Construction Has a Data Problem
Every project generates thousands of documents. Almost none of that knowledge survives to the next bid.
A project manager told me recently that his team had won a large commercial project using almost identical scope to one they'd built two years ago. They had no idea. They estimated from scratch. It cost them margin they never recovered.
The data was there. It always is. Locked inside a finished project folder that nobody opened.
This is the real data problem in construction. Not that the data doesn't exist. It does — in enormous quantities. The problem is that it disappears the moment a project closes.
The Numbers
95% of construction data is unstructured — PDFs, emails, photos, handwritten notes
13% of working hours on a project are spent just searching for information
€1 in €3 of re-estimation costs could be avoided with better use of historical data
The documents are there. The knowledge isn't.
Think about everything a single project generates. Tender documents. RFI logs. Submittal packages. Change orders. Daily reports. Inspection records. Subcontractor correspondence. Site photos. Meeting minutes. Punch lists. Handover documents.
Each one contains decisions. Context. Lessons. The kind of institutional knowledge that normally lives in someone's head and walks out the door when they change jobs.
The industry doesn't have a data collection problem. It has a data access problem.
Every finished project is a library of answers to questions your estimators are about to ask again. The problem is nobody catalogued it.
What gets lost when the project closes
When a project finishes, three things usually happen with the data. It gets archived in a folder structure nobody ever standardised. The people who understood it move on to the next job. And the next tender starts from a blank spreadsheet.
It's not negligence. It's just how the industry runs. There's no time between project close and next bid to do anything with what you learned. So the knowledge stays buried.
Until the same scope comes up again. And the same mistakes get estimated in again. And the same surprises show up on site.
The shift that changes everything
The construction firms that are getting ahead of this aren't building data warehouses or hiring data scientists. They're asking a simpler question: what documents do we already create, and can we make them searchable and useful?
Tender documents are a perfect example. They contain exactly the kind of structured scope information that trips up estimators — special conditions, performance requirements, liability clauses, scope boundaries that aren't obvious on first read.
When you can search across past tenders for similar language, similar exclusions, similar risk clauses — you stop re-learning lessons you already paid for.
You don't need a data transformation. You need one better question.
The firms winning on this aren't doing a five-year digitalisation roadmap. They're asking: what's one place where we're making decisions without the data we already have?
For most teams, the answer is the bid. It's where the highest-stakes decisions get made with the least access to institutional knowledge.
That's the place to start. Not with a platform. Not with a strategy. With one project folder and the question: what did we learn last time we built something like this?