4 ms·
Perhaps the biggest lesson here: never give someone a gut-level estimate of anything. The challenge is that a lot of clients/bosses/whatever will press you for
by jonnathanson 15y ago
Perhaps the biggest lesson here: never give someone a gut-level estimate of anything. The challenge is that a lot of clients/bosses/whatever will press you for one. Some people just need to hear a spot-estimate in order to feel comfortable.
In those cases, I've found a fairly effective verbal counter. It goes something like this: "Listen, Bob [assuming the client's name is Bob]. I don't want to mislead you here. I'd rather work up a proper estimate than throw out some number on the spot that might be incorrect. I want to do this the right way, for your benefit and mine."
If Bob presses you further, respond with "I've worked on similar projects in the past. Some have taken 2 days, and some have taken 10 days. It really depends on X, Y, and Z. I want to take a proper accounting of X, Y, and Z in this case, so that I'm not misleading you up front."
If pressed further: "Bob, obviously time is of the essence for you and for this project. I understand that, and I will do everything I can to ensure that I complete the project as quickly as possible. In the end, this will be much quicker if we get it right the first time. Give me a little bit of time to do some planning, and it'll save us a lot of time in the long run."
All of this is true, of course. It's better to aim twice and shoot once, as the saying goes. However, there will be clients or bosses who just won't accept this sort of conversation up front. It'll piss them off. So be it. FAR better to have a little bit of confrontation at the outset than a lot of confrontation later on (i.e., if you toss out a gut estimate and then have to crawl back and amend or delay it).
- alex_c 15y agoAddendum: If you absolutely HAVE to give a gut-level estimate, multiply your gut level estimate by 4 or 5 before you let it escape your mouth.
- galaktor 15y agoThat's kind of what bgruber pointed out in his comment - Scotty later reveals in TNG that he always told Kirk the work would take longer than he thought it would.
- jonnathanson 15y agoWhich is basically the time-honored "Underpromise, Overdeliver" (UPOD) method. UPOD is pretty reliable in a corporate setting, but it can be tricky in a client-contractor relationship. If you're a contractor, and you have reason to suspect that you're being vetted against other contractors, underpromising to too great an extent can work against you. So it becomes a delicate dance between promising enough to win the bid, but not promising so much that you set yourself up for failure. BTW: that was a great episode!
- gujk 15y agoWhich then earns you an accusation of "sandbagging".
- gaius 15y agoI roll a D12 then pick the units (hours/days/weeks) as appropriate.
- gujk 15y ago"Awesome. We'll have that Perl-> Rails port with MySQL -> Cassandra conversion and 1 million paid accounts migrated in at most 12 weeks, ready to handle 5 million new accounts the day of our grand re-launch? Great! Then I don't have to worry about finding a SDET or QAE or Datacenter admin to help parallelize the work and hit our release date. Marketing has the press conferencne all planned, we just needed the word from you before we filled in the Octo..Nove..Decem..January 24 launch date! Thanks!"
- gaius 15y agoOr months ;-)
- jonnathanson 15y agoI feel like you also need to factor in Charisma and Constitution bonuses there.