The Three Levels of Product Work
This is one of the most quietly useful mental models I have adopted, because it explains a thing that otherwise looks like personality clash. Shreyas Doshi argues that all product work happens at three distinct levels, and that most of the friction on product teams comes not from disagreement about what to do but from people unknowingly operating at different levels.
Impact: The actual value created, for users and the business. Did the thing move an outcome that matters? In the long run this is the only level that counts. Execution: The work of building and shipping: the specs, the coordination, the velocity, hitting the plan. Necessary, visible, and the level most PMs default to because it is concrete and you can feel productive doing it. Optics: How the work is perceived, by your team, your stakeholders, your execs. The narrative, the demo, the credit, the framing.
#Why operating at different levels causes conflict
Here is the insight that earns this model its place. Picture three people in a launch review. One asks "did it actually move retention?" (Impact). One asks "did we ship on schedule?" (Execution). One asks "how did this land with the exec team?" (Optics). None of them is wrong. They are just at different levels, and each one thinks the others are missing the point. Map a tense meeting onto these three levels and a startling amount of "they don't get it" dissolves into "we were answering different questions." Naming the level is half the fix, and it is one of the more useful moves you can make in cross-functional disagreement.
#Optics is not a dirty word, and that's the hard part
The instinct of the earnest PM, mine included early on, is that Impact is real and Optics is politics, so you do great work and let it speak. It will not. Optics is the level where great work gets recognized, resourced, and continued. If excellent work is perceived as mediocre, it gets defunded and the impact never compounds. Doshi's point is not "play politics." It is that managing perception is part of the job, and refusing to do it is self-sabotage dressed up as integrity. The senior move is to operate at all three deliberately: do the impactful thing, execute it cleanly, and tell the story so it survives.
| Level | The question | Failure mode |
|---|---|---|
| Impact | Did it move what matters? | Skipped, confusing motion for outcomes |
| Execution | Did we build and ship it well? | Over-indexed into a feature factory |
| Optics | How is it perceived? | Ignored (sabotage) or gamed (all spin, no substance) |
The balance is the craft. All Optics and no Impact is a con that unwinds fast. All Impact with zero Optics is a martyr whose work quietly dies. The PMs who compound are the ones who refuse to treat any level as beneath them.
#How it connects
This is the natural partner to Doshi's LNO framework: Leverage tasks are usually Impact-level work, and a lot of Overhead is Optics you can satisfice. It also gives teeth to Outcomes Over Outputs, because "outputs over outcomes" is exactly the disease of getting stuck at the Execution level. And the Optics level is the honest, named home for the executive communication and managing-up work that the rest of the Atlas treats as a skill. Here it is a level of the work itself, not an afterthought.
#Continue Reading
- The LNO Framework is Doshi's effort-allocation companion, where Leverage is usually Impact-level work.
- Outcomes Over Outputs is the output trap named: it is what getting stuck at Execution feels like.
- Managing Up for the discipline of operating well at the Optics level with leadership.
- Executive Communication for Optics done with substance: telling the story so the work survives.
- The Frameworks Toolkit for the full index of tools and when each earns its place.