One piece of work can have several results, and all of them can be true. Say you fixed a problem that saved your sales team hours of chasing every month. Your manager will care about those hours. But the same fix may also be why new customers now start four days sooner, and that is what a VP of sales cares about. The onboarding team might care most that they no longer apologize to customers. So the result you lead with matters, because it decides who can use what you tell them. If you lead with one the listener doesn’t care about, they won’t remember it and they can’t pass it on.
But when you ask people what they’ve been working on, a lot of them describe the work itself. They tell you what they built, fixed or ran. That’s accurate, but it leaves the listener to work out why it mattered. They rarely have the time or the context to do that for you.
This post is about what you say. It isn’t about how to get noticed (our older piece on visibility at work covers that), and it isn’t about how to ask for the raise. Instead it’s the narrower follow-up to our guide on growing your career without becoming a manager. First you’ll see one piece of work described for four different listeners. After that we cover how to word your part of something a whole team did. The last part is what to say when the work has no number attached.
What does it mean to communicate your impact at work?
Communicating your impact at work means telling someone what changed because of your work, in terms they already care about. Impact here is simply the change your work produced that matters to someone other than you.
That definition only works if you pull apart three things that usually get blurred together. The first is what you did. That’s the meetings, the analysis and the afternoons spent rebuilding a spreadsheet. The second is what you handed over, such as the report, the new process or the fixed form. The third is what’s different now because of it. That might be fewer errors, a faster turnaround or a team that no longer has to chase anyone. Most descriptions of work stop after the first two, because those are the parts you lived through and remember best.
Julia Evans described this gap on her engineering blog in her 2019 guide to writing a brag document:
It’s easy to accidentally write “I shipped $feature” and miss the follow up (”… which caused $thing to happen”).
She was writing for software engineers, but the same gap turns up in every function. A finance analyst says they “rebuilt the reconciliation.” An HR coordinator says they “launched the new leave form.” Both sentences are true, but both stop just before the part anyone else needs.
It’s tempting to leave that part for your manager to fill in, but you can’t. The same post makes the point that if you don’t remember everything important you did, “your manager (no matter how great they are!) probably doesn’t either.” That isn’t a failing on their part. Your manager is holding several people’s work in their head, so they simply have more to remember than you do. Even so, they’re the one who has to explain your work to people who never saw it.
That’s why it helps to keep a record, which is a running log of what you did and what changed. Our guide to self-evaluation in performance reviews shows how to build one. This post starts a step earlier, with the person who’s listening. It also asks a different question from our post on making an impact at work. That post is about doing work that matters in the first place, while this one is about describing it.
How do you turn the same work into outcomes for different listeners?
The easiest way to see this is to follow one piece of work through four conversations. The example is illustrative, so Leonie and her colleagues are made-up people. The company, the dates and every number are invented for the purpose as well.
Leonie is a sales operations analyst at a software company of about 400 people. She spent most of last quarter on one problem. Whenever a deal closed, the account executive handed it to the customer onboarding team through a free-text note in the CRM. But the note was often missing what onboarding needed: the billing contact, the number of sites and the contract start date. So onboarding would go back to sales to ask, and the new customer would wait.
Leonie did four things about it. First she agreed with the onboarding team which details they needed, and only then did she build a handoff form that made those three fields required. She also set up a 15-minute check every Friday to look over each deal closed that week before it moved across. Last, she fixed a report so a missing billing contact gets flagged on the day the deal closes. Ask her what she’s been working on and she’ll tell you exactly that: a new form, a weekly check and a fixed report. Every word is accurate, but none of it says what changed for anyone.
Here’s what was true by the end of the quarter. Sales closes about 40 deals a month. Requests from onboarding back to sales for missing details fell from about 30 a month to about 8, and Leonie took that count from onboarding’s ticket queue. The average time from a closed deal to the customer’s kickoff call also went from about 12 working days to about 8. But onboarding hired a coordinator that quarter, so not all of that drop belongs to the form. Leonie also asked three account executives how long each of those requests used to take them to chase down. They said somewhere between half an hour and an hour. Those are the facts, and every version below uses them and nothing else.
How to find out what each listener cares about
Before you can pick the right outcome, you need to know what the person listening is measured on or worried about. What follows is our editorial advice. It’s drawn from how performance conversations tend to work rather than from any study. It’s practical all the same, and you can do most of it without asking anyone anything.
Start with their goals, if they’re written down anywhere. You might find them in a team plan, a set of quarterly targets or the slide their own boss presents each month. Next, listen to the questions they ask in meetings. The question someone keeps asking is often the one they keep getting asked by the person above them. It also helps to know what changed for them recently. If they’ve had a new target, a reorganization or a bad quarter, that’s usually where their attention has moved. What they repeat tells you something too. If your manager has mentioned the same customer complaint three times this month, that complaint is on their mind.
You can also just ask, and for some listeners that’s the only way to find out. Ask a partner team what’s been hardest for them about the thing your work touched. Ask your manager what their own manager keeps asking them about. Neither question is awkward, and both tell you something you couldn’t have guessed. Will Larson’s template for engineering promotion packets puts a version of the same question to the people around you: “What do they value about your work?” He wrote it for staff engineers, but it works in any function.
This is the practical end of influence, because you can’t move someone with a point they don’t care about. It doesn’t matter how true the point is.
Your manager
Leonie’s manager is Conrad, and he runs sales operations. He’s measured on how much of the sales team’s time goes on selling rather than admin, and on how clean the CRM data is. His own manager asks him about one thing more than any other: how much capacity the account executives have. So when Leonie talks to Conrad, the outcome to lead with is the time the sales team got back.
In their next 1:1 she might put it like this: “Onboarding’s requests back to sales for missing details went from about 30 a month to about 8. The reps I asked said each one took them half an hour to an hour to chase, so that’s somewhere between 11 and 22 hours a month back for the sales team.” The range matters, because Leonie doesn’t know the exact time each request took. If she picked a single number, Conrad would be the one stuck defending it when someone asked where it came from. A range with its basis stated is something he can repeat without worrying. The count is safe for the same reason. It comes from onboarding’s ticket queue, so he can check it himself.
After that she adds the part Conrad is least likely to know, which is how the work got done: “Glenn’s onboarding team agreed the three fields with me before I built anything, and the Friday check takes 15 minutes. Anyone on the team could run it.” Neither sentence is about the result, but both matter to him. The first tells him the form isn’t something sales imposed on onboarding, so it won’t turn into an argument between the two teams next quarter. The second tells him the fix doesn’t depend on Leonie being there. That’s what a manager who’s responsible for the team’s capacity wants to hear about any new process.
That gives Conrad a version he can use. He can put the 11-to-22-hour range in his own update to his boss, and he knows where the count lives if anyone asks. He also knows the Friday check is a task he can hand to someone else when Leonie moves on to the next thing.
Your skip-level
Delia is Conrad’s manager and the VP of sales. Leonie assumed Delia would care about rep capacity just as Conrad does, so she prepared a sentence about hours. But that was only a partial fit. When Leonie asked Conrad what Delia had been asking about lately, he told her Delia is judged this year on how many new customers renew after their first year. He also said Delia keeps asking why new customers take so long to get going.
That means the outcome that matters to Delia is the kickoff time, which went from 12 working days down to about 8. The trouble is that it’s also the number Leonie is least sure of. Onboarding’s new coordinator played a part, and she only has one quarter of data. So when Delia asked how the handoff work was going, Leonie led with the kickoff time. In the same breath she said onboarding’s new hire is part of it and that one quarter tells her nothing yet about renewals. Delia asked to see the figure again after the next quarter.
A skip-level is one step further from the work and has less context than your manager. That’s why a single sentence about the thing their organization is judged on does more than a summary they’ll skim. Delia doesn’t need to know what a handoff form is. She needs to know whether new customers are starting sooner.
The partner team
Glenn leads customer onboarding. His team lives with the effect of Leonie’s work every day. So when anyone wants to know whether the change was real, his is the view they’ll ask for. What matters to Glenn is what got easier for his team, but that outcome isn’t Leonie’s to claim. His team’s faster kickoffs come partly from his new coordinator and partly from their own scheduling. If she told him the form had sped up his team, he’d know it wasn’t the whole story.
