One evening a few years ago, the manager of an offshore oil and gas platform in the Gulf of Mexico called me unexpectedly. Her new facility was experiencing a dangerous quality issue. Nearly all of its more than 2,000 bolted piping connections were leaking. She was in crisis mode and needed immediate help diagnosing and fixing the problem.
Many cups of coffee later, we figured out what had happened. Workers at the shipyard that built the platform had tightened most of the bolts by hand, instead of using calibrated torque wrenches required by the standard operating procedure. Somehow, the loose bolts were not detected during initial inspections. More than 75% leaked once the systems were pressurized.
Unfortunately, stories like this are not uncommon in modern construction and manufacturing.
In fact, loose bolts have been in the news quite a bit following the Jan. 5 Alaska Airlines door plug incident—and from subsequent quality investigations at manufacturer Boeing Co. and its suppliers. Even the most experienced workers make mistakes, and intense schedule and cost pressures make errors more likely.
The construction industry faces a persistent 30% rework rate on projects, which adds an average of 12% to to total project costs, says a 2023 study published in the journal Quality & Quantity. Rework is also the cause of 39% of all worker injuries, according to BBI Services, a construction industry consultant.
High rework rates contribute to the larger problem of projects being consistently over budget and behind schedule. These are not caused by a lack of knowledge of how to build things.
Shortfall in Worker Quality Data
The industry has proven methods for constructing and maintaining the built environment. Rather, it is the systems for training and managing field workers—providing them the right information when they need it and verifying that work was performed correctly—that have not kept pace with other technology-driven industry improvements.
Many planning, scheduling and turnover processes in construction have been digitized, but fieldwork is still mostly tracked on paper, if at all.
Improving work quality requires changes to both culture and process.
For culture, we can use many of the same strategies employed to improve jobsite safety. Leaders should normalize talking about and learning from quality failures, without fear of retribution. Workers and supervisors who identify quality risks should be rewarded, not shunned. For process, leaders should look to ISO 9000, the “gold standard” for effective quality management systems,
It boils down to three basic tenets: Write down what you do; do what you write down; and make sure you are doing it.
In construction, that means contractors should have written procedures for all critical work processes; train crews on these procedures; ensure a system is set up to confirm the correct procedure was followed for each work activity; and conduct regular audits to confirm all of the above.
Unfortunately, many believe that balancing quality and productivity is a zero-sum game, so quality suffers when time is short. But this does not have to be the case. Quality and productivity can be improved at the same time.
Technology now makes it possible to track fieldwork with the same level of precision available to track Amazon package deliveries. Over the past five years, data platforms have been used to track more than 7 million work completions at the individual worker level. In addition to improved work quality, data has yielded surprising insights for users in boosting workforce productivity.
One user’s data showed that 80% of field work on its projects was performed by 40% of workers, with no material difference in work quality. Some workers had low rework rates but moved slowly. Others were fast but had rework rates well over 30%. With this data, the user was able to improve both quality and productivity by optimizing its workforce to balance efficiency and quality. High performers were rewarded, and underperformers were either corrected or removed from the project.
Capturing detailed workflow data also allows managers to identify process bottlenecks with more specificity. One user who found that work was taking 35% more time than planned on average was able to identify that the culprit, almost always, was the QC step in the workflow.
Workers generally completed tasks on schedule but spent more time than planned waiting for a certified inspector’s work signoff and unable to move to the next task, exacerbating the delay.
These process inefficiencies were not visible in legacy project management tools, but new data helped project managers spotlight an improved QC process—cutting wait times by 60% while maintaining below-average rework rates so projects got back on track and stayed there.
Construction quality and productivity challenges will compound as a generation of veteran tradespeople retire and are replaced by less experienced teams. Inefficiencies such as the above examples are often invisible or obscured but are easily solved once observed.
Collecting real-time worker-level quality data will help upskill the industry workforce, complete projects on schedule and ensure work is done right the first time and every time.
Matthew Kleiman is co-founder & CEO of Cumulus Digital Systems, a Cambridge, Mass., cloud-based platform enabling connected work in construction and manufacturing and author of Work Done Right: Using Systems Thinking to Guide Your Digital Transformation. He can be reached at email@example.com