Advanced analytics with QuantumBlack

Amid a competitive landscape, top performing firms are increasingly distinguished by their ability to harness data. Advanced analytics is one of the most powerful tools available to today’s capital-projects stakeholders—McKinsey research suggests it could produce savings of up to 25 percent—yet many struggle to maximize its potential.

What we do

McKinsey Analytics, which includes QuantumBlack, a specialized data analytics group that uses data, analytics, and design to create competitive advantage, enables clients to achieve transformative, sustainable performance improvements across the infrastructure and capital-projects value chain.

Our interdisciplinary teams combine domain expertise, change management techniques, and analytical horsepower. We gather vast amounts of data from disparate sources—from project-cost trends to email traffic—and analyze it to identify top performance drivers or obstacles to capital-project success.

The impact

• Enhancing real-time monitoring and predictive maintenance. An engineering company needed to consolidate its data gathering processes across more than 250,000 sensors on a £15 billion project. We automated the basic analysis, allowing engineers to focus less on detecting issues and more on interpreting broader patterns, and improved risk management via machine-learning techniques. This added 7 days to their forecasting capability and reduced monitoring costs by 20 percent.
• Increasing productivity. A global engineering firm sought to increase the productivity of its oil and gas division. We quantified the firm’s productivity gap and analyzed teams across six product lines and more than 100 geographies, identifying opportunities to reduce project disruptions. This resulted in a 22 percent increase in overall productivity.
• Predicting network performance. To improve the overall network performance for a leading network service provider, we ingested more than 250 gigabytes of data from sources spanning historical performance to local weather information. By using a combination of key performance indicators and dynamic thresholds to define success, we built a predictive model to forecast network degradation, improving prediction capabilities by 40 times the baseline.