Bad Estimates Destroy Careers (Here’s What to Say Instead)

Your manager asks for a timeline. You

don't even know what the project is, but you can feel the pressure. So, you

basically say two weeks, which I think is something that all devs say when they're not exactly sure. You just lied.

In 2 months, when the think is smoking in the corner, that lie is on you. Let

me tell you what to say instead. Giving

a deadline for something you don't understand is suicide. I've seen careers end because of that. People burn out and projects implode all because you said two weeks of I need more information.

You were set up to fail. I'm making sure it doesn't happen again. You're in a meeting or a Slack DM. Someone asks for a date. You know they don't have a clue

a date. You know they don't have a clue what the heck they're asking for. Maybe

it's a PM who hasn't seen the legacy mess. Maybe it's a stakeholder that

mess. Maybe it's a stakeholder that can't tell the difference between a stack and a hay stack. You do some best case scenario math. Then you say it two weeks. The second you say that, you just

weeks. The second you say that, you just sign a contract that you can't keep. 2

weeks assumes the world is perfect. It

never is. Those people aren't trying to trick you. They're just anxious. They

trick you. They're just anxious. They

want a number because a number feels like a plan. But a number based on nothing is just BS. And you're the one taking the hit when it's late. Rule one,

never give a single date. Ever give a range. Tell them this is 3 to 6 weeks.

range. Tell them this is 3 to 6 weeks.

That's the truth. Early on, your estimate can be off by quite a bit.

That's just reality. But the more you learn, the range gets smaller. Something

vague earns a vague timeline. If they

want precision, they're going to have to give you precision. If they push, tell them you need 2 days for research. It

protects you. It shows them they don't actually know what they want. Reframe

the problem. Ask what they're actually trying to solve. Solving the outcome is better than building a complicated mess.

If they don't take a range, use one of these scripts. The risk transfer. If I

these scripts. The risk transfer. If I

give one date, I'm probably 70% wrong. I

can say 4 weeks, but if nothing goes wrong. Is that what you want? The

wrong. Is that what you want? The

discovery proposal. I can't give a real number yet. Give me 2 days to dig in.

number yet. Give me 2 days to dig in.

I'll give you a range I actually believe in by, let's say, Friday. Or D. No, with

a pathway. I can't do 2 weeks without shipping junk. I can finish my current

shipping junk. I can finish my current project on time or I can pull in another dev. Which one? Saying no to a lie is

dev. Which one? Saying no to a lie is your job. Saying yes to a fantasy helps

your job. Saying yes to a fantasy helps nobody. Bad estimates have a real cost.

nobody. Bad estimates have a real cost.

Half of all projects go over budget because of them. You end up working weekends and cutting corners. High

pressure code has 15 times more bugs.

You're not saving time. You're borrowing

at a massive interest rate. This is

about the business. Bad estimates cost money and trust. Your job isn't to make people feel good with a fake date. It's

to give them the truth so they can decide. If you take on impossible

decide. If you take on impossible deadlines, you'll burn out and you'll start to hate that job. Next time you're cornered, remember the defense. One,

give a range, not a date. Two, push for info. Three, reframe the problem.

info. Three, reframe the problem.

Memorize this. I hear you need a timeline. Right now, it's between X and

timeline. Right now, it's between X and Y. If you need it tighter, I need more

Y. If you need it tighter, I need more info. Does that work? You're not being

info. Does that work? You're not being difficult. You're being a professional.

difficult. You're being a professional.

Stop making people comfortable with lies. Start making them uncomfortable

lies. Start making them uncomfortable with the truth. That's how you do the