0 comment

A new project management measurement: the role landscape

I may have just stumbled across a new way of looking at project/resource information: the role landscape. When you need to see a single role’s workload over the lifetime of a project, this is the model to reach for.

As a resources manager and project manager (among other hats), I spend my working life staring at MS Project task lists and Gantt charts, and other project management tools and documents, such as PRINCE2 boxes-and-arrows charts, time-tracking reports, QA logs, risk logs, and time/cost budgeting spreadsheets. I need a lie-down just looking at that list.

There are loads of project management and time-tracking tools and online applications out there (Basecamp and Harvest being favourites), but not much available for resource management, to solve issues like:

  • Managing the distribution of several team members’ workloads over several projects
  • Seeing what team member has any free time in any given week

Yes, you can use the resources stuff with MS Project, but it’s hard to get a sense of how different roles’ workloads relate to each other over time, rather than just as a single snapshot. I’ve yet to find an elegant visualisation of project data and resource data.

Enter the Role Landscape (using capitals for effect – impressed?).

What’s a role landscape?

The role landscape is a model that plots the amount of time spent per day on a project, over the lifetime of the project, per role (see below). You can also chunk the duration into the project phases. The rough example below shows a column chart of daily time-blocks for a Business Analyst role over six (generalised) phases of a web application development project.

A rough example of a role landscape as a column chart

Now stay with me on this: using either the column chart or line chart, you can easily visualise areas of high workload versus low/ambient workload. The peaks indicate when that role is flat out; the flat areas are where you hope they’re involved in other work at those times. By changing the columns/lines to a curved line (the example below again uses a Business Analyst as an example role), you get a true sense of a ‘landscape’ appearing.

A second example of a role landscape, using a curved line plot

Taking it further

This model becomes more useful when several roles are plotted at once (below). At a glance you can see workload relationships areas of high activity across multiple roles (or team members for that matter). When there’s combined peaks (or overlaps) like this, you know that there are going to be more meetings, more emails and phone calls, and more versioning of project assets (requirements documents, reports, code, you name it). Seeing these patterns allows you to anticipate — and plan for — these times of increased activity and stress in a project.

A third example of a role landscape, showing an overaly of three roles: Business Analyst, Senior Developer and Senior Designer

A few notes of the rough example above:

  • The Senior Developer gets no down-time for a long time; if this always happens, then this should ring some resourcing/team management alarm bells
  • The multiple green peaks of the Senior Designer landscape within the Senior Developer landscape represent the two working together to solve design/development issues; this can either be a regular structured scheduled scrum-like interaction, or ad-hoc
  • I threw in the last peak in the Launch phase because there’s always some mad scramble right before the big green ‘go’ button is pressed
  • Of course the designer and developer are going to be working more than 8 hours a day ;)

Rolling your own role landscapes

So how can you derive this sort of data to make your own role landscapes? Any time-tracking application worth its name should be able to give you the amount of time per day per resource; it’s then just a matter of importing it to Excel (or similar) and charting it. For online solutions, there’s a lot out there already for coding charts in X/HTML and CSS, as well as using Google Charts API.