Autodesk experts share a look at the state of the construction industry, including a few insights you may not expect.
There’s no question that data plays a more important role than ever in the construction industry. And with good reason: the right insights can improve everything from project cost and timeline accuracy to reducing the risk of lawsuits and disputes, and even winning more business.
There’s much to learn from quality construction data, and we recently held a webinar featuring a group of experts and data scientists at Autodesk who revealed some surprising insights on the state of the construction industry, and what might be in store in the years ahead. Below, we recap some of the findings and insight revealed in the virtual session.
“In simple terms, data is unorganized, factual, raw figures,” said Allison Scott, an award-winning strategist, technology translator, and Director, Construction Thought Leadership & Customer Marketing at Autodesk. The magic of data doesn’t lie in one number or individual stat; it’s about what you can do with a larger set of information.
To understand how data insights can inform success outcomes for the construction industry, Scott refers to a framework of Data - Information - Knowledge - Wisdom. Under this framework, you start with a data point. An example of a data point might be the number of overhead mechanical racks installed on the job that day, or logging the root cause of a safety incident. Then, you turn that data into information when you start to make it useful. From here you can apply more context and processes to an increasing number of information sets, which means you can start to connect the dots and identify patterns.
This information eventually translates into knowledge, like figuring out the average time it takes to install an overhead rack based on information collected from multiple jobs. Or implementing new safety training to mitigate worker injury risk You can then use that knowledge to anticipate and better manage risk for future jobs.
Which brings us to wisdom, or the ability to turn knowledge and experience into quality decision making and judgments that allow you to better understand driving principles. As an example, Scott cites identifying that the average installation rate of one trade partner is faster and has less errors than another, which gives you an edge when deciding which company to award a bid to on the next project.
In the COVID-19 era, data is playing even more of a vital role in helping contractors mitigate project risk, Scott stressed.
“The pandemic has upended business as usual, and to ensure the health and safety of people, many contractors have started to use technology and data to better understand potential areas that may put their workforce at risk,” she said, adding that the pandemic has also increased interest in artificial intelligence solutions, which have the benefit of boosting efficiency in translating information into knowledge.
Indeed, according to a recent report, nearly two-thirds of IT decision-makers expect to increase their budgets for automation. In the year ahead, this will not only help monitor construction progress, but it will also accelerate the tracking of more leading indicators for health and safety, quality, and productivity. Overall, the industry is shifting toward more automation fueled by the insights emerging from construction data.
Another recent change is in the way that data gets processed into information and knowledge. For construction, this has traditionally been done through a manual process. However, data visualization tools, like Autodesk’s Dashboards, Reports, Project Home, and machine learning technology like Construction IQ are now doing this for us, organizing, interpreting, and uncovering patterns in a faster and more comprehensive way than before. And it’s not just construction, many other industries are adopting these strategies as well. From music to healthcare, a number of industries are turning to machine learning to analyze their data, making businesses more efficient and productive in the process.
From millions of data points, Autodesk’s data experts were able to uncover some notable insights affecting numerous processes across the lifecycle of a construction project. Led by Pat Keaney, Director of Product Management for Autodesk Construction Cloud Intelligence Products, and Manu Venugopal, who oversees Construction Data and Analytics for Autodesk Construction Cloud, our team of data scientists revealed fresh insights to improve your current risk management strategy.
To uncover this insight, our data experts started with a simple question: can we “reverse engineer” the behaviors of highly successful projects? First, they had to define success. “We used profit margin to distinguish between successful and unsuccessful projects,” Keaney clarified.
Next, the team compared apples to apples. “You don’t want to compare an airport terminal to a hospital project,” added Keaney. “We separated projects by project types, and then we created bands based on project size.”
Once these comparisons were made, the team was able to reveal five surprising patterns:
Our team of data scientists found that projects with more than 6% of construction value in change orders showed margin erosion.
Change orders in construction are not necessarily a bad thing, and are almost an expected part of any job. But it’s when projects hit a tipping point of having too much of their construction value in change orders, this pattern of losses can emerge.
