Published
When I graduated in the early 90s, I worked for a technology firm which was obsessed with data; largely due to the influence of its Japanese owners and their process optimisation thinking from their manufacturing origins.
The analysis of data to improve performance permeated every aspect of work. We never called it data analytics, but quality. “Quality for the individual” was one of the first modules on the graduate training programme. It taught us the importance of objective setting, the measures used to assess performance in relation to objectives, and the need for baselines from which performance would be judged. Every stationery cupboard in our 80+ offices was filled with graph paper for us to track aspects of our work to find ways to continuously improve. The fact that graph paper was easier to get hold of than notepads was just one example of the company’s obsession with data.
When I say the obsession with data permeated every aspect of work, I found there was one exception – projects. We tracked cost and schedule performance and that was about it. I found it puzzling at the time as other aspects of work would have dozens of ‘standard’ measures to track performance. But in projects we basically had just two, and even then, these two measures were not standardised – it was down to the project co-ordinator to decide how best to track them.
When I left that company I discovered ‘data poverty’ on projects was common place, perhaps universal. It has always sat uncomfortably with me as managing projects without data isn’t project management, it’s just winging it. Given the amount of money invested in projects and the high expectations for their successful outcomes, it’s critical that we stop winging it and start using data to guide our decisions.
Consider the case of an infrastructure owner/operator I worked with which had around 600 concurrent projects. They were moving from using four different project management frameworks to just one. The existing frameworks all had a gated lifecycle, but with a different number of gates ranging from 4 to 8. I was asked to recommend the optimum number of stage gates to minimise overhead and maximise performance. It was impossible to work out, based on their historical projects because they hadn’t collected data on process performance for each the frameworks, nor was there a standard measure of project success and hardly any projects had a useable baseline on benefits. We had to rely on modelling and extrapolation to come up with a ‘best guess’ of five. In truth we couldn’t be sure whether five was the right answer, so it was a leap of faith. However, we could at least start to measure process performance alongside project performance from that moment onwards to ensure any future decisions on gates would be data driven and not belief driven.
For the project world to exploit the rapid advances in technology and data analytics we need to move from a position of data poverty to a position of being data rich. Projects are temporary and, for major projects, involve multiple organisations – so data is often siloed or usually dumped after the project completes. We need to take a collective approach to change this. The good news is there is already momentum through the Project Data Analytics Taskforce which developed a manifesto for data-driven projects.
The manifesto, emerging applications and lessons from the world of data analytics are the focus of a panel discussion on 8th March 2023. The mixed panel of practitioners and academics panel will discuss how our use of project data is changing every aspect of delivery and will explore:
- The challenges around the adoption of new technology and data analytics, in particular how to build trust in the technology and encourage people to embrace the opportunities it presents, rather than simply putting it into the too scary/too hard box.
- The barriers and enablers to adoption and how concepts such as perceived ease of use and perceived usefulness could help us overcome some of these challenges
- How to transfer real case project data, which are normally unstructured and noisy, into a workable format that could be used in Machine Learning models. Real-world experience in project data analytics and whether it is feasible to use Machine Learning to predict project performance.
This is a great opportunity for project management professionals to learn about the importance of data in project management and how we can leverage it to drive better outcomes for our organizations.
So mark your calendars and join us for this exciting discussion!
P.S. all parts of this blog are true, but some parts were written by ChatGPT.
Andy Murray, Executive Director,
Major Projects Association