Project Data Analytics 101

Advanced project data science and analytics have the potential to change the face of project management and project delivery forever. But the project management profession is slow to catch on.

I was invited to write an article for the PMI newsletter and I have copied it below. I hope you find it useful. I’ve also copied the text below for those who may wish to refer to it.

PMI Article Page 17

PMI’s recent report on ‘Success in Disruptive Times’, describes the challenges that lie ahead for project managers as the rate of disruption escalates. Machine Learning and Artificial Intelligence (AI) are having a huge impact on other professions, but our profession appears to be lagging behind!

My enthusiasm for project data analytics, isn’t widely shared. When I discuss the subject with fellow practitioners only around 10% of them ‘really get it’ – the reaction of others is somewhere between ambivalent and sceptical.

Many argue that every project is different, so how can a machine help to improve outputs and outcomes? Most have enough challenges on their own ‘to-do’ list without wishing to add more. The PMI Pulse of the Profession report states: “If organisations are struggling with the challenges of today, will they be adequately prepared for the disruptive environment of tomorrow?” Foreword as seen from Pulse of the Profession Report


Five steps to gaining traction and building momentum

1. Develop your data strategy

Start by clearly defining your objectives, and decompose them into a set of user stories, develop the data model to underpin them, and map your data against the model. In some organisations data is collected to ‘feed the corporate machine’, rather than provide value to those creating it; which leads to a degradation in quality and lack of completeness. Some of the data is ‘dark data’, a term used by Professor Whitehorn to describe data that is invisible in an organisation. It may exist in transactional systems but because it hasn’t been used for basic analytics and reporting yet, it hasn’t earnt the trust of the stakeholders.

2. Demonstrate the art of the possible

Sharing good practice, horizon scanning and understanding the inherent challenges in advanced data analytics methods. When our eyes begin to open up to the possibilities that data can offer, then data is seen as a valuable business asset rather than a by-product of project delivery.

3. Develop a data sharing culture

The key to effective analysis is data. Lots of data. If you have a busy pipeline of similar projects then advanced methods are likely to be more successful, than a mere trickle of disparate projects. But if we pool data in a trusted environment, then one company’s ‘one off project’ becomes part of a dataset of (similar) one offs. Patterns in this data enable us to predict future performance and influence decision making.

4. Understand the science

Data analytics is awash with jargon, and it is complicated. Even so, it is important to develop a grasp of the fundamentals. There are plenty of online courses, books and blogs available, but I recommend you absorb the key principles before jumping into the complexities of statistical analysis. Understand your own use cases and begin to get a feel for how emerging methods could satisfy them.

5. Build an active community

The first step is to develop a community of practitioners, bridging traditional stovepipes between project delivery, data science and business analysis. This enables us to; raise awareness, share a common language, cross-fertilise skills and tackle problems. The London Project Data Analytics MeetUp is designed to do this. Join, it’s free.

Applying data analytics to the project business case

Take the example of the business case – recognised as the corner stone of any projects. When the business case is written there are many bold claims of the range of benefits that will be delivered, but…
• How many of these have actually been delivered?
• What is the cost/benefit ratio of each?
• Is there a correlation between management action taken and the magnitude of the benefits delivered?• Which benefits are most difficult to deliver, and which factors have the greatest influence on success?
• Is there an event or milestone after which successful benefits management degrades?
These insights can all be extracted from the data. Simply collecting benefits management data is insufficient; we need to understand the relationships between; benefits, costs, schedule, risk and a range of other parameters, that influence successful outcomes via a network of connected data.


A whiff of organisational change

Some organisations are moving away from traditional spreadsheet-based reporting systems towards PowerBI, Tableau and other analytical tools. This is a great step forward on the maturity journey but for most, only provides a ‘descriptive analytics’ capability; i.e.: the ability to grasp what is happening. A positive outcome of this change is it enables organisations to understand the inherent potential of data, that helps to drive up data quality and completeness. The aspiration for many is to move towards understanding what is likely to happen: predictive analytics and how to influence the outcome: prescriptive analytics.

Whether you are leading or lagging on project data analytics, you need to know that it will change the face of project delivery – forever! Granted it may take several years in your own organisation, to collect sufficient data to reach the promised land of prescriptive analytics – understanding what is likely to happen and how to influence it – so why not begin your own journey now?

PMI Article page 18-19