August 11, 2015 at 3:30 pm
David.Poole (8/11/2015)
There is a level of hell where you get committed to estimates before you find out the subject of the estimate.Does anyone have any experience of estimates that have proben to be reasonable predictors of actual activity?
Does the whole bureaucracy surrounding estimates actually justify its own existence?
Accurate estimates? Oh yes. Most of my small estimates are pretty spot on for known systems. 😀
There's a reason the question "how long will it take you to add a new user to the system" is much easier to answer than "how long will it take you to create a database for our new order tracking system". And it's not just knowns versus unknowns, it's also that small things are easier to estimate than large things.
I actually prefer the more "agile" method of relative estimating for large projects, because it actually related to work that's been done before ("this is some order of magnitude bigger or smaller than <prior piece of work>"). But to use that you have to have buy-in at all levels of the company so that everyone understands it. And in the end, all managers are going to want an ETA and a cost basis. They need them to do their jobs.
It's an ongoing educations effort as I see it. And I try to always include the words "based on what we know right now", both verbally and in print. It helps cement that an estimate is just that - an ESTIMATE. Not a commitment.
Huh. That was a longer answer than I thought it would be. :hehe:
Steph Brown
August 11, 2015 at 3:43 pm
I put the assumptions down in writing so make the post-mortem and blamestorming as quick as possible. Its hard to undergo a hand-ringing exercise when you can put a cross next to the key assumptions on which the estimates were based.
The bit I find frustrating is when the person asking for the estimate is theoretically the owner of the key metrics on which the estimate should be based but in actual fact cannot supply any of those metrics. Imagine asking the guys responsible for your infrastructure if they know how many servers they have and how many requests/sec at busy times they have going through those servers and they haven't got the foggiest idea!!!
August 11, 2015 at 5:08 pm
David.Poole (8/11/2015)
There is a level of hell where you get committed to estimates before you find out the subject of the estimate.Does anyone have any experience of estimates that have proben to be reasonable predictors of actual activity?
Does the whole bureaucracy surrounding estimates actually justify its own existence?
To wit, I take my best estimate, raise it to the next order of magnitude, and then double that. It pretty much covers unnecessary meetings, interruptions, added scope, and hand holding. It's usually pretty much on. 🙂
Note that if someone else designed whatever it is I'm supposed to be doing, I use quadruple the previous estimate instead of just doubling it because they usually have no clue as to what actually needs to be done and "oops, I forgot" changes and "can you add this one last thing" changes will occur every day until it finally comes out of UAT.
--Jeff Moden
Change is inevitable... Change for the better is not.
August 12, 2015 at 5:14 am
I also suffer from the "forgotten" requirement and "just one more thing" at the last minute that's never been heard of before, but it's what they wanted all along. It's just always there and seems to change from time to time, just to keep me guessing. I think most people are afflicted with syndrome from time to time...or all the time. I do estimates for one thing and, by the time the scope has crept from here to the moon and morphed 7 times, I end up delivering something completely different. Hence, my estimates aren't usually very accurate, but I usually beat the time that's allocated for me.
August 18, 2015 at 7:38 am
My experience with estimates is that we tend to estimate too small. Even after 4 decades of programming I still can't get it right. How could it possibly take a year when it feels like it can be done in 4 months, but ends up taking two years. I never get it right.
August 18, 2015 at 8:07 am
Ed Wagner (8/12/2015)
I also suffer from the "forgotten" requirement and "just one more thing" at the last minute that's never been heard of before, but it's what they wanted all along. It's just always there and seems to change from time to time, just to keep me guessing. I think most people are afflicted with syndrome from time to time...or all the time. I do estimates for one thing and, by the time the scope has crept from here to the moon and morphed 7 times, I end up delivering something completely different. Hence, my estimates aren't usually very accurate, but I usually beat the time that's allocated for me.
There are occasions when the business won't allocate budget for pure refactoring and optimization, so it get's slipped into deliverables, totally out of scope.
Come on folks, fess up. 😉
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
August 18, 2015 at 9:47 am
Jeff Moden (8/11/2015)
David.Poole (8/11/2015)
There is a level of hell where you get committed to estimates before you find out the subject of the estimate.Does anyone have any experience of estimates that have proben to be reasonable predictors of actual activity?
Does the whole bureaucracy surrounding estimates actually justify its own existence?
To wit, I take my best estimate, raise it to the next order of magnitude, and then double that. It pretty much covers unnecessary meetings, interruptions, added scope, and hand holding. It's usually pretty much on. 🙂
Note that if someone else designed whatever it is I'm supposed to be doing, I use quadruple the previous estimate instead of just doubling it because they usually have no clue as to what actually needs to be done and "oops, I forgot" changes and "can you add this one last thing" changes will occur every day until it finally comes out of UAT.
I've always liked this approach:
http://dilbert.com/strip/2009-12-07
When that doesn't work, I then follow something close to your method - and attach the "cone of uncertainty" with an X on the far left side (i.e. this STILL has a +/- 100% uncertainty)
----------------------------------------------------------------------------------
Your lack of planning does not constitute an emergency on my part...unless you're my manager...or a director and above...or a really loud-spoken end-user..All right - what was my emergency again?
August 18, 2015 at 10:35 am
Like I said earlier, if estimating is part guesswork, then the business should be the one providing the guesses. The business should define what is a Sale and then estimate the number of Sales per year. Then the developer can draft a physical data model and estimate storage required.
