What should project management cost?
We've been looking for the right benchmark for the cost benefit of project management for decades. No-one doubts the value of project management. But we all want confidence when assessing its cost.
All our clients wonder what a project manager's contribution should represent as a percentage of the project budget. If they've invested in a PMO, they want a reasonable market comparison against which to gauge the value it offers.
But finding a measure that's simple enough to remember, and which accounts for the multiple, conflicting considerations involved, isn't easy. Separating the cost of all project management (the PMO, the administrator, the reporting and decision process, the programme and portfolio integration costs etc.) from the simple salary line item for the project manager's role, is the start.
Then there's the complexity of the project itself, which tends to drive cost too. Finally, how do you account for differing levels of maturity between PM environments, to ensure your benchmark acts as a meaningful comparison across the board?
As an organisation specialising in estimation services for project management, this question has been front of mind for a while. We've referred always to the standard available texts, which anchor a cost range at the 15% mark, but now our own maturity and business network has enabled a fresh, more confident opinion through concerted work.
We've consolidated the learning and experience of our estimation and consulting and research teams and matched that to our global partnerships and associations, connecting the database and tool experts we're privileged to work with ... and think we have an answer worth publishing.
Working with a range of blue chip, mostly financial service clients, with a combined spend of billions on software development projects in different environments and types of project, we stretched the orthodox range both ways, to between 5% and 25%. As estimation specialists, we started digging.
By accessing long-term data using evidence from some 19 000 projects from across the world, collated in the knowledge base of an internationally recognised estimation toolset, and by refining our query through tighter specification of project type (we stuck to software development), project size (we analysed a range using function point thresholds of 100, 500, 1 500 and 2 500) and, finally, maturity of environment (applying the globally accepted CMMI strata), we narrowed the range to a significantly revised 5% to 10%.
By cross-referencing these results, moreover, with input from another close tool partner, SEER by Galorath [http://www.galorath.com], we refined a slightly higher range between 8% and 11%. Cross-matching results arrived at a benchmark 9.1%.
We should be clear, the meaning of the benchmark shifts within your own commercial context. You need to know the major contextual drivers pulling either way are important. Perhaps it's where the real value in the benchmark lies.
Research from Gartner, Forrester and The Standish Group puts optimal involvement for a project manager in 2.5 projects at any time, provided only one project is in Execution and the others are either in Feasibility or Closure phases.
As we all know, size matters. Using labour costs as our measure, we know that projects with labour costs less than R5m have high levels of success (71% succeed), while projects with labour costs between R5m and R25m have reduced levels of success (38% succeed) and projects with labour costs greater than R25m have minimal success. Obviously, labour cost is a reflection of complexity.
The implication is that projects with labour costs between R5m and R25m should have dedicated project managers, while projects greater than R25m require more than one project manager or a programme structure with sub project managers.
And this is where the maturity of your PMO environment kicks in. Because 9.1% is defined by what 100% is. It turns out PMO maturity can drive your costs down by threefold between only two CMMI levels.
We're quite proud of this breakthrough. It's been a long time reaching this level of analytical and cross-context certainty, to be able answer that standard question, "How much should my PM cost?" Software development as an industry has spawned software tools to measure itself. It's time to use them as PMO standards.
Interrogating your plans and programmes (and portfolios) scientifically, using the ever-more refined services and tools on offer tends to answer more than just the simple questions about costs, trajectories, integration and efficiency. Sometimes it can answer question we didn't know we'd asked.