"Soft skills" survives as a category because it is vague enough to be unarguable. Nobody disagrees that communication matters. Nothing in that sentence tells you what to do on Tuesday.
Promotion decisions, particularly at the step from individual contributor to senior and from senior to lead, turn on a much narrower set of behaviours. Managers who run promotion cases consistently describe the same ones. None are personality traits. All are habits that can be built in a quarter.
Writing that survives forwarding
The highest-leverage writing skill in most organisations is producing a document that still makes sense when it is forwarded to someone who was not in the conversation.
Most work writing assumes context only the author has. The subject line is a fragment, the recommendation is in the fourth paragraph, the numbers are in an attachment. It gets forwarded, and the recipient asks three questions it should have answered.
What survives has a predictable shape: the ask in the first two sentences, the reasoning below it, the detail below that. Someone can stop reading after twenty seconds and still act correctly.
How a manager observes it. They forward your document upwards without rewriting it. That is the whole test.
How to make it visible. Write the summary before the analysis. Put the decision you want in the subject line. When you send a long piece of work, lead with three bullets and attach the rest.
Running a meeting that ends in a decision
Most meetings end with a vague sense of agreement and no record of what was agreed. Whoever reliably converts a meeting into a written decision becomes structurally important very quickly, regardless of seniority.
The mechanics are unglamorous. Send a one-line purpose beforehand stating what needs to be decided. When the conversation loops a second time, name the options and ask people to pick. In the last three minutes, state the decision, the owner and the date out loud.
How a manager observes it. Things stop coming back. The same topic does not reappear in three weeks with everyone's position slightly changed.
How to make it visible. Send the summary yourself, to everyone present, in the same thread. Over a quarter those summaries are the clearest evidence of coordination you can produce.
Giving bad news early
This separates people managers trust from people managers monitor, and it is almost entirely about timing.
A slipping deadline flagged nine days out is a scheduling problem. The same slip flagged on the day is a crisis with an audience. The information is identical; only the room to act differs.
The barrier is rarely ignorance of the slip. It is the hope that it will recover, combined with a reasonable fear of looking incompetent. Early bad news reads as control, not failure — provided it comes with a proposal.
The form that works: what has gone wrong, what it means for the date, what you propose, what you need. Four sentences, no apology paragraph.
How a manager observes it. They are never surprised by your work in front of their own boss.
The one habit to start with
For the next quarter, escalate every problem at the point you first think "this might be fine". That thought typically arrives five to ten days before the problem becomes visible to anyone else. Bring one proposed option with it, even a bad one — a problem with an option attached is a different conversation from a problem alone.
Managing sideways
Most work now depends on people who do not report to you and are not measured on your outcomes. Getting things from them is learnable and almost never taught.
Find out what the other team is measured on, and frame your request in those terms. Ask for the smallest thing that unblocks you, not the ideal thing. Give them a deadline and a reason for it. Then tell them what happened afterwards — the step nearly everyone skips, and the one that makes the second request easier than the first.
Escalation as a first move works once and costs you the relationship.
How a manager observes it. Other teams mention you unprompted. Your requests stop arriving on their desk.
Absorbing ambiguity
Senior roles differ from junior ones mainly in how much of the problem arrives undefined. The junior version of good work is executing a clear brief well. The senior version is turning an unclear brief into a clear one, then executing it.
In practice this means writing down an interpretation and circulating it. "Here is what I think we are trying to achieve, here is what I am going to do, tell me if that is wrong." Most people cannot articulate a goal from scratch but can correct a draft immediately.
How a manager observes it. They stop having to specify. The brief they give you gets shorter over time, which is the clearest promotion signal there is.
Nobody is promoted for being easy to work with. People are promoted for being predictable under uncertainty, which is a different thing and much rarer.
Making the case
All five are observable only over time, which is why promotion conversations go badly when they start a month before the review. Keep a running file — one line per week: what you decided, what you unblocked, what you flagged early. It turns a subjective argument into a list of dated events.
How much you are expected to advocate for yourself varies by market. US organisations generally expect explicit self-nomination and a written case. The UK and much of Northern Europe still lean towards the manager raising it, and aggressive self-promotion reads badly. The evidence is the same; only the delivery changes.