“I think the takeaway here is that you need to expect some change but you also need to invest more time and resources into the upfront, earlier phases of project planning to keep that change in the right zone,” Keaney said. “So the next question we asked is what led up to those changes? We worked backwards to RFIs, which are sometimes early warnings of change.”
In analyzing the RFI behaviors of a number of teams across five different software systems, our data experts found that projects that prioritized closing more critical RFIs faster were more successful.
The devil is in the details. The most successful teams were not closing out all RFIs faster, they were closing out critical RFIs faster. “This made a lot of sense—the best managers intuitively know which RFIs are the most important and they prioritized those,” Keaney said. “They knew which RFIs had dependencies or could cause schedule delays and they got those resolved faster.”
The nuance that this data reveals is that prioritization matters. Many firms are collecting data that can’t necessarily predict project results. What’s worse, measuring this data could drive the wrong behavior. But when it comes to prioritization, how do firms know which RFIs matter most? This brings us to the next insight our team revealed.
When our data scientists examined the common root causes of a group of unsuccessful projects*, we were not surprised to learn that these projects had 50% more RFIs with a root cause of coordination problems than successful projects studied.
“Autodesk has been providing best in class coordination products for years, so we were not surprised to learn this,” Keaney commented. “But, like the industry, we did not previously have this clear data-driven insight to quantify the value and importance of coordination.”
Specifically, now that we know poor coordination erodes project margins, the industry can focus more on solutions to improve and standardize their coordination processes, leading to more project success in the future.
*In this case, unsuccessful projects are classified by profit margins.
One of the most overwhelmingly strong insights our data scientists uncovered was that more than 70% of RFIs could have been resolved in design review.
“I am sure both GCs and design professionals would agree with this,” said Venugopal. “Our industry is on a mission to figure out how to minimize project cost overruns. And what we are realizing is that one way to head off unexpected changes is to have a robust design review process. Today, many GCs that I talk to have teams devoted to reviewing design documents and drawings, and identifying potential problems and solutions in a more collaborative environment.”
To reach this insight, Venugopal’s team decided to measure the impact of the design review process with machine learning models that would automatically tag RFIs with root causes. The findings showed that roughly 70% of RFIs stem from design and documentation errors and omissions.
“A more robust design review gives you the opportunity to identify and mitigate a majority of these problems early, and make sure they don’t reach the field,” Venugopal added. “I think the biggest theme we saw come out of all of this, which is something we hear all the time, is the earlier you catch the potential problem, the cheaper it is to fix.”
When it comes to quality and safety, the primary insight our data scientists gleaned is that standardization is the key to project success.
As just one example, our data team found that projects that used standardized checklists for both their quality and safety programs had a higher rate of success and were able to proactively manage problems vs. reactively managing these issues as they arose.
The team saw the same findings with safety—that standardization in capturing and reporting information is critical to safety programs.
“We see vastly different ways project members log incident information and observations, making it nearly impossible to analyze the true root cause of those safety issues,” Venugopal said. “This is where machine learning has played a major role.”
Venugopal cited safety issues like fall risk and social distancing and PPE compliance as areas where machine learning can run predictions and help improve safety protocols.
“I was talking to a superintendent on a jobsite visit—they had 1,400 open issues,” recalled Venugopal. “Imagine how long it would take for him to read all those issues logged by the team. This is where machine learning tools, such as Autodesk’s Construction IQ, can help identify and prioritize those issues automatically so that teams can focus on the ones that are the most critical.”
Moreover, the data our team studied showed that projects that closed out higher risk issues faster typically had more positive outcomes.
There’s so much that data and machine learning can teach us about how we work and how we can work smarter. According to McKinsey, leaders who apply the learnings from these rapid advancements in analytics and machine learning will be best positioned to tap deeper into the value these tools can unlock, especially in today’s uncertain environment.
If you have not already completed the first step in your innovation strategy by adopting a company-wide initiative to move to digital technology, it’s time to harness the power of data and machine learning to support your project processes.
Do your current processes give you the data you need to capture critical project information, benchmark performance, and maintain quality control across your construction projects?
Take this free assessment to see how your workflows stack up.
