Say the Real Number
Picture a contractor who tells you the bathroom will take three weeks. It takes two. You're delighted, for about a day. Then a small, unflattering thought arrives: if he knew it would take two, why did he say three? And the next time he gives you a number, you quietly shave a third off it in your head.
That little mental adjustment is the whole problem with one of the most repeated pieces of work advice going. "Under-promise and over-deliver." It shows up in onboarding decks, in mentoring chats over bad coffee, in the margins of sales training. It sounds humble. It sounds safe. It feels like the grown-up way to handle expectations.
I think it mostly isn't. Not always, and I'll get to the exceptions. But as a default habit, padding your promises so you can beat them teaches the people around you something you didn't mean to teach.
What the padding actually says
A promise is supposed to carry information. When someone asks when a report will be ready, they aren't asking for a performance. They're asking for a number they can plan around. Maybe they need to book a meeting, or tell their own boss, or decide whether to start something else in the meantime.
If you say Friday and you privately mean Wednesday, you've handed them a bad number on purpose. They plan around Friday. They push back the meeting. They tell their boss Friday. Then Wednesday comes, the report lands, and they get a pleasant surprise that is also, if you squint, a small inconvenience. The slack you built in was slack they paid for.
Do it a few times and something worse happens. They learn your exchange rate. Your "Friday" becomes their "probably Wednesday," and your estimates turn into a code that has to be translated before anyone can use them. People who work with a chronic under-promiser end up doing arithmetic on every sentence. Two days means one. "I'll try" means yes.
Careful people should find the next part the most uncomfortable. Once the discount is priced in, you lose the ability to give a real warning. Suppose one week the job genuinely will take until Friday. You say Friday. They hear Wednesday. Now you're late against a number you never gave.
There's also a slow drift in your own head. If you pad every estimate, you stop knowing what your honest estimate even is. The padding becomes the number. Ask yourself how long something will take and the answer that comes back already has the cushion folded in, so you can't tell anymore whether you're fast or just careful with your words. That seems like a real loss to me, though I'll admit I can't prove it with anything better than watching how estimates tend to go.
Saying the real number out loud
The alternative sounds almost too plain. Say what you actually think will happen. If it's Wednesday, say Wednesday. If you're unsure, say how unsure: "Wednesday if the numbers from finance arrive Monday, Friday if they don't." That sentence is longer than "Friday," and it's far more useful, because it tells the other person what the risk is and where it lives.
People resist this for a reasonable reason. Accurate promises get missed sometimes. That's what accurate means. If your honest guess is right about half the time, then about half the time you'll be a little late, and being late feels worse than being early by a mile. The under-promiser never has that bad moment. They've traded it for a steady leak of credibility that nobody mentions out loud.
The bad moment has a cheaper fix than padding: tell people early. A promise you revise on Tuesday, before anyone has acted on it, costs almost nothing. "Heads up, finance is slow, this is now Friday." Most people take that well. What they don't take well is silence followed by a miss. If there's a habit worth building, it's the early text.
Think about the weather forecast. Nobody wants a forecaster who says 90 percent chance of rain every day so they'll look good when it's sunny. You'd stop checking within a month. The forecasts people trust are the ones that say 30 percent and mean it, which means sometimes it rains on a 30 percent day and the forecaster was still right.
None of this means you should promise the best case. Optimism is its own kind of dishonesty, and it's probably more common than sandbagging. The goal is the middle number, the one you'd bet your own lunch on. Then, if you finish early, fine. If you finish late, you said so in advance.
I can think of places where the old advice still earns its keep. A brand-new client who has been burned before may need a few easy wins before they believe anything you say. A situation full of unknowns you can't even name might deserve a wider range than usual. But a wider range, said honestly, is different from a smaller promise you secretly expect to beat.
A reasonable test is to imagine the other person could see your private estimate next to your public one. If the gap would embarrass you, it's probably too big. If they'd shrug and say, "Sure, makes sense," you're fine.
The contractor, in that version of the story, says two weeks, maybe three if the tile backorders. It takes two and a half. You aren't thrilled. But the next time he gives you a number, you write it on the calendar exactly as he said it, and that turns out to be worth more than a surprise.
Jonathan Endicott (Jonathan Weston Endicott) is an entrepreneur and creator who writes about building, momentum and discipline. More at about Jonathan Endicott, on Instagram at @jonathanwendicott and at linktr.ee/jonathanwestonendicott.