By {{Article.AuthorName}} | {{Article.PublicationDate.slice(6, -2) | date:'EEEE, MMMM d, y'}}
{{TotalFavorites}} Favorite{{TotalFavorites>1? 's' : ''}}

Major capital projects are unique and incredibly complex undertakings. However, when it comes to bringing together the most critical element of a project’s success—the project team—too much often is left to chance or intuition. With machine learning and a small data-driven approach, teams can be assembled using an algorithmic analysis that uncovers hidden meaning in project data. The result: dramatically improved project cost, schedule and safety performance.

What is machine learning?

Machine learning, a branch of the field of artificial intelligence, does not have a single, convenient definition. Arthur Samuel, a pioneer in AI, offered the following informal definition in 1959: "[machine learning is] the field of study that gives computers the ability to learn without being explicitly programmed.”

The field of machine learning is broad and deep. It ranges from small to large, from simple to complex. Today, modern processors, distributive computational resources, reams of rich data and a little know-how have made machine learning a convenient tool that generates insights in real time.

Lowering costs with small data

Project management systems generate a lot of data. It’s not big data, it’s small data, but there is a lot of it and it’s valuable. Small data tends to be well structured and packed with meaning; data from project controls, finance, contract and asset management, and other information systems can be combined to provide a rich data set. As data sources proliferate, individual data points become increasingly available and connected. Extracting meaning out of complex and related webs of data is precisely where machine learning excels.

Up to 30 percent of project cost is consumed in the interfaces between companies and individuals. By leveraging small data, organizations have the opportunity to significantly lower this percentage with fewer errors, omissions and rework in design and construction. Better planning and coordination of activities also reduces schedule, cost and safety incidents.

Thinking ahead, more successful projects lead to higher employee satisfaction and talent retention which, in turn, leads to improved project performance across a portfolio of projects and higher financial returns.

Assembling the project team

Project teams are a unique collection of skilled resources. For a single project they are brought together from disparate companies. Many have never worked together before. Historically, this is done somewhat arbitrarily based on what skills are required, availability and “gut feel” on who is good and who is not.

Using small data to inform important team structure decisions can have a substantial impact on project outcome. The composition of the project team should be based on many factors such as historical performance, personality traits and personal values.

Few items are too small when it comes to machine learning. Data that is not captured in existing enterprise systems such as financial, procurement, project controls and document control could be immensely powerful. Emails, texts, chats and meeting recordings all contain data that reflect how well project teams are performing and where potential or actual problems exist.

Transcending the status quo

Change may not be easy, but it’s necessary. Think “Moneyball.” Architecture, engineering and construction management is dominated by a very conservative demographic—not unlike the professional baseball managers of old who knew what a good player looked like based on their extensive experience in the game. The Oakland A’s, Billy Bean and his data scientist proved how wrong merely acting on a hunch could be. Now, most MLB teams have analytics departments with dozens of data scientists that provide management data to evaluate their teams and prospects. 

There is untapped potential in data. Eventually, data capture and analytics from machine learning for small data will become a historical model that enables a company to refine and correlate performance with team structure. This will prove a competitive advantage for owners to take a more data-driven approach (versus low bidder) when building their project teams. For contractors, using small data will not only lead to more project wins, but also a more stable project pipeline aligned to available resources.

As companies begin to leverage small data to improve forecasting and resource management for major capital projects, it is important to remember that projects are made successful by their teams. Applying the same rigor to designing those teams as designing the assets they build is a low-cost, high-value technique for delivering better projects.


 Comments ({{Comments.length}})

  • {{comment.Name}}


    {{comment.DateCreated.slice(6, -2) | date: 'MMM d, y h:mm:ss a'}}

Leave a comment

Required! Not valid email!