You've been running point on the thing that used to need someone senior's sign-off, for the better part of a year now. Your name is the one people go to when something breaks, and you were in the room making the call at the last planning cycle, not reading someone else's summary of it afterward. Nobody has said the word "promotion" out loud. You keep assuming that's next, because from where you're sitting it looks obvious the title hasn't caught up to the job.
You've been telling yourself the case will make itself. The work is good, people rely on it, and eventually someone senior will connect the pattern and put your name forward. That belief has been quietly reinforced by exactly the kind of work that has a habit of disappearing the moment it stops causing problems, which is the reason "they'll notice eventually" doesn't hold up as a strategy, and the reason nothing has moved yet.
Promotions get decided by people who weren't in the room
A promotion isn't a slow accumulation of goodwill that eventually tips over into a title change. It's a specific decision, made in a specific window, a calibration meeting, a headcount review, a comp cycle, by people assembling a case from whatever evidence exists at that exact moment. Your manager might genuinely want to advocate for you in that room, but the room itself runs on documented scope and outcomes, not on a felt sense that you've been carrying more than your job description says.
This is why reliable, quietly excellent work is the hardest kind to get promoted on. The same instinct that makes automation invisible once it stops breaking makes a year of expanded scope invisible too, for the same reason: nothing about it generated an event. No outage, no escalation, no moment that forced the org to notice you'd been operating above your title. The work simply continued to be fine, and "fine" doesn't get raised in a calibration meeting on its own.
The second problem compounds the first. Even a manager who has genuinely noticed still has to make your case to people who've never seen you work directly, the exact translation problem from explaining technical work to people who don't care how it works: the committee doesn't want the story, it wants a defensible summary it can compare against a level definition. If your manager is translating from memory instead of from material you've handed them, the case gets thinner every time it's retold.
Three moves that build the case instead of waiting for it
None of these require you to become a different kind of person at work, and all three are things you can start this week.
Keep a running case document, not a pre-cycle scramble. The one-sentence trail you're already leaving after anything you ship is the raw material, the goal now is to fold it into something structured enough to hand someone. A dated doc with three columns does it: what you owned a year ago, what you own now, and the outcome where you have a number for it. Add one line the day something happens, not the week before the cycle opens, or you'll be reconstructing a year from memory under a deadline.
Ask what the bar actually is, in words, not vibes. "I'd like to be considered for [level] at the next cycle, can you tell me specifically what's different between where I am and that bar?" This does two things: it converts a vague hope into a checklist you can actually close gaps against, and it puts your manager on record before the cycle, which matters if you still get passed over and need to know whether it was the case or the timing.
Hand your manager material, not a memory. They're the one translating your case upward to people who've never watched you work. Send the running doc ahead of the cycle with something like: "Ahead of calibration, here's a summary of what I've owned since [date], in case it's useful in the room." You're not asking them to remember your year. You're giving them something to read from.
This week
Open a blank doc today and write three bullets: three things from the last month that were genuinely a level above your current title, one line each, what it was, what it replaced, what happened as a result. That's the whole task. You're not committing to a weekly ritual yet, just proving to yourself the material exists and it's more concrete than the vague sense that you've been doing more.
That's the whole arc, five posts end to end: making invisible work visible first, then translating it for someone who doesn't care how it works, then pricing the quick favours before they became a second job, then declining what didn't fit without it reading as can't cope, and now turning all of it into the case for the title you've already been operating at. None of it works as a one-off. It's the same discipline, aimed at five different rooms.
Frequently Asked Questions
Isn't this just bragging?
It reads that way right up until you notice the difference: bragging asks people to be impressed, this asks people to have the information they need to make a decision that's already on their calendar. A promotion case built from dated, specific evidence isn't a performance, it's a document someone in that calibration room needs whether or not you supply it, the only choice is whether it's accurate and yours, or reconstructed secondhand by someone translating from memory.
What if my manager genuinely doesn't seem to care about advocating for me?
That's worth knowing directly rather than assuming, and the "what's the bar" question above is exactly how you find out. A manager who deflects a specific, reasonable question about a specific level is telling you something different from a manager who answers it clearly and then doesn't follow through. The first is a signal to route around, the second is worth one direct conversation about what's actually stopping it before you conclude the whole system is broken.
What if there's no formal calibration cycle where I work?
The mechanism still applies, the timing is just less predictable. Somewhere, a decision about your level gets made by someone other than you, at some point that isn't fully in your control, which makes the running document more important, not less, because you can't time it to a known window. Keep it current and ask the bar question whenever headcount or budget conversations are visibly happening, not only on a fixed calendar date.
What if I ask and the answer is no?
Then you've converted a vague uncertainty into a specific gap, which is strictly better information than silence. A clear no, with the actual missing piece named, tells you exactly what to build toward next cycle. A silence you never asked into just leaves you rerunning the same "they'll notice eventually" loop this whole arc has been arguing against.
How far ahead of the cycle should I actually start this?
As early as the work itself starts looking like a level up, not a quarter before the meeting. The running doc costs one sentence per event, so there's no real cost to starting today even if the next cycle is months out, and a doc with six months of dated entries reads completely differently from one assembled the week before.
Does this work the same way if I want a management track instead of a senior IC track?
The mechanism is identical, only the evidence changes shape. An IC case is built from scope and outcomes, a management case needs the same discipline applied to things like who you've grown, what decisions you made about the team's direction, and what happened as a result. Track the events, not just the title you're aiming at, and the document works the same way either track.