That’s why the translation runs the other way with Glenn. Instead of telling him the outcome, Leonie asked him what had changed for his team since the form went in, and he told her: “We used to open half our kickoff calls by apologizing for asking customers things they’d already told sales. We’ve mostly stopped.” She couldn’t have measured that outcome from her side of the handoff. She thanked him and told him she was glad the fields were the right ones. She also asked whether he’d be willing to say the same thing in writing when her review came around, and he said he would.
Will Larson makes a related point in the same promotion-packet guide: “Peers are often better at identifying your strengths and contributions than you are, and they are closer to your work than your manager might be.” The conversation helped Glenn as well. He now has an answer ready for the next time his own manager or anyone else asks whether sales operations has been any help.
The promotion or calibration committee
A committee is a group of strangers comparing many people on paper, and you’re not in the room. The self-evaluation guide covers how those meetings run and what your manager can carry into one. The translation changes here because the readers can’t ask you anything. So you have to give them everything they need on the page.
The description has to stand on its own, so Leonie explains what the onboarding team does and why a missing billing contact holds up a new customer. It also has to show scope. Her work covers every deal sales closes across two teams, which is about 40 a month. It has to carry proof the committee can check, and here that’s the ticket-queue count before and after. Last, it has to show the change lasted. Leonie can show that for the count, because she handed the Friday check to a newer analyst in the quarter’s last month and the count stayed at about 8. But the kickoff time hasn’t yet held for a second quarter, and any link to renewals is a year away. Her write-up says exactly that.
The strongest part of her packet is the paragraph from Glenn about the apologies that stopped, in his own words. It carries weight because a committee reading about someone it has never met tends to trust the person who lived with the change more than the person who made it. That paragraph exists only because Leonie asked him months earlier. How much weight the committee puts on scope also depends on the level you’re aiming at, which our guide to growing as an individual contributor explains. Making the case itself is a separate conversation, laid out in how to ask for the project or the promotion you want.
The four versions side by side
| Listener | What they’re measured on | The change they’ll care about | What to attach |
|---|---|---|---|
| Conrad, her manager | Sales time spent selling; CRM data quality | 11 to 22 hours a month back for the sales team | Ticket-queue count; how the hours were estimated |
| Delia, the skip-level | First-year renewals | Kickoff in about 8 working days, not 12 | The coordinator caveat; one quarter only |
| Glenn, onboarding | Kickoff quality and speed | No more chasing sales; no more apologizing to customers | Nothing; the account is his to give |
| A committee | Comparing many people on paper | Scope, proof and a change that lasted | Before-and-after count; Glenn’s paragraph |
Keeping the four versions honest
All four versions rest on the same facts. The only thing that moves between them is which true change Leonie puts first and how much context she attaches. She adds nothing, and she leaves out nothing that would embarrass her in front of the others.
You can test this by reading each version as the other three listeners would. Say Leonie had told Delia that kickoffs were four days faster because of her form. Say she had also told Glenn his coordinator had made the real difference. Each sentence flatters the person hearing it. But if Delia and Glenn ever compared notes, they’d find two stories. After that, neither of them would be as ready to trust the numbers in the next one. That’s where translation turns into spin. So before you say any version out loud, imagine the other listeners sitting next to you while you say it.
How do you claim your part of a team result without overclaiming?
Most results at work belong to several people, which makes describing your part awkward in both directions. If you claim all of it, the people who know better can quietly mark you down. But if you hide your part inside “the team,” nobody can tell what you did. We couldn’t find a study that measures how often either happens. So there’s no figure here, just the wording.
The three situations below come from one illustrative HR operations team at a manufacturing company, and the names and numbers are invented. This year the team ran benefits open enrollment. Afterwards payroll had to make about 50 corrections, down from about 140 the year before. Three people had a hand in that drop, each in a different way.
Odette’s result was mostly her own. She rewrote the enrollment FAQ, and questions to the HR inbox during enrollment fell from about 300 to about 120. She can say that plainly, as long as she names the one thing she depended on. The IT team put a link to the FAQ on the enrollment page itself, and far fewer people would have found it without that link. Saying so costs her nothing, and anyone who knows about the link will see she hasn’t left it out.
