[consulting] Estimation-Blowout case-studies wanted

Sam Cohen sam at samcohen.com
Sat Feb 21 14:26:26 UTC 2009


On Sat, Feb 21, 2009 at 3:52 AM, Victor Kane <victorkane at gmail.com> wrote:

> Interesting topic!
>
> First of all, this is true of all software projects, not just Drupal!
>

No doubt this is true, but I do think Drupal presents unique challenges that
makes estimating almost impossible  -- that it unless you build in lots of
padding.

To start, if a client wants any non-core functionality, there's the issue of
whether or not a module exists that can do this.  And if there is,  how
close will it get.

It's often not possible or practical to do the research it would take to
determine these things up front, so there's an enormous amount of guess work
involved.

The other killer is the bugs that come up.  I started a very complex site on
Drupal 5 last summer that almost bankrupted me because I did it at a fixed
rate.  The client wanted very long content types with dozens of fields of
all types and views will lots of exposed filters.  I made the mistake of
thinking cck and views would handle their needs, but discovered bug after
bug which added dozens of hours to the project.  I know longer offer fixed
rates on anything.

I'm finally getting a little smarter in how I approach projects.  I spend a
lot more time educating the clients about the process and explaining why
projecting costs is so difficult, and how, as we go, I can present options
on various ways to do this (this way will take an hour, this way might take
2 days, it's your call)

The thing I'm also finding helpful is the working backwards approach, where
we have a fixed budget up front, but I keep the specifics as vague as
possible and do as much as I can within the time they have allocated.

Sam
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.drupal.org/pipermail/consulting/attachments/20090221/0751ed10/attachment.htm 


More information about the consulting mailing list