Why organisations need intelligent forecasting

Blue Down Arrow

According to the Oxford Global Projects Database (OGPD), which holds data on nearly 12,000 different projects, almost every complex major project is destined to cost more and be delivered later than planned. Less than one in ten megaprojects are delivered on time and on budget, and full realisation of all project outcomes is rare. The reasons behind such failure rates are themselves often complex, but analysts repeatedly stress the importance of forecasting.

For example, Professor Bent Flyvbjerg, executive chairman of OGPD, says forecasting failures can arise from human biases, blaming ‘uniqueness bias’ (the assumption that nothing similar has ever been attempted before, so lessons can’t be learned from similar megaprojects elsewhere), ‘black swan’ events (rare but devastating disasters – COVID-19 is, of course, a very topical example), and communications breakdowns (early warning signs get overlooked or aren’t dealt with).

The size, uncertainty, complexity, urgency, and institutional structures of megaprojects have variously been identified as failure factors, but academics at University College London have found that no single factor can account for poor performance. Their analysis (see Project Management Journal) identified six themes – decision-making behaviour; strategy, governance and procurement; risk and uncertainty; leadership and capable teams; stakeholder engagement and management; and supply chain integration and coordination – were all of equal importance in analysing why megaprojects seem doomed to fail. The UCL literature review particularly highlighted that forecasting is important in relation to decision-making behaviours. Unfortunately, optimism bias, misrepresentation of the truth, and ‘escalating commitment’ (the perception that, once started, a megaproject is too big to fail and too costly to stop) are common in many complex project organisations.

Delivering ‘more for less’ through intelligent forecasting

As we look forward with some trepidation to a ‘new normal’, post-COVID world in which the major economies will have to recover from deep recessions, the future of some megaprojects will be questioned. Those that do continue may have to deliver ‘more for less’, underlining the need for continuous performance monitoring and measurement. Combining both historic and real-time information sources will be critical to maintaining visibility and control over megaproject outcomes.

A large-scale infrastructure programme can involve 1000s of people from 100s of organisations; teams may make short-lived contributions to very specific aspects, resulting in a constant churn of people, systems and information. However, one thing that remains consistent over the megaproject lifecycle is the creation and accumulation of data – complex programmes are data-producing machines. This data needs to be constantly collated, connected, integrated and analysed.

As programmes evolve, managers need to be able to tap into this data, understand past decisions, test the assumptions made, and then work with the latest data to provide a reliable foundation to assess future risks. Intelligent forecasting builds on historic and real-time data, identifying the information and processes critical to whether programmes succeed or fail.

Lessons from Network Rail

Oakland Group has already shown how intelligent forecasting can help infrastructure organisations. Network Rail (NR) turned to us to help them extract more value from their data. It was no easy undertaking. There was a lot of data, and we had to overcome ‘siloed’ reporting, individuals, departments and regions with different agendas, and inconsistent use of data (client Murray Leach, NR’s head of infrastructure projects systems and supports, talked about our work at Big Data LDN – see Information Age, November 2019).

We also worked with NR to apply machine learning and predictive techniques to their data. Data feeds have been automated; key feeds have been identified for use on different projects (for example: the construction of a new station, electrification, a new tunnel, improving bridges, etc); and we can classify these and differentiate between green- and brown-field sites.

NR has benefited from actionable insights, improved ‘in year’ forecasting, better targeted assurance activities, and improved process compliance. Three years on from the start of our proof of concept, data quality has been improved, and reporting and forecasting is no longer reliant on disconnected Excel and Powerpoint — managers can drill down to the sources of their programme data.

As a major UK infrastructure provider managing capital works costing almost £7 billion per annum (FY 2018-19), NR is responsible for maintaining and expanding capacity on transport infrastructure that is critical to the UK economy. Reliant on UK government financing for much of its activities, NR has to be transparent about its activities in reporting to the regulator, the Office of Rail and Road. Building a robust internal reporting and intelligent forecasting framework has helped NR deliver its programmes of work with greater confidence and reliability.

Previous Post: COVID-19 Sees Cloud Demand SoarNext Post: The Chief Data Officer - Adapting to life in a changed world