Anton built a defined piece of a larger result. He wrote a check that compared the benefits vendor’s file with payroll’s records before the first payroll ran, and it caught 38 mismatches. The team’s result is the drop from 140 corrections to 50, but his piece is the 38 caught before they could become corrections. This year’s total would have been closer to 90 without his check. So he says it in that order. First comes what the team achieved, and next comes which part was his. He finishes with what would have been different without it. He doesn’t say he cut corrections by two-thirds, because he didn’t.
Isla’s part is the hardest to word. She ran short briefings asking line managers to get their teams to enroll in the first week rather than on the last day, and enrollment was much more spread out this year. The trouble is that the benefits manager also moved the deadline away from month-end, so nobody can separate the two effects. That means the honest version is for Isla to say what she did and what she saw. The benefits manager owns the result, so Isla lets her say what she thinks caused it. Sometimes your part can’t be measured, and no wording changes that.
So an honest claim has three parts: what you did, what the team’s result was and your portion of it. Let the middle part come from whoever owns the result if you can, such as the benefits manager’s report. That way the listener relies on you only for the part that’s actually yours.
Don’t take the credit when it would be inaccurate or when your piece only worked because of someone else’s decision. You can also leave it alone when the listener will ask the owner of the result anyway. But don’t swing the other way and over-credit the team (“honestly, it was everyone”). That’s the same mistake in reverse, because it leaves the listener just as unable to tell what you did. If your problem is someone else taking credit for your work, our guide to managing up covers the credit-taker.
One study is relevant here, though only within limits. Christine Exley and Judd Kessler ran experiments with “over 4,000 participants from online labor markets and over 10,000 school-aged youth.” They reported the results in an NBER working paper, The Gender Gap in Self-Promotion. Its abstract says: “Women subjectively describe their ability and performance to potential employers less favorably than equally performing men.” It goes on: “Even when all incentives to promote are removed, however, the gender gap remains.” But these were online and school experiments, not workplace studies. So what follows is our own judgment, not the paper’s. People may describe the same performance differently for reasons that have nothing to do with the work. If so, it’s safer not to start from how confident you feel. Start from the listener’s goals and the checkable facts instead, because that guards against both overclaiming and underclaiming. If the hesitation feels more like not deserving the credit, our post on imposter syndrome and promotions is the better place to start.
What if your work has no number?
A lot of useful work produces no figure of its own. Some of it prevents something that then doesn’t happen. Some keeps something running or makes other people faster, and some maintains a thing nobody notices until it breaks. Tanya Reilly gave part of this a name, “glue work,” in her talk Being Glue. She describes it as work that “makes the team better,” though “people aren’t always rewarded for doing it.” Her examples come from software teams. We think the idea applies just as well to operations, finance, HR and marketing. That part is our argument, not her evidence.
Three questions help turn this kind of work into something you can say. First, what would have happened without it? That question works because prevention is invisible until you describe the thing prevented. Second, who noticed and what did they do differently? Asking that moves the evidence from your view to someone else’s. Third, what got faster or cheaper? What gets repeated less now, or is avoided altogether? That’s the question most likely to find a number if there is one. Mostly the questions give you statements rather than figures, but statements are often enough.
Evans offers a similar scaffold for what she calls “fuzzy projects that can feel hard to quantify”: “explain your goal for the work,” “list some things you’ve done towards that goal,” and “list any effects you’ve seen of the work, even if they’re a little indirect.” Here’s how that looks for Clive, another illustrative example. He’s a senior procurement analyst who checks every Monday which supplier contracts are due to renew in the next 90 days. His goal is that no contract renews without someone choosing to renew it. He gets there by sending each budget owner a list of their renewals and the date by which they’d need to give notice. The effects are concrete. Four contracts that would have renewed automatically this year were canceled or renegotiated by their owners instead. One owner in marketing wrote to him that they’d have paid another year for a tool nobody used.
