January 2, 2009

Alignment of project dimensions: A balancing act

When writing the promised piece on agile methods and management information, I found that I wanted to present some thoughts about alignment of different dimensions of projects to avoid having to make the other article longer than necessary. Thus this post first.

I have found it useful to think about the alignment of some project dimensions before signing the contract and kicking off the project. The dimensions I will consider here are contractual terms (pricing), collaboration and strength of relationship, and the software development method to be used. Your selection of variables may vary, but I insist that trying to take such a bird’s eye view is useful when assessing the risk of project failure.

Below (click to see it in full size) is four actual projects placed in these dimensions. They are all software development projects, but differ somewhat in terms of complexity of integrations, amount of front end vs. back end programming, etc. as well as size. So they are by no means a perfect academic sample set, but they are projects that I know because I have worked and lived them.

As you might guess from the topic header, my hypothesis is that some combinations of the dimensions fit better together than others. Which of the projects above would you be afraid of taking on as project manager or team member? Which did you think get into trouble?

In retrospect, one of the possible “doomed to fail” projects actually didn’t (number 4). I try to cover the reason why in my next post. The other prime suspect, number 3, did actually fail spectacularly in terms of profitability. In fact, I took over this project midway (that is when just about 110% of the budget was used…) and even without drawing out the diagrams, the missing alignment was striking. However, the end product developed was to the customer’s satisfaction and new projects will be conducted that hopefully will be better business as well.

Another way of viewing the above dimensions could be in a 2x2 matrix (I am currently a consultant, after all!). Below I have tried to do that, with development method on the vertical axis and contract/pricing horizontally. The color (blue=”colder” customer/supplier relationship, red=strong partnership) of the dot indicates the form of collaboration, to avoid fiddling with three dimensions.

In danger of over-simplifying, I’d like to make the following remarks about these quadrants:

  • The bottom left quadrant is just fine, although maybe a little bit old-fashioned. You could consider discussing with the client if a more flexible approach could be more suitable (moving up and right in the chart) and delivering more value.
  • You should be extremely careful if you are in the upper left quadrant. You must then be absolutely sure that you and your client have found a good way to manage scope adjustments. I would also never conduct a project in this space if it is not a strong relationship with the client, i.e. a "red dot" above. Without these two in place, this is the “doomed to fail”-quadrant.
  • The upper right quadrant if the place you probably want to be operating most of the time with agile methods. Especially with a strong relationship with your client (red dot). Be sure to manage expectations both upfront and at every increment, though. Often it doesn’t help to have the legal contract on your side if your most important client is not close to satisfied after all funds are burnt doing iterative improvements.
  • The lower right quadrant is probably just fine, but it indicates that although there is client/supplier trust, indicated by the choice of time&material pricing, there is actually no belief in the gains promised by agile development methods. Thus, one gets the formal overhead without really needing it to handle pricing and legal issues. But formal specifications can absolutely be a way to handle the expectations management mentioned in the previous bullet point. Some target price regimes might fall into this quadrant, but not all (for example, project 4 was a target price regime).

Another related way of looking at this, is through the classical project management pyramid – TRQ, time, resources and quality/scope. The basic idea is that you cannot change one without adjusting the other accordingly. So if you fixate time and cost, certainly the details of scope can be a variable throughout the project. This view fits quite nicely with what DSDM preaches and on the development level this is the way I find working with Scrum-like approaches most fruitful.

Of course, my comments about the sample projects at this point are with the power of perfect hindsight. But let’s look again at project number 3, where we obviously tried to do flexible development with slim specifications with a completely new customer (collaboration form is not always a function of length of the relationship, but the two are usually strongly correlated and in this case it was a new customer) at a fixed price as if it was a classical waterfall approach we were using. Was it very likely to succeed if judged upfront? I think not.

Please do not read my comments only as a way for vendors to maximize their profits and minimize their risks. The assessments of the project dimensions should rather be discussed openly with the client before kicking off the project. It is after really understanding your customer’s need you will be able to make the best trade-offs for that particular case. Project management methods, and legal contracts for that matter, has large elements of tailoring after all.

There is a lot of fuzz about the specifics of each agile method proposed. Maybe too much. To me, agile software development is actually just as much about mindset and approach as it is about specific methods. And I do think that one should make sure that the mindset and the more formal frameworks (contract, form of collaboration and actual development methods being the three sampled here) match and discuss this upfront between customer and supplier.

In an earlier draft, I considered using project size as one of the variables. This could easily have been included for example through the size of the dots in the diagram notation. I considered about 10 projects I had managed myself and a few others that I know fairly well and I did not really see size (ranging from a few hundred to many thousand hours) being as important as the variables I have proposed above when considering if agile methods will be suited. But you might beg to differ? Let me know what you think by leaving any comments you have below. I hope to learn a lot from them.

No comments:

Post a Comment