In the previous two screencasts (see links below) about the “Project Portfolio” I described the process of project portfolio prioritization using constraints. In this screencast called “Part C” I’ll introduce the time dimension to the modeling process. We still need to select the best project mix to maximize benefit in order to meet our objectives, but in this next step we’ll consider how resources are used over time. This differs from my first two screencasts in this series where we looked at resources without consideration for when projects start and how they interact over time.
Here’s the same subset of a larger project portfolio which I’ll use to illustrate modeling resources over time. It will take $50M to fund these projects and require these resources.
I’ve transferred the projects from decisionAccelerator into fastProject in order to add a schedule for each. Here are my seven projects that are all scheduled to start simultaneously. I’ve allocated my two engineering resources to each project, following the same distribution of time over the duration of each project. I can always allocate resources at the next level of detail to get even more granularity.
Each project in this macro model has one level of break-down, further work break-down will be done later by my project teams.
Based on resource allocation to each project, this is the distribution of both resources groups over the one year planning window. This shows 150 Design Engineers and 70 Test Engineers. If all projects would start at the beginning of the year then we have a slight problem in the first quarter. We are 20% short of resources.
This histogram shows the usage for each resource individually. Again, this information is based on individual resource assignments to each project.
Lets see if we can level out resources, while still doing all my projects. The green “funded” indicates the order of funding to maximize benefit.
We use this as a “starts-control” tool to better distribute resources over time. I’ve leveled out the projects over time based on the cost-benefit curve and funding priority. The leveling shifts one year of work over two years in order to slow the rate of spending and even out finite resources.
This is what my resource distribution looks like now once the projects are leveled. The top view shows “man-days” of effort, percent allocation, and budgeted cost for each quarter based on the resource allocation to the projects. The bottom view shows percent allocation based on 150 and 70 engineers.
This is the cumulative cost curve or spending profile required to fund the project portfolio. I’ve spent about half the budget in the first year. And this is the non-cumulative cost profile by quarter. Again, this is based on my leveled two year project schedule for the portfolio.
The problem is, that I must release some of these projects within the first year in order to remain competitive. Which projects should we focus on given that I am also facing a 50% budget reduction? As we saw in the previous screencast, Projects C and F are dropped and the benefit drops to 77%.
Lets see how the resources are distributed over time given this new focused list of projects. The new portfolio mix requires less people, 77 Designers and 28 Test Engineers. Clearly, doing all these projects at once is problematic. I lack the resources in the first two quarters to start all the projects at once.
Using the cost-benefit curve, we can stagger these projects over time.
I’ve chosen to start each project in this order over the one year window.
Now my resource problem is pushed to the third quarter, specifically in these months. This is what each individual resource looks like during these problem months.
This is my cost distribution for the year - in other words this is the spending rate needed to fund this project portfolio.
I’ve decided to approve this mix and these start dates for the project portfolio. Later each project team will detail these projects down to the next levels of granularity.
I’m willing to assume some risk knowing that I have a potential resource bottleneck in Q3, but I believe my teams can even this out between Q3 and Q4, since the magnitude of the gap is low. 20-30% over allocation is within acceptable boundaries. What I’m looking for are the large gaps for long periods of time.