Clive could produce a number here, but the honest version is careful about which one. The four contracts were worth about $60,000 a year, and that comes straight from the contracts system. Still, he can’t say he saved $60,000 because some owners might have caught their renewals anyway. What he can say is that the contracts were worth about that much and came up for a decision instead of renewing on their own. There are a few honest ways to give a number like that. You can give a range, an “about” or an estimate with its basis. A before-and-after count from a system works too, as does a quote from the person who benefited. Whichever you use, name where the number came from when you say it aloud. That matters because a figure the listener can’t trace can make them question the whole claim, including the parts that were solid.
If no number exists, say so. A plain, specific account from the person who felt the effect is better evidence than a figure you made up. That’s why the message from the marketing budget owner counts.
Where does each version go: a 1:1, a message, a review or a packet?
The translation itself doesn’t change from one venue to the next. What changes is how long you have and how much shared context you can rely on.
| Venue | Who reads it, and when | Length | What changes | Where it’s covered |
|---|---|---|---|---|
| 1:1 | Your manager, live, with time for questions | A sentence or two, then conversation | Lean on shared context; know where the number came from | Running useful 1:1 meetings |
| Written update or message | Manager and others, skimmed between meetings | A few lines | The change first; link the proof rather than paste it | Structuring an update as a story |
| Self-review | Your manager, then the calibration room | One entry per piece of work | Name where the evidence lives; state the baseline | Self-evaluation examples |
| Promotion packet | A committee that doesn’t know you | A paragraph per project | Context that stands alone; scope; partner accounts | Making the promotion case |
The punchline-first habit from managing up applies in a 1:1, so say the change first and give the detail if they want it. Three of the four venues are written, which is why written communication carries more of this than people expect. A written record matters for another reason too. Reviews tend to drift toward whatever happened recently or was most visible, which is one of the common biases in performance reviews. But a sentence written down in March is still there in December. If the conversation you’re preparing for is about pay, our guide on how to ask for a raise picks up from here.
Where this breaks
What follows is our editorial judgment, and none of it gets tidied up at the end.
Some work doesn’t change much for anyone. A quarter spent on a project that was canceled is one example, and a report nobody opens is another. Work like that has no outcome to translate, and careful wording won’t create one. That’s because the method only finds a change that’s already there. If you push it past that, it produces exactly the inflated claims the rest of this post warns against.
The listener’s priority may also be one you think is wrong. Delia cares about first-year renewals, but Leonie might believe the real problem is the discounts sales gives away at the close. Putting the change in Delia’s terms doesn’t mean agreeing that her priority is the right one. You can say the change in her terms and then make a separate point about what you think matters more. Keeping those two apart is what lets you do both without sounding as if you’re dodging her question.
Even a well-worded sentence can’t change who is in the room. Some managers don’t pass anything up, and in some companies credit follows politics more closely than it follows work. The evidence behind this post is thin, too. Evans and Larson wrote as practitioners at engineering companies, and the one study cited is about how people describe their own performance in experiments. We found no study showing that describing your impact this way gets people promoted or paid more. So treat it as sensible practice, not as proof.
You can also overdo it. If you turn every update into a case for yourself, people stop listening. Often they’re the very people you most need to reach. Most weeks, an update is about the work and what the team needs. The storytelling guide’s section on when not to tell a story covers that ground.
Start with one piece of work and two listeners
Take the last quarter of your work and pick one piece of it, ideally the one you’d currently describe most vaguely. Name two people who’ll hear about it: your manager and one other person, perhaps someone on the team that lives with the result. Write down the change from this work that each of them would care about, along with the one detail that would let them check it. If you can’t fill in the second person’s line, ask them what’s been hardest for them lately. The answer usually tells you. If the line you wrote for one of them would surprise the other, go back to the facts before you say either out loud.
Talk your version through with Merlin
Writing those lines down is the easier half. Saying one to a VP who has three minutes and a question of her own is harder, because the oral communication side of it only improves with use. Talk it through with Merlin. Tell Merlin what you worked on and who you’re about to speak to. Ask it to play that person, and see which question comes back first. Merlin runs natively in Slack and Microsoft Teams, so you can do it in the ten minutes before the conversation.