A new Revisto report surveyed 2,006 AEC professionals across eight markets and found that 96% of projects still overrun on cost – with 92% reporting overruns of 6% or more. But the real story isn’t a technology failure. It’s a pricing problem that starts years before anyone breaks ground. Budgets are set too early, estimates are too optimistic, and the contracting model practically guarantees the numbers won’t hold. More BIM isn’t going to fix that.
There’s a number that gets thrown around a lot in ConTech circles right now: 96%. As in, 96% of construction projects still overrun on cost, despite years of digital investment across the industry. Revisto put it front and center in their latest report, surveying 2,006 architects, engineers, general contractors, and subcontractors across the US, UK, Germany, France, Switzerland, Australia, Saudi Arabia, and the UAE.
On the surface, it reads like a damning verdict on the industry’s digital transformation efforts. All those Procore licenses, all that BIM coordination, all those dashboards and data flows – and nine out of ten projects are still blowing the budget. The implication from some quarters is that the industry needs more technology, faster adoption, better collaboration tools.
That reading is, at best, incomplete. At worst, it’s being used to sell software to people who’ve already bought the wrong solution to the wrong problem. The budget problem in construction is much older and more structural than any software gap. It starts with how projects are priced – and it won’t end until that changes.
Budget overruns in construction are largely a pricing and contracting problem, not a technology gap. Projects are estimated years before they start, using budgets that bear little relationship to what work actually costs by the time shovels hit ground. Add change orders, litigation strategy, and the incentive to underbid, and you have a system that is structurally designed to go over.
The Stat That Needs Interrogating
Who benefits when 96% looks like a technology problem?
The first question worth asking about any industry report is a simple one: what does the company publishing it sell? Revisto makes collaboration software for design teams. That’s not a disqualifying fact – it’s just context. And that context matters when their headline stat implies that poor collaboration is behind most budget failures in construction.
Dustin DeVan, who built and sold BuildingConnected for $275 million and now runs Ediphi (an estimating platform), spotted the problem immediately: nobody in the report actually defines which budget they’re talking about. Over the course of a typical project, there are many. An owner and architect might agree on a $100 million concept budget. A GC then prices it at $200 million. Did the project “overrun”? Or did it just get priced properly for the first time?
“Perhaps we need better ways to price work, because it might just be that’s what it costs – rather than the design was broken.”Dustin DeVan, CEO of Ediphi, on the Bricks & Bytes podcast
The distinction matters enormously. Attributing a budget gap to design coordination failures, when the real problem is that the initial number was never grounded in market pricing, sends the entire industry chasing the wrong fix. Better model coordination won’t close a gap that was baked in at the concept stage.
How Budgets Go Wrong Before a Shovel Hits the Ground
The timing problem that no software solves
On public projects especially, the sequence is almost perfectly designed for overruns. A budget gets set by a cost consultant – often someone without direct pricing experience in the specific trade packages involved. That number gets approved, sometimes by a government body. Then the project sits in planning, permitting, and procurement for two or three years. When it finally goes to tender, the market has moved, material costs have shifted, and the original number has aged out of relevance. The price balloon isn’t a surprise. It’s math.
Owen Drury, who spent 12 years estimating construction projects before founding Bricks & Bytes, made a point on the podcast that most people in the industry already know but rarely say out loud: if you estimate a project correctly, you probably won’t win it. There’s almost always a competitor willing to price lower – whether because they’re more aggressive, less experienced, or because they’re pricing in day-one change orders as part of their commercial strategy. Litigation-funded claims on construction projects aren’t rare. They’re a recognized business model for some firms.
- Concept budgets set too early. Initial numbers are often agreed between owners and architects before any builder has priced the work. The gap between those figures and GC pricing can be enormous.
- Multi-year gaps between budget and build. On public projects especially, a budget approved today may not be tested by the market for two or three years. By then, costs have moved.
- Underbidding as strategy. Pricing low to win, then recovering margin through change orders and claims, is a known pattern. The “budget overrun” is sometimes the plan.
