Alternative example - one crucial component is out of stock, but manufacturer has a different model of a microcontroller, that is pin compatible but has half of the RAM. You run an analysis and your firmware should run fine, but you may need to do some tweaks here and there. You place an order with an alternative part.
Suddenly you have to support multiple versions of the same product. Then imagine your product team comes up with a new feature that also users ask for. Problem is that it may not work with the revision with half the RAM. Would you offer an update only to a section of customers? Would you replace the boards for free once they become available? Or do you make a completely new version of a product (which is going to piss customers who just bought the "old" version).
Etc. Etc. These supply chain problems make any planning not possible. The common "solution" is to only start making a product once you have all parts physically available. But this can be an enormous upfront cost, as you can't order parts as pre-orders for the product come in. Then let's say you created 1000 units, sold all. In the meantime 5% of crucial components go out of stock and lead time is 52 weeks and growing. You can probably sustain the company for a year and re-design the product to use different parts, but experience shows that they may get out of stock too. You are going to have to focus 90% of your time chasing components than developing the product. This is not viable. Many businesses are closing now because of this.