Posted by sethlbrown on September 7, 2011 at 10:10pm
Does anyone know of a study or survey on usage of different project management methodologies. In particular, I'm wondering how many Drupal projects use Scrum, or some other Agile variant, and how many use Waterfall. If anyone has a good study, or good survey data, happy to send them a new Lullabot t-shirt!
Comments
Not yet
Hey Seth,
Nope, but I'd be interested in the same results. I asked a similar question at DC London's PM BoF. In our little group consensus was that europe uses waterfall more than we think, even w/ Drupal for various reasons. I got the impression (and this jives w/ past work experience in Paris as well), that it's b/c of the lack of PM culture round these parts.
Again, this is based on little info, but a survey with geographical information would be very interesting.
-Shan
PM Methodology based upon project type/budget
I think that it comes down to the type of work your doing - irrespective of whether or not you're using Drupal as your development platform.
From my experience running my own service firm, as well as working for others, a truly agile approach is really hard to do when working with clients. It takes a really informed client to understand how the agile process works - and a very trusting/patient client to give in to the milestone/budget process involved. For really large client projects, or for product work, it clearly has a lot of benefits.
But let's face it, on a project budget of less than 300-400 hours how many development sprints are you going to have? How much data do you have to come up with a velocity score? As soon as you know your velocity, the project's completed....
We've become big fans of upfront, paid discovery - with a fixed bid for the implementation phase. I hate to say that we're a "waterfall shop", as that term has somehow become derogatory over the last few years, but I guess that's what best describes our work. We organize tickets into "MVP" releases, a backlog, and icebox though. So at least we're sorta cool. ;)
Cheers,
Sean
ThinkShout.com
Totally anecdotal (I don't
Totally anecdotal (I don't know of any studies) but though there are a lot of people who seem interested in agile, seems like most shops still do waterfall style. I've run some agile PM bofs at cons and camps before, and that's generally the mix - probably 15% or so actually do agile and the rest just interested in it and how it could work.
Everyone who currently does agile seems really to have a good experience with it. There are some naysayers who've tried it in the past and had bad results, of course. Had some interesting comments on an old post here http://affinitybridge.com/blog/managing-budgets-and-billing-while-practi...
Those who tend to actually do agile (we do at Affinity Bridge) seem to mostly be mid-sized (ie. 10-20 person) fairly hardcore development-oriented companies, as in heavy on the technical side, lots of complex development, longer/ongoing and larger budget projects, often with more contributions back to Drupal/contrib.
Agile, but...
Hi,
I agree with seanberto sentiments where it really depends on the client, not the agency. As a PM I plan the phases of the project waterfall style, helps get it all into place visually with the world of Gantt charts - clients don't need any explanation here. This then gives you milestones for various deliverables for the client (eg wireframes, design concepts, beta site etc).
But then it turns internal. With many different projects underway we run the development in Agile across the development teams. We have have recently ditched strict scrum and moved to Kanban for the simple reason that clients want features, they don't want burndown charts of points accepted through a sprint. Using Kanban has been really good to focus the devs on delivering completed features to the client for demo, rather (in our case) two weekly sprints of not quite completing things in reality because of the diverse day to day demands an agency has.
In Kanban we are using Releases to tie back up in with the intial waterfall presentation to the client.
So far, so good.
Cheers
Andy
..in the UK
Nice, Andy
You've got me thinking with respect to internally managing projects with Kanban - as I think that it could line up really nicely with Features-driven Drupal development. Thinking about features as Kanban "cards", you could also start to chalking out rough internal time estimates for Features that are simliar across projects (ie, an events-management feature implementation generally takes us 7 days, customizing our Blog feature for a particular client takes 4 days, etc.)
Our little team does also try to make as many upstream Drupal contributions as we can - specifically by getting our clients to work together to collectively fund releases of the modules we support. Kanban could facilitate that approach too, as it seems to provide a good balance between flexibility and accountability.
I appreciate the post.
Cheers,
Sean
Finding Good Statistics
I'd be interested in seeing the data from a survey for project management methodology usage and effectiveness as well. I did a smaller survey on Drupal project success factors about two years ago that my co-worker Matt followed up and improved for his masters paper, Stakeholder Perceptions of Drupal Project Success. Neither of these are quite what you are looking for but I thought that having some (possibly irrelevant) statistics might get us close to a free t-shirt. :)
I think that the tricky part of accurately measuring which project management methodology is most effective would be controlling for developer and stakeholder Drupal experience, the size of the project, and the technical feasibility of the project.
My guess would be that if you found a way to control for those variables you would find that methodologies with faster iterations are likely to finish on-time more often. When undertaking Drupal projects, it's so easy to count on X module to meet a need and then discover weeks later that it doesn't quite behave as expected.
In terms of just popularity, my guess would be the same as arianek's that the mid-size and larger shops are the ones not doing purely waterfall development. It's hard to do scrum when your team consists of one person (not that it can't be done).
Julia