How can a humble IT developer reliably guess how many sales orders will be entered a year from now, if the VP of Sales wouldn't even care to hazzard a guess?
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
August 18, 2015 at 10:43 am
Eric M Russell (8/18/2015)
Like I said earlier, if estimating is part guesswork, then the business should be the one providing the guesses. The business should define what is a Sale and then estimate the number of Sales per year. Then the developer can draft a physical data model and estimate storage required.How can a humble IT developer reliably guess how many sales orders will be entered a year from now, if the VP of Sales wouldn't even care to hazzard a guess?
+ 100!
Separately, I'd like to thank Matt for the Dilbert link (had to copy / paste, but it got me there). One of my teammates suggested we use $3,583,729.28 for every "no requirements" estimate we are asked for. I seconded the motion. :w00t:
Steph Brown
August 20, 2015 at 7:33 am
Stephanie J Brown (8/18/2015)
Eric M Russell (8/18/2015)
Like I said earlier, if estimating is part guesswork, then the business should be the one providing the guesses. The business should define what is a Sale and then estimate the number of Sales per year. Then the developer can draft a physical data model and estimate storage required.How can a humble IT developer reliably guess how many sales orders will be entered a year from now, if the VP of Sales wouldn't even care to hazzard a guess?
+ 100!
Separately, I'd like to thank Matt for the Dilbert link (had to copy / paste, but it got me there). One of my teammates suggested we use $3,583,729.28 for every "no requirements" estimate we are asked for. I seconded the motion. :w00t:
Sorry Stephanie - I was trying to recreating the link markup by hand (apparently my work browser security prevents using any of the editing shortcuts already provided on the posting screens here. Let me see if I can fix that for ya.
----------------------------------------------------------------------------------
Your lack of planning does not constitute an emergency on my part...unless you're my manager...or a director and above...or a really loud-spoken end-user..All right - what was my emergency again?
August 20, 2015 at 8:40 am
Matt Miller (#4) (8/20/2015)
Stephanie J Brown (8/18/2015)
Eric M Russell (8/18/2015)
Like I said earlier, if estimating is part guesswork, then the business should be the one providing the guesses. The business should define what is a Sale and then estimate the number of Sales per year. Then the developer can draft a physical data model and estimate storage required.How can a humble IT developer reliably guess how many sales orders will be entered a year from now, if the VP of Sales wouldn't even care to hazzard a guess?
+ 100!
Separately, I'd like to thank Matt for the Dilbert link (had to copy / paste, but it got me there). One of my teammates suggested we use $3,583,729.28 for every "no requirements" estimate we are asked for. I seconded the motion. :w00t:
Sorry Stephanie - I was trying to recreating the link markup by hand (apparently my work browser security prevents using any of the editing shortcuts already provided on the posting screens here. Let me see if I can fix that for ya.
Here's the link to the Dilbert cartoon. http://dilbert.com/strip/2009-12-07
Kindest Regards, Rod Connect with me on LinkedIn.
August 20, 2015 at 1:15 pm
Rod at work (8/20/2015)
Matt Miller (#4) (8/20/2015)
Stephanie J Brown (8/18/2015)
Eric M Russell (8/18/2015)
Like I said earlier, if estimating is part guesswork, then the business should be the one providing the guesses. The business should define what is a Sale and then estimate the number of Sales per year. Then the developer can draft a physical data model and estimate storage required.How can a humble IT developer reliably guess how many sales orders will be entered a year from now, if the VP of Sales wouldn't even care to hazzard a guess?
+ 100!
Separately, I'd like to thank Matt for the Dilbert link (had to copy / paste, but it got me there). One of my teammates suggested we use $3,583,729.28 for every "no requirements" estimate we are asked for. I seconded the motion. :w00t:
Sorry Stephanie - I was trying to recreating the link markup by hand (apparently my work browser security prevents using any of the editing shortcuts already provided on the posting screens here. Let me see if I can fix that for ya.
Here's the link to the Dilbert cartoon. http://dilbert.com/strip/2009-12-07
The copy/paste worked, but thank you for fixing it for everyone else. I left that cartoon up all day - it just made me chuckle every time I saw it. We've been getting a lot of requests for SWAGs (silly or super wild anatomical guesses) lately, so this conversation and the cartoon have been SO appropriate!
Let the insanity continue. :crazy: (Humor helps)
Steph Brown
March 7, 2016 at 2:02 am
jay-h (8/11/2015)
...Escher...
My favourite artist since I was a child.
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
March 7, 2016 at 2:18 am
A problem with estimates that I find is that often effort and elapsed time are considered equivalent.
[PM] How long will it take you to do X?
[Me] 5 days.
[PM] OK. Start it tomorrow.
Tomorrow morning first thing.
[PM] Don't start X. Get on Y immediately!!!
[Me] OK.
The next morning.
[PM] How have you got on with Y.
[Me] Done. It took me all day.
[PM] Great. How long until X is done?
[Me] 5 days.
[PM] How come? You told me it would take 5 days and it is two days later!!!
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
March 7, 2016 at 10:04 am
Aw, don't be so hard on them. Not everyone can do math word problems. :hehe: :w00t:
Steph Brown
Viewing 15 posts - 31 through 44 (of 44 total)
You must be logged in to reply to this topic. Login to reply