Most promotion cases fail for the same reason: they describe the candidate rather than demonstrate them. "I'm ready for the next level" and "I consistently show strong technical judgment" are claims, and claims from an interested party carry very little weight with a committee that's never worked with you directly. What actually moves a promotion decision is evidence specific enough that someone with zero context could read it and reach the same conclusion independently.
Start collecting evidence before you need it
The biggest mistake is starting to assemble a promotion case the month before the cycle closes, trying to reconstruct six months of impact from memory. By then the details are gone — the exact metric improvement, the name of the person who said your design doc changed their approach, the specific decision you made under pressure that avoided an outage. Keep a running document, updated in the moment, not reconstructed later. When you ship something meaningful, resolve a hard incident, or get unsolicited praise from a peer, write it down immediately with enough detail that it's still legible in eight months.
Evidence beats adjectives
Every level's rubric uses words like "significant," "consistently," and "independently," and the temptation is to just restate those words back at the committee. That doesn't work because the committee's job is to judge whether the words apply — repeating them back is circular. Replace every adjective with a fact that implies it. Instead of "led a complex cross-team project," write "coordinated the migration across three teams, wrote the technical plan that became the reference doc for two other org migrations, and delivered it two weeks ahead of the original estimate." The second version doesn't need to claim leadership; the facts make the claim for you.
Quantify where you honestly can, but don't force numbers onto things that don't have them — a fabricated or stretched metric is worse than no metric, because a sharp committee member will notice the stretch and it undermines the rest of the packet. Qualitative evidence is fine when it's specific: a direct quote from a peer's feedback, a concrete description of a decision and its consequence, a before-and-after description of a process you changed.
Third-party corroboration matters more than self-report
A committee weighs a peer's unprompted praise far more heavily than your own description of the same event, because self-assessment has an obvious incentive problem and peer feedback doesn't. Actively collect this — ask people you've worked closely with for specific, written feedback while the work is still fresh in their memory, not vague endorsements after the fact. A manager or skip-level who can speak to your impact from their own vantage point, with specifics, is worth more in a packet than three more bullet points you wrote yourself.
Calibrate scope against the actual bar
Before writing anything, get an honest read on what the target level actually requires at your company — the written rubric is a starting point, but the real bar is often set by recent examples of people who were and weren't promoted. Ask your manager directly what's missing, not just what's present. A packet stuffed with evidence for the level you're already at, no matter how well-written, doesn't make a case for the next one; it just proves you're solid where you are. The strongest cases show a small number of examples that clearly demonstrate next-level scope, judgment, or impact, rather than a long list that mostly demonstrates competence at the current level with more items.