- Change orders will happen. Clients change things. They see the building going up and want something different. That cost doesn’t belong in the original budget – but it shows up in the overrun numbers all the same.
The Public vs. Private Divide
Not all overruns come from the same place
Private projects and public projects have very different overrun profiles, and lumping them together in a single stat doesn’t tell you much. Private developers tend to have more direct exposure to cost risk, which creates stronger incentives to get the budget right from the start. A developer who goes over by 20% on a residential scheme feels it immediately in returns. That feedback loop tightens behavior.
Public projects operate under a completely different set of pressures. Budgets are political, procurement is bureaucratic, and the people setting the initial number are rarely the same people who have to live with the consequences if it’s wrong. Add to that the design-bid-build contracting model – where builders are excluded from the design process and then handed a set of drawings to price with no opportunity to flag constructability or cost issues early – and overruns start to feel less like failures and more like features of the system.
Martin, the structural engineer on the Bricks & Bytes podcast, made a useful point here. A project with detailed 3D models and complete documentation can still overrun badly if the contractor misses sequencing complexity or makes wrong assumptions about what the drawings actually require. Conversely, a contractor with 30 years of experience pricing similar work off 2D drawings might hit the number almost perfectly. The technology is a leveling tool – it can bring more people up to a baseline of competence – but it doesn’t substitute for commercial judgment built up over a career.
What Actually Needs to Change
Pricing certainty, not just design coordination
If the budget problem is fundamentally a pricing problem, then the solutions worth backing are the ones that attack pricing directly. That means better estimating tools that reflect live market data rather than historical averages. It means earlier contractor involvement in design, so that buildability and cost are stress-tested before the drawings are finished. And it means procurement structures that don’t reward the most optimistic bid.
Interestingly, this same challenge is playing out in construction insurance – a sector with obvious parallels. Justin Levine, CEO of Shepherd (which just closed a $42 million Series B), described a problem that mirrors the budget overrun issue almost exactly: contractors are building insurance assumptions into bids years before they actually buy the policy. By the time they go to market, rates have moved – often significantly. The result is that contractors are underwater on bids they’d already committed to at the earlier rate.
“Year over year, the cost of insurance continues to increase. And when contractors don’t have predictability, they get in a position where they’ve committed to a rate – and then they’re underwater on a bid.”Justin Levine, CEO of Shepherd, on the Bricks & Bytes podcast
Shepherd’s response is to offer rate certainty year over year, using behavioral data from tools like Autodesk and Procore to assess risk and lock in renewal rates for contractors who demonstrate consistent, safe practices. It’s a pointed example of a technology solution that’s actually solving the pricing visibility problem rather than just speeding up data entry. The parallel to what the best estimating platforms are trying to do is hard to miss.
Get the construction briefing that doesn’t sugarcoat it
Every week, 2,500+ construction executives and ConTech founders read Bricks & Bytes for the takes the industry press won’t write.
Join 2500+ ReadersThe Technology That Actually Helps
Tools that solve the pricing problem, not the coordination one
None of this is an argument against technology adoption in construction. It’s an argument for getting clear on what problem you’re actually trying to solve. If the core issue is that budgets are set without reliable pricing data and then locked in years before the market can validate them, then the tools worth prioritizing are the ones that close that gap: estimating platforms with live cost data, early contractor engagement tools that pull pricing intelligence into the design phase, and insurance products that offer genuine rate predictability over multi-year project timelines.
The more interesting question for ConTech founders is whether the industry’s contracting structures will bend enough to let those tools work. An estimating platform that prices work accurately is only useful if winning the job doesn’t require submitting an artificially low number. That’s a procurement reform problem as much as it is a technology problem. For founders targeting the budget overrun issue, understanding that distinction – and building around it rather than against it – is probably the sharpest edge available.
