J. Colin Cagney, a director at KPMG Major Projects Advisory, knows that while most companies want to use data analytics to increase business, they’re often not sure where to start. When Cagney and his team work with a client, the first goal is to find a quick win for the company, something that leverages data they have already gathered.
Recently, he has worked with a Top 50 Contractor with $2 billion in revenue. The company sought significant growth. KPMG’s first move was to dive into the company’s data, usually housed on separate platforms. Even with data spread across platforms, there are tools that use the existing data to produce actionable information.
Even with data spread across platforms, there are tools that can use existing data to produce actionable information.
Cagney doesn’t hide his enthusiasm for this kind of thing. “I'll give you a killer example of something we were able to help them with—we integrated all of their safety data across at least 6 different data sets.” Data analytics enabled the creation of a “safety predictive model” that includes weather data and predicts a high or low risk of a safety incident in the next 3 days, based on the type of work.
People often don’t realize how powerful the data they already collect can be.
“It predicted incidents with 89% accuracy—I mean it kind of blew everyone’s minds.” He adds that the incidents that were missed altogether amounted to less than 1% of the errors.
Part of the planning process involves helping a company identify the value of its various parts. “We went through and worked with them to identify value across safety, scheduling, productivity—all the key areas that are most critical.” Companies can only gain from optimal scheduling, better estimating data, and better safety prediction.
Part of the planning process involves helping a company identify the value of its various parts.
Getting the quick win with existing data helps avert skepticism and fear of data analytics as something that might only work for companies with the most up-to-date data collection. “People assume that their data is too low quality, which is sometimes the case but not always. I think you have to spend time and effort to get to the bottom of that and truly find out.”
It helps if you have someone trained in analyzing data to take this on. “Unless you truly investigate the data and figure out why it's not clean, it's hard to get it clean in the future,” he says. “Usually it all comes down to human error–somebody not entering it in right or not following the data governance.” In addition, new tools can automate data clean up and and prepare data for current analytic tools.
Before beginning a data road map, Cagney seeks input from key players such as project engineers, project managers, and superintendents—the people who are impacted by changes in everyday operations.
“You have to dig down deep in an organization to enact change,” he says. With buy-in from people whose lives could be easier as a result of effective data analytics, real change can happen across the isolated parts of a company. “People who are involved in the day-to-day business really have to buy in.”
“You have to dig down deep in an organization to enact change,” he says.
Once a company has cleaned up its data and sees how it can be used to improve processes, what happens next is all about priorities. What is the first project that will receive the most effort? Cagney says a company might, for example, decide to use data analytics in the first phase of its plan to improve its performance in safety, scheduling, and estimating. It might target work acquisition in the second phase. “You plot out each of these key areas.”
In many cases, companies that build a data governance plan to set the stage for further data analytics enablement. When data is more predictably uniform, it's easier to use the right tools to deliver actionable information.
Cagney doesn’t recommend replacing all data collection systems as a company might risk losing data in the process. Existing data from different systems can offer a complete view. “We've found the most impactful analysis is when you're integrate datasets together.” He advises a clear plan for doing this in a multi-system environment. “There’s very little risk if you’re doing something on top of your core systems. You're really just building on top of existing infrastructure.”
As to the question of whether companies should pursue a data road map now or sometime in the future? “I think it's definitely a bigger risk to not head in this direction,” Cagney says.
If you liked this article, here are a few eBooks you may enjoy: