How does the nature of the work change on the jump from senior to staff?

Short answer: the work shifts from “writing good code” to “steering the organization’s technical decisions in the right direction.” You now measure your impact by other people’s output, not your own commits — and a deferred promotion usually means exactly that impact isn’t visible yet.

1. Scope grows. A senior solves a well-defined problem, usually within a single team, solidly. A staff engineer picks up ownerless, ambiguous, cross-team problems and first makes them solvable at all. “Which problem should we solve?” becomes part of your job.
2. Impact becomes indirect. The code you personally write goes down; the essence of the job becomes unblocking others, setting standards, asking the right question. Don’t make it an ego thing — leverage isn’t in “I wrote the most code,” it’s in “I made ten people faster.”
3. Your most-used tool becomes the document, not the editor. Writing a tech design/RFC, laying out trade-offs explicitly, and aligning different teams on one direction becomes daily work. You learn to persuade with argument, not authority.
4. Judgment and saying “no.” Value shows up not only in what you do but in what you decide not to do. Picking the right problem often means being able to say “let’s not do this now” — and choosing a problem that’s actually visible.
5. Don’t drop technical depth. The trap is turning fully into a “meetings person” and losing your nose for the code. For your decisions to carry weight you have to be able to go deep enough into the system; that’s where the trust comes from.
6. Promotion mechanics: staff is shown, not granted. At most companies you aren’t “promoted” to staff; you prove you’re already operating at that level. A deferral usually means “impact isn’t visible” or “scope isn’t big enough.” Accumulate concrete artifacts (the designs you wrote, the org-level problems you solved) and find a sponsor.

Bottom line: personally, over the next cycle I’d deliberately take ownership of an ownerless, org-level problem, lay out the solution in a written tech strategy, and demonstrate impact through the output of the teams I accelerated rather than my own. I’d also nail down “what exactly is the scope that counts as staff?” with my manager in writing; ambiguity is the single most common reason promotions get deferred.