For a broader look at how construction tech pricing models are evolving across the vendor side, the Bricks & Bytes deep dive on ConTech pricing models covers the tension between value-based pricing and traditional per-seat structures in some detail. And for a VC perspective on where AI in estimating fits into the broader investment thesis, this piece on a venture capitalist’s hot takes on construction tech is worth the read alongside this one.
| Root Cause | Commonly Blamed On | What’s Actually Driving It | Tech That Helps | Tech That Doesn’t Help | Who Needs to Act |
|---|---|---|---|---|---|
| Concept budget vs. market price gap | Poor design coordination | Budgets set before builders are involved | Early-stage estimating platforms, integrated delivery models | BIM coordination tools (design problem, not cost problem) | Owners, developers, architects |
| Multi-year pricing drift | Project delays | Public procurement timelines vs. live market rates | Real-time cost data feeds, dynamic estimating | Project management dashboards | Procurement bodies, GCs |
| Underbidding to win work | Bad estimating | Contracting model rewards low bids; margin recovered through claims | Integrated project delivery, collaborative contracting | Any tool that optimizes bid speed | Clients, legal reform, procurement |
| Insurance rate volatility | Uncontrollable market forces | Rates priced annually; bids priced years in advance | Multi-year rate certainty programs (e.g. Shepherd Savings) | Standard annual renewal processes | Insurers, carriers, brokers |
| Change orders during construction | Poor site management | Clients change scope; unforeseen site conditions | Owner-facing change management tools, geotechnical data | Most – this is behavioral, not technical | Owners, PMs, contract structures |
The Revisto report puts the figure at 96%, but the stat lumps together a wide range of budget types across very different project categories and contracting models. The core drivers are structural: budgets are typically set at concept stage, before any GC has priced the work, using numbers that may not reflect live market conditions. By the time a project actually goes to tender – often two or three years later – those numbers are stale. Add change orders, insurance rate drift, and the commercial incentives around low bidding, and overruns become almost inevitable. The problem isn’t primarily a technology gap. (Source)
Partially – but only for a specific subset of overruns. Better model coordination can reduce errors that generate costly rework and design-related change orders during construction. But it doesn’t address the earlier, more foundational problem: budgets being set years before the market is consulted, by people without live pricing data. A perfectly coordinated design package can still be priced wrong if the concept budget was unrealistic to begin with. The technology is most useful when it’s combined with earlier contractor involvement and real-time cost data.
In a competitive tender environment, the contractor who prices a job accurately will often lose to someone willing to bid lower. Some firms price low deliberately, planning to recover margin through change orders and, in extreme cases, through litigation. Third-party litigation funders who back these claims in exchange for a share of any settlement are part of this ecosystem. The practical result is that the “budget overrun” is sometimes part of the commercial strategy from day one – not a failure of execution.
Insurance costs are one of several line items that contractors estimate at bid time but actually procure much later – sometimes years later. Since commercial insurance is typically priced annually and construction markets have been running hard for the last six years, the rate a contractor assumes at bid time can be significantly lower than what they actually pay when they go to buy the policy. Shepherd, which raised a $42 million Series B in 2025, is building toward multi-year rate certainty as a product feature to help contractors price more accurately. (Source)
Generally, yes – and for structural reasons. Public projects tend to use design-bid-build procurement, which excludes builders from the design phase entirely. The concept budget is often set by a cost consultant with limited live market data and approved by bodies that won’t bear the direct financial consequence if it’s wrong. Multi-year permitting and procurement delays then give the market time to move well past the original figures. Private developers, who feel cost overruns directly in their returns, tend to apply tighter commercial discipline from an earlier stage.
Shepherd, the AI-native commercial insurer targeting construction, is working toward what they call autonomous underwriting – a system where their AI ingests submission data and gets to an initial price without manual data entry from underwriting assistants. Where a typical carrier takes two to three weeks to respond to a submission, Shepherd aims to do it in under two days. For contractors, faster and more consistent pricing means better data when building bids. For the industry broadly, it’s a demonstration of what genuine process reimagination (rather than just faster automation) looks like in practice.
Related Articles
The Wild World of Pricing Models – How ConTech Companies Are Finding Their Numbers
A Venture Capitalist’s Hot Takes on Construction Tech
