Project management and the scientific method
So many fervent discussions on different project management styles baffle me…
I was raised to be a scientist. One of my earliest memories is being about 5-6 yr old and mom sitting me down in front on a set of 6 test tubes and a lab book where she had written numbers on the top of each page. Explained to me the scientific process, how to test w the 5 senses (told me that taste was okay THIS time but not normally) She also gave me Ph indicator strips cause they are fun to play with and easy to practice measuring and recording. She then left me to figure out what the 6 substances were and keep track in my lab book.
Hypothesis, testing, record data, evaluate hypothesis, repeat
Really this isn’t very different from most project management…
After all aren’t most creations in their own way experiments?
You have an idea, you create it, you test it, you evaluate how it did and you evolve…
All the other parts to be are just frosting… Agile means you fail faster; Waterfall is when you can’t afford to fail. So deciding which time frame to use just means do some risk evaluations beforehand.
I find it interesting that people get so religious on these discussions. For me, if you can do real risk analysis because any failure cost is high then I spend more on design time. If it is a get it done situation where expectations aren’t so rigid esp in the case of experimentation because it hasn’t been done before, then less design and more experimentation is the order of the day.
I guess in the true spirit of Feynman – I just believe in constantly challenging all assumptions and designs 
September 28th, 2010 at 8:18 am
“I just believe in constantly challenging all assumptions and designs.” Yes! I love how you think
I’m a big Feynman fan, too.
Thanks for sharing the story of your science memories, my children 4 & 6 have come to call dinner time, Dinner Science Theater. As a matter of fact, we did a pH experiment just a couple of weeks ago. Some kooks at the fair were giving away “miracle cure” alkaline water…anyways long story short I was able to teach the kids about questioning authority and how to assess things for themselves with a simple pH test. The water was the same pH as our tap water…go figure. Whenever anyone makes an amazing claim the first words out of my kids’ mouths are “Let’s test it Dad!”
BTW-Glad you are posting again
September 28th, 2010 at 9:57 am
If waterfall is for when you can’t afford to fail, then I see this as saving all your hypotheses until the deliverable, when they are all tested. In its purest form, this seems to go against the idea of “constantly challenging all assumptions”. Of course, in reality, “waterfall” projects tend to test hypotheses early on, during the design stage, as well as during development through testing. A big difference between waterfall and agile is in how these hypotheses are tested – agile tests by showing to a client, and waterfall tests against requirements, assuming the requirements are correct.
If it’s a question of which is more scientific, I agree that it depends on the task. For example, if the project is building a mars rover, the laws of physics won’t change between the requirements gathering phase and the deployment, so waterfall makes sense, as the requirements have to be sound. But for something more subjective, like web design, we just don’t have the formal language to gather sound requirements, so requirements are often stated in terms of client needs + objectives. So for this type of work, agile is more scientific, as it tests that the needs were correctly understood.
September 28th, 2010 at 12:06 pm
The problem with management “science” is that markets, schedules, workers, customers — in short, *people* — just aren’t as reproducible as pH readings: you can’t ever do that critical “repeat” step meaningfully or comparably. The problem afflicts all the social sciences, but since management science is overwhelmingly done by amateurs (amateurs in the “science” part), it’s especially problematic.
One result seems to be that most debates over management styles aren’t about efficacy or effects at all, but rather are proxies for something else, something possibly unrecognized even by the debaters. Sometimes it’s “leave me alone, let me do my own thing.” Sometimes it’s “I don’t trust you.” Sometimes other things, often things that wouldn’t be made explicit even if they were conscious. And that’s why the debates can get so heated: no experiments, no evidence, no reproducibility, no external verification, and lots of hidden agenda.