Questions · heavy civil estimating

Why is estimating software so bad in heavy civil construction?

We have extraordinarily powerful computers. Why are estimators still spending hours moving numbers from one box to another?

I have sat in that seat. A takeoff produces quantities. Those quantities get typed again into the estimate. Bid items, crews, production rates, and cost codes get built by hand. Sub and supplier quotes get copied in. Bid leveling is a person comparing quotes side by side. The same numbers then shuttle across PDFs, Excel, HeavyBid or HCSS, and whatever accounting or ERP the office uses.

The interesting part is not that the work is hard. Judgment is supposed to be hard. The interesting part is how much of the day is data transformation: getting a number from a place that already has it into a place that needs it.

I have not found a credible industry survey that says estimators spend X hours a day on re-entry. That number would be useful, and I do not want to invent it. What we do have is a long, consistent picture of spreadsheet-heavy, poorly integrated workflows.

What the surveys actually show

The 2019 JBKnowledge Construction Technology Report, as summarized by ConstructConnect, found estimating workflows still relied on spreadsheets at 64.9%, up from 63.2% in 2018. Only 54.6% reported dedicated estimating software. Those two numbers sitting next to each other are the tell. Software and spreadsheets coexist. That usually means re-entry and parallel workflows, not a clean handoff.

The same pattern shows up in later ConTech coverage. The 2020 JBKnowledge report still listed estimating among the workflows companies were most likely to run on spreadsheets instead of software, at 62%. I am using the published report here, not a new survey of my own. Estimating stayed in the 60%+ spreadsheet range even as dedicated estimating software remained common.

FMI’s Procore-sponsored study, Where Construction Firms Are Finding Value, asked about technology problems. The top mentions were low user adoption (34%), poor ease of use (17%), and lack of integration (16%). Buying a tool is not the same as getting data to move, or getting people to live in one system.

McKinsey Global Institute’s construction productivity work is the wider frame. Global construction labor-productivity growth averaged about 1% a year for two decades, against about 2.8% for the world economy and about 3.6% for manufacturing. Construction sits among the least digitized sectors. That is not only an estimating problem. Estimating is one of the places you can feel it with your hands.

Their construction-tech ecosystem research describes a fragmented market. More than half of the companies they mapped covered only one or two narrow use cases out of 38. In an earlier cut of the same research, just 13% of solutions spanned more than one of the three clusters they used: field, collaboration, and back-office. Data stays siloed because many tools were never designed to talk. Estimating, in the day-to-day of a heavy civil office, lives in that isolated preconstruction and back-office world.

What vendors say, labeled as such

I prefer the industry surveys above to vendor claims. Vendor copy is still useful as a signal of the problem they are selling against.

HCSS markets takeoff-to-HeavyBid integration specifically to reduce manual entry and errors. That is a product page, not a survey. On an HCSS video page, an estimator describes quick sub and material entry cutting “probably about fifty percent of the data entry time.” That is a vendor-hosted testimonial. I am including it because the industry’s own software vendors are openly selling the reduction of re-keying.

What I am actually asking

I build construction software. I am not writing this as a pitch. I keep coming back to the same question: if the computers are this good, why is so much estimating still a person carrying numbers between systems?

Is it the products? The data? The way bids are structured? The way firms buy and adopt tools? Some mix of all four?

If you have a source, a better framing, or a day-in-the-life number that is actually measured, I want it.