Different scale.
Same operating principles.

Andy Jassy’s operating principles at Amazon map to the rules we use to turn around 25-to-300-500 person B2B software companies.

Whether you run a trillion-dollar tech giant or a mid-market software business, the laws of delivery don't change:

1. Question the structure, not the headcount. When delivery slows down, boards add people. Usually, an outdated structural assumption is the real bottleneck. Amazon: Used a fulfillment crunch to scrap a 25-year-old national network for a regional one—cutting cost and delivery time together.
2. Surface the truth in the metrics. Once a problem shows up in the numbers, fixing it is just a matter of priority. If your board pack only tracks billed revenue, you stay blind to the unbilled custom work quietly draining R&D.
3. Small, accountable teams beat big org charts. Over-specialization slows delivery to a crawl. Small teams with clear end-to-end ownership ship what bloated departments can't. Amazon: Rebuilt an inference engine in 76 days with a small cross-functional team—work that once took 40 specialists a year.
4. Rebuild cleanly instead of bolting on fixes. You cannot patch your way out of ten years of one-off client workarounds. Separate custom client work from the main product and rebuild the foundation properly. In our world: One flagship platform failed three rebuild attempts. By fixing how decisions were made first, the fourth shipped in 18 months.
5. AI raises the starting point; judgment sets the ceiling. AI speeds up routine coding, but senior engineers must still own the architecture, quality checks, and how the pieces fit together. Automating a messy workflow just creates tech debt faster.
6. Never let an outside party set your roadmap. For Amazon, the trap was waiting on a dominant supplier (Intel), so they built their own chips. In B2B software, the trap is letting your loudest enterprise client dictate your product roadmap.
7. Build from the pain you’ve lived. Amazon built AWS to solve its own internal scaling headaches first. Our operating model came from the same place—running a custom dev shop bent around a few anchor clients until we redesigned how the work got done.

Most struggling software companies try to start at #4 or #5 (starting a rewrite or buying AI tools). If you want the fix to stick, start at #1 and #2.

Andy Jassy’s operating principles appear as a third-party graphic in the original LinkedIn post; it is not reproduced here.

First published on LinkedIn on 30 October 2026. The original post.

The full mechanism — how absorbed custom work prices as a services business, and how the shape is resolved — is in the report →