Construction’s Data Swamps
In recent weeks, we’ve been diving deep into the world of construction technology at several industry conferences. Unsurprisingly, the buzzword on everyone’s lips is AI—artificial intelligence for everything, everywhere.
The backdrop? A flood of data, more massive than we can truly comprehend. By 2025, global data volume is expected to skyrocket to 175 zettabytes, according to Autodesk and Deloitte Access Economics. That's enough to fill over 175 billion iPhones. But the real question is: with this tidal wave of information, are we equipped to channel it effectively? The construction industry, particularly, stands at a crucial crossroads. While 65% of businesses report using generative AI regularly—double the rate just ten months ago, per McKinsey's latest survey—construction companies face unique challenges in using this data to fuel a competitive edge.
The potential here is immense. Construction sites generate invaluable data at every turn: material deliveries, worker attendance, weather conditions, project timelines, costs—the list goes on. And yet, much of this data sits dormant. Deloitte's research suggests that the construction sector is still in the early stages of data analytics. FMI Corp.’s 2018 findings indicated that a staggering 95.5% of daily data produced by the industry goes unused, a statistic that hasn’t improved much with time.
Why? Because construction data is notoriously siloed. Information is often isolated by site, project, or business unit, limiting its potential to drive broader business insights. The result? An industry eager to become "data-driven" but largely unable to act on it. FMI estimated that while 74% of construction firms aim to harness data, only 29% successfully translate it into actionable insights. Adding to this complexity, the past few years have seen a proliferation of hyper-specialized tools designed for niche tasks. While useful in isolation, these tools often fragment data further, creating technology fatigue and bottlenecks in overall workflows.
So, why hasn’t AI transformed the construction landscape already? The answer lies in the nature of the data itself. Often, aggregated construction data lakes lack context, organization, and governance - murky “data swamps” - that stifle meaningful insights. A 2020 FMI and Autodesk survey estimated that poor data management cost the construction industry a staggering $1.85 trillion, with 14% of rework linked to bad data.
If we’re to unlock the value buried in these data swamps, we need a structured approach to data management. The recent McKinsey survey emphasizes that generative AI, while promising, accounts for only about 15% of a successful AI solution. The other 85%? Ensuring robust data frameworks, skilled personnel, and secure, organized data streams.
The journey to data enlightenment begins with unifying data sources—a lake where all streams merge. While APIs can bridge various systems, they’re not a substitute for a central, consistent source of truth. With a clean, secure data lake in place, organizations can create “data warehouses” for advanced processing. Here, AI and machine learning (ML) can finally shine, uncovering insights that drive productivity gains across complex, interconnected construction sites.
In many ways, the construction industry today resembles the health sector pre-ACA and pre-Electronic Medical Records (EMRs). EMRs, though not mandated by the ACA, became essential to streamline reporting, coordinate care, and shift to value-based models. This aggregation transformed healthcare, driving an explosion in medical technology innovation. Construction stands on a similar brink: once data is unified and accessible, the innovations could be transformative.
Autodesk’s projections reinforce this. Data leaders in construction can expect a 50% boost in profit growth compared to beginners. This shift represents Contech 2.0—not just collecting data but strategically organizing and analyzing it. The road is challenging, but the opportunity is profound.