It Takes Two Weeks to do Anything

Estimating is hard. So hard that many engineers don't really try; everything they estimate will take two weeks. Change a button from red to blue: two weeks. Write a specification for a new software feature: two weeks. Build a rocket for human flight: okay, most engineers are smart enough to estimate at least 3 weeks, but you get the point. 
​
Of course, “two weeks” usually isn’t really an estimate at all. So, what do we mean when we say something will take two weeks?

• “I am a good engineer. I can envision a solution, so I can probably implement it pretty quickly.” But seeing a path is not the same thing as walking that path, yet we fall for this trap easily.​​​

• “It takes what it takes. Don’t try to pin me down.”
  
• ​“My job is engineering. Let the project manager estimate things.”
• “I have no idea how long it will take, but it won’t be done tomorrow, so go away.”

We’ve probably seen all of these excuses used. We’ve probably used some or all of these excuses.
Estimating is hard, but it is a skill that is very important to your success, your company’s success, and your customer’s success.

What techniques do you use to estimate more accurately?