Really good piece. Do you have a sense of when fragmented production changes from a benefit to a cost? What differentiates a TV from a car in terms of how easy production is to coordinate?
That's a good question. When products mature, you probably want there to be relatively few producers; we know roughly how a microwave or bookshelf or toothbrush should work, so stamping out lots of units matters more than innovation/flexibility. But that model can also depend on the components; if a product is changing by getting smarter (more options for the user, a screen to interact with), then a) you run into the scale economics of chips, displays, etc., which pushes for mass production, and b) you can get mass customization at the software level instead of hardware. So the winning move in that case is to have excessively good hardware, an always-on Internet connection, and continuously updated/customized software.
So as a shorthand, it depends on:
1. How much does the product change each year?
2. Is there a fixed capital cost for making X of it, and economies of scale for chunkier versions of that cost?
3. Is it cheap to ship relative to its value? (Tends to lead to a global monopoly)
4. What is the cost curve for the most expensive component, over time and units, and is that component used in other products? (anything that uses parts that also show up in smartphones is getting continuously cheaper to make, so you'd expect more experimentation).
Really good piece. Do you have a sense of when fragmented production changes from a benefit to a cost? What differentiates a TV from a car in terms of how easy production is to coordinate?
That's a good question. When products mature, you probably want there to be relatively few producers; we know roughly how a microwave or bookshelf or toothbrush should work, so stamping out lots of units matters more than innovation/flexibility. But that model can also depend on the components; if a product is changing by getting smarter (more options for the user, a screen to interact with), then a) you run into the scale economics of chips, displays, etc., which pushes for mass production, and b) you can get mass customization at the software level instead of hardware. So the winning move in that case is to have excessively good hardware, an always-on Internet connection, and continuously updated/customized software.
So as a shorthand, it depends on:
1. How much does the product change each year?
2. Is there a fixed capital cost for making X of it, and economies of scale for chunkier versions of that cost?
3. Is it cheap to ship relative to its value? (Tends to lead to a global monopoly)
4. What is the cost curve for the most expensive component, over time and units, and is that component used in other products? (anything that uses parts that also show up in smartphones is getting continuously cheaper to make, so you'd expect more experimentation).