What's the best way to give estimates for engineering work?

PayPal / Eng beepboopp
Mar 7 8 Comments

Accounting for:
- time spent debugging/errors
- unforeseen interrupting work
- wanting to deliver quickly but not set up unrealistic expectations

comments

Want to comment? LOG IN or SIGN UP
TOP 8 Comments
  • Google goooooooo
    “It will be done when it’s done”
    Mar 7 4
    • Western Digital xfcs35
      Wow, the way things get done in G
      Mar 7
    • Intel UGeJ58
      ^Isnt that always better than just 3xing and sit on your hands 1/2 of the time?
      Mar 7
    • Western Digital xfcs35
      No is not. When you hire a contractor to do a job in your house you do not expect them to answer "it will be done when it is done" . You don't expect that answer either from a mechanic fixing your car charging you by the hr. If those workers can give an estimate, why you cannot?
      Mar 7
    • Intel UGeJ58
      Yes, let's try to broadly stroke the brush and apply this to unskilled labor. The contractor can explain what they will do and you will understand 100%

      If you have a novel problem that has no previous solution and apply a hard timeline, you will get the quality of unskilled labor. That, or just shit the bed and wonder what went wrong at the end of the day.

      If G explained all of the nuances of the problem and you can understand 100%, then you shouldn't pay him top dollar.
      Mar 7
  • FactSet
    roadies19

    FactSet

    PRE
    Bridgewater Associates
    roadies19more
    Whatever you think x1.5/2
    Mar 7 0
  • New / Eng sparked
    1. Have well defined requirements
    2. Have a comprehensive knowledge of existing product as to understand any “gotchas”/design rework
    3. Have an understanding of what is involved in implementation
    4. Understand the complexities of testing/validation

    If you don’t know all of the above it’s best to plan for a small investigative task to document the above before giving any reliable estimation

    Then double your estimation.
    Mar 7 0
  • Oracle now@google
    My rule of thumb is to guess then multiply x3. After that create a detailed list of stuff that fits the estimate.
    Mar 7 0