How to measure your projects more effectively

Andreas Bang Leed gives five top tips for getting measurement right and avoiding the traps that derail most projects 

According to Oxford Global Projects’ database of more than 20,000 projects, only 0.4% of projects meet all three criteria of delivering on budget, on time and with the promised benefits. Less than one in 200 projects meets its own targets. The other 99.6% fall short on at least one dimension. So why do the vast majority fall short? And why, despite decades of new standards, certifications and software solutions, do we see no meaningful improvement in project outcomes across the decades? 

In researching our new book, How to Measure Anything in Project Management (which I co-authored with Douglas W Hubbard and Alexander Budzier), we found that inadequate measurement is at the heart of the problem – not too little measurement. Many organisations are drowning in KPIs and dashboards. The problem is the wrong measurement, applied in the wrong way. 

Here are five practical tips for measuring projects more effectively: 

1. Start with decisions not dashboards 

Most project measurement is what we call ‘discovery-driven’. You collect data because it seems useful, hoping it might inform a decision later, but no specific decision has been identified. Dashboards fill up with metrics we glance at without knowing what action they should trigger. 

The alternative is ‘decision-driven’ measurement. Before measuring anything, ask: what decision could this change? If a metric would not alter any choice you might make, its information value is zero, no matter how impressive the chart looks. 

One project leader on a UK defence project told us: “We measure this KPI because someone from the House of Lords once requested this information. To this day, all our projects measure this KPI.” This is measurement for its own sake, not measurement that drives better outcomes. 

Before adding any new metric, identify the specific decision it would inform. What would you do differently if the number were high versus low? If you cannot answer that question clearly, reconsider whether to measure it at all. 

2. Beware the analysis placebo 

In clinical drug trials, patients sometimes feel better simply because they believe they are being treated. This is the placebo effect. We have found something similar in project management: the ‘analysis placebo’. 

This is the feeling that some analytical method has improved your estimates and decisions, even when it has not. Research across multiple fields, from sports betting to clinical psychology to investment analysis, consistently shows that more information and more collaboration can increase confidence without improving actual accuracy. 

It might seem that experience would guard against being fooled by an analysis placebo. As decision psychologist Paul Schoemaker put it: “Experience is inevitable. Learning is not.”