behaviours. The flywheel that Doctorow and Giblin talk about in "Chokepoint Capitalism". It seems like a huge big emission to talk about getting bigger making getting bigger easier without mentioning that they're so big that we can't approximate their share on the markets as zero, like we could with a small company in a highly competitive market. In some ways that should make it harder for them to get bigger as they've already captured the market and there's less space to grow into, but it also makes it easier as they're able to make it hard to operate competitors and they benefit from network effects.
So many scary facts laid out so clearly... and orgs still insist in change requests board, hierarchical leadership styles, org wide policies and rules for developers...
Things seem to be getting worse with often multiple cloud platforms and multiple CI/CD systems per startup not to mention anti-patterns like a FaaS per every http endpoint in effect.
That would all be fine, if those would have contract testing. If you have a schema that says email: optional, address: optional - but the business logic says it needs at least one method to contact a user = the trouble starts. But lets not kid ourselves… most startups today have trouble to not use php, rust, python etc for a simple website. Let alone system design.
A great Thoughtworks commercial - a lot of bolting together of much more elegant thought leadership (methods/concepts/practices). Good try, though, muddy thinking, especially around complex adaptive systems and flow. Relying on the early days of what used to be a good reputation for thought leadership.
This guy’s scarey smart and he communicates extremely effectively. Wow.
Interesting line of thought. 37:35 James Lewis asks the question of why Amazon's growth is accelerating with time whereas most other companies' growth decelerates with time. That's the difference between marketplaces where the winner takes all most of the time, and product/services businesses.
And flywheel effects. Monetizing costly achieved data, customer relations, marketing / word of mouth etc etc. Marketplaces are just an overarching echelon of those principles.
@@dinoscheidt Right. In addition to its marketplace business, Amazon also has AWS which is a platform business, and platforms also show flywheel effects. If you have petabytes of data stored on AWS and your software is dependent on AWS services, you are unlikely to move onto another cloud. Also you are more likely to find hires who are skilled in AWS vs other clouds. In addition, Amazon and AWS have received a lot of VC money, which helps when it comes to marketing and execution capability.
Also at 11:45 he wondered why each new employee was bringing in more money. I think an even easier way of explaining this is just 'economies of scale'. Yes the interactions between the teams are vital to enable this over the long term but I think the general behaviour is well understood and not that surprising
Yes, I thought James glossed over Amazon's monopolistic & monopsonistic behaviours.
Law of diminishing marginal utility is a universal principle.
I'm sure the software architecture and organisation structure is a big part of it, it seems like some of the super-linear growth of Amazon could be explained by it's monopolistic & monopsonistic behaviours. The flywheel that Doctorow and Giblin talk about in "Chokepoint Capitalism".
It seems like a huge big emission to talk about getting bigger making getting bigger easier without mentioning that they're so big that we can't approximate their share on the markets as zero, like we could with a small company in a highly competitive market. In some ways that should make it harder for them to get bigger as they've already captured the market and there's less space to grow into, but it also makes it easier as they're able to make it hard to operate competitors and they benefit from network effects.
A big like 👍 Thanks for sharing this 🙏
Great like would love to know more about how to build organization based on social network
So many scary facts laid out so clearly... and orgs still insist in change requests board, hierarchical leadership styles, org wide policies and rules for developers...
Good talk. Would of loved to spend more time on real world examples
what is the next big thing wrt organisational design / structure after Team Topologies? What to inspire next?
what does it mean by building organization like social network? How?
Thank you for the lecture, quite fascinating and informative.
Things seem to be getting worse with often multiple cloud platforms and multiple CI/CD systems per startup not to mention anti-patterns like a FaaS per every http endpoint in effect.
That would all be fine, if those would have contract testing. If you have a schema that says email: optional, address: optional - but the business logic says it needs at least one method to contact a user = the trouble starts. But lets not kid ourselves… most startups today have trouble to not use php, rust, python etc for a simple website. Let alone system design.
A great Thoughtworks commercial - a lot of bolting together of much more elegant thought leadership (methods/concepts/practices). Good try, though, muddy thinking, especially around complex adaptive systems and flow. Relying on the early days of what used to be a good reputation for thought leadership.