Loading ...
Sorry, an error occurred while loading the content.

Re: [XP] Re: Customer doesn't trust our estimates

Expand Messages
  • wecaputo@thoughtworks.com
    ... My experience is that ideal days (not to mention load factor) unnecessarily complicate things. Velocity and our Estimates represent two variables that
    Message 1 of 50 , Apr 1, 2003
    • 0 Attachment
      Kiel:
      >The thing that appealed to me about first estimating ideal time and
      >then adjusting via Velocity is the seperation of concerns it allowed.
      >This seems like it will yield more reliable estimates than trying to
      >lump together our "time spent elsewhere" with how long the task or
      >story should take, etc. Do peoples experiences not confirm this?

      My experience is that ideal days (not to mention load factor) unnecessarily
      complicate things. Velocity and our Estimates represent two variables that
      correct for each other. Simply estimating in relative points (that story's
      a one, this looks about the same, twice as big, half as large etc) provides
      the same result, but without the misleading notion that our estimates by
      themselves say something about schedule. IOW speaking in days, or even
      having points derived from days sets an expectation that we can adjust
      estimates to days.

      Rather, I find totalEstimates / currentVelocity to provide a much better
      prediction tool (which is the point of estimating in the first place.
      Simply keep adjusting estimates as you learn more (well this looked the
      same size, but there is this sub-thingy that needs done, so all of these
      others are two's as well) and let velocity do its thing (no need even for
      averaging, the average is already in there if you adjust the estimates) and
      the plan seems to converge on an accurate measurement in a few weeks. Once
      that happens, you can look for ways to increase velocity, reduce scope or
      extend deadlines, instead of the constant pressure to "fix the estimates".

      Best,
      Bill

      William E. Caputo
      ThoughtWorks, Inc.
      --------
      Cynicism is not realistic and tough. It's unrealistic and kind of cowardly
      because it means you don't have to try. -- Peggy Noonan
    • davechaplin
      Is your customer technical? If so, then ask them to show you how they would do it quicker. Admit that you don t know. You know yourself that you are working
      Message 50 of 50 , Apr 9, 2003
      • 0 Attachment
        Is your customer technical? If so, then ask them to show you how they
        would do it quicker. Admit that you don't know. You know yourself
        that you are working very fast using XP, and that the quality
        investments you are making will pay off. You can gamble on that. Make
        sure your estimates accurate though, since no-one likes a deadline
        that isn't hit. Under promise /commit, and over deliver.

        Don't tell her you are coding 6 hours a day regardless. Tell her you
        are coding 8, then make sure you only code spikes and R&D stuff in
        the other 2 hours. Personally, we stop touching production code after
        about 4pm. Our brains are a bit dead. Then we work on previously
        marked TODO's/STUB's in the code that are low brain activities to fix.

        Regarding the problems with times estimates. Give a range if you are
        unsure. e.g. between 5 and 15 days. Tell her you need to do some R&D
        for 2 days to get an accurate estimate. If she takes the lower one
        that is her problem and you have a project management problem you
        need to solve. Openly refuse to commit to deadlines unless you are
        sure of hitting them. I suspect you had project overuns in the past
        because someone else other than the developers was promising
        something that couldn't be delivered. Seen that before. It was solved
        by the particular project manager being marched out the building by
        security. That solved it!

        Hope that helps.

        Dave.

        --- In extremeprogramming@yahoogroups.com, "rodriguez_jose48"
        <rodriguez_jose48@y...> wrote:
        > Now, after we've managed to have some separation of
        responsibilities
        > between programmers and Customer, our main problem is that the
        > customer does not trust our estimates.
        >
        > She thinks we are giving her too big estimates, and that we are
        > working too slowly. She is right in thinking like this because in
        the
        > past this project didn't went so well, and schedule overruns were
        > a 'normal' thing.
        >
        > What can we do to solve this problem?
        >
        > This has a couple of bad consequences over the process.
        >
        > For example we cannot tell her we are programming only 5-6
        hours/day.
        > I've seen on this list that others have a similar period of
        effective
        > coding each day, so I believe it is a normal thing (anyway we are
        > tired after 5-6 hours of effective coding).
        >
        > Another problem is that we fear to have the Customer with us during
        > the planning. We fear that we will not be able to give an accurate
        > estimate in a short time (planning a story shouldn't take long, I
        > guess) and if we give a shorter estimate, she will not agree to
        > change it in the future.
        >
        > Besides, during the planning game, we, the programmers, have to be
        > careful, not to tell anything about the mapping real time/ideal
        time
        > (we are estimating in ideal days and weeks).
        >
        > Any ideas?
        >
        > Jose.
      Your message has been successfully submitted and would be delivered to recipients shortly.