Data-Driven Property Work: From Energy Label to Decisions

Most property decisions rest on a foundation that is more scattered than anyone cares to admit. The planning conditions sit in one place, the building data in another, title and easements in a third, and the energy label in a fourth — and when the decision finally has to be made, the answer is typically held in the hands of a single employee with a spreadsheet that cannot be reconstructed six months later. Data-driven property work is not about gathering even more data. It is about turning what already exists in the public registers into figures, reports and valuations you can act on — and doing it consistently enough that two employees reach the same conclusion from the same property.

The difference between having access to data and working in a data-driven way is enormous. Access is essentially free in Denmark: Plandata.dk (the planning data portal), BBR (the Buildings and Dwellings Register), tingbogen (the Land Registry), matriklen (the cadastre), CVR (the Central Business Register) and the energy-label register are all publicly available. The hard part is binding them together into one consistent picture of a property, and doing it fast enough that it actually feeds into the decision rather than becoming an after-the-fact rationalisation. This article walks through what a data-driven property looks like in practice — from the raw registers, through data quality, to a unified decision basis where the energy label and ESG belong as an integrated part rather than a separate compliance track.

What “data-driven property” actually means

The term is used loosely, so it is worth being precise. Data-driven property means that the decisions that cost the most — buying, selling, developing, refinancing, allocating within a portfolio — are made on a basis that is:

• Traceable. Every figure can be traced back to a source and a date. When an investment committee asks “how do we know that?”, the answer is available without making phone calls.
• Reproducible. The same property yields the same basis, regardless of who pulls it and regardless of when.
• Comparable. Two properties can be set side by side because they are described with the same fields and the same definitions.

The opposite — what we might call anecdote-driven property work — is not necessarily wrong in the individual case. An experienced developer often guesses right. The problem arises when the organisation has to scale: when a portfolio has to be screened, when a team has to hand over a case, or when an external party (a lender, an auditor, a buyer) has to be able to verify the valuation. This is where the undocumented gut feeling turns into a risk rather than a strength.

Rule of thumb: If a key figure in your decision cannot be traced to a source and a date, it is not data — it is an assumption, and it should be treated as one.

The public registers — and what each one answers

A unified picture of a property requires knowing which register answers which question. The most important public sources each cover their own layer of reality:

Planning basis — what may be built?

Plandata.dk gathers municipal and local plans. This is where you find the actual permitted use, the plot ratio, the number of storeys, and the constraints that determine what a site can carry. The local plan takes precedence over the municipal plan, and a valid local plan is binding until it is changed. The indicative plot ratios — typically 30 for detached housing, 40 for low-density terraced housing and 60 for multi-storey development — are the starting point, but the specific plan may set something different, and deviations require either a dispensation under section 19 of the Planning Act or a new plan. Reading the plan correctly is a discipline in itself; we have set out the approach in our walkthrough of how to read a local plan correctly.

Building data — what stands there now?

BBR describes the existing buildings: areas, use, year of construction, roof covering, heating installation. BBR is strong as a structural description but weak on data quality, because many records are old or were reported by the owner themselves. The area concepts in particular are often confused — footprint, total living area and commercial area are not the same figure, and an analysis that mixes them becomes misleading. The typical traps are worth knowing in advance; we have described them in our walkthrough of the most common pitfalls in BBR data.

Title and encumbrances — who owns it, and what is charged against it?

Tingbogen is the legal source for ownership (title), mortgaging (charges) and easements (burdens). An easement concerning right of way, building lines or a ban on subdivision can materially change a property’s real usability without appearing in either the plan or BBR. The Land Registry’s sections — title, charges, burdens — should be reviewed before any decision of consequence.

The cadastre and the companies

Matriklen establishes the geometric and legal boundaries of the parcel, and CVR ties ownership and company relationships together so that ownership chains and background can be traced. Together, the registers answer two different questions that are often confused: what is the property? (the cadastre, BBR, the plan) and who is behind it? (the Land Registry, CVR).

The combined exercise — pulling all these layers and setting them against one another — is the core of a thorough preliminary investigation. We have assembled it as an operational checklist for site due diligence that can serve as a framework for the data collection.

Data quality is the real bottleneck

It is tempting to think the problem is access to data. It rarely is. The real bottleneck is data quality — and the ability to assess it. Three factors are decisive:

1. Currency. When was the record last updated? A BBR area reported in 1978 and a local plan adopted last month carry vastly different weight.
2. Definition. What is the field actually measuring? “Area” means different things in BBR, in the Land Registry and in an estate agent’s sales particulars. A data-driven basis keeps the definitions apart.
3. Conflict. What do you do when two sources disagree? When BBR states one number of storeys and the local plan allows another, the disagreement is not noise — it is often the interesting signal itself, because it points to either a registration error or a development potential.

A mature data-driven practice therefore does not treat all figures as equally certain. It works with confidence: some fields are hard facts (cadastral number, registered title), others are soft indicators (an old BBR area, an estimated market rent). Being explicit about that difference is what separates a professional valuation from a number guess with two decimal places.

Where the energy label and ESG belong

The energy label is a good example of a register that has traditionally been treated in isolation — a document you obtain at sale and otherwise file away in a folder. In a data-driven decision basis, the energy label is instead a structural property of the building on a par with area and year of construction: it indicates operating costs, renovation needs and — increasingly — access to financing and tenants.

The energy label runs from A to G and says something about the building’s calculated energy consumption. It is not a snapshot of actual consumption but a standardised calculation, which is both a strength (comparability) and a limitation (it does not capture user behaviour). What the letters cover, and how they should be read, we have set out in our explanation of what a building’s energy label means from A to G.

ESG work builds on the same data. The environmental dimension of ESG is directly tied to the energy label, heat supply and building age — all fields that already exist in the public registers. The point of a data-driven approach is that ESG is not a separate data-collection project but a reformatting of data you already have into the frameworks a reporting requirement demands. Where to begin, and how to avoid making it heavier than necessary, we have described in our guide to getting started with ESG reporting on properties.

For anyone managing a portfolio, it becomes genuinely operational when energy label and building data can be pulled across many properties at once — for example, to find the buildings where a low energy label and a high age together point to the greatest renovation or risk need. We have covered that discipline separately in our article on screening a property portfolio on energy label and building data.

From property data to report — the last, decisive step

Even a perfect data foundation creates no value until it becomes something a decision-maker can read in five minutes. The path from property data to report is the step where most organisations lose time — because it is typically manual: copy-paste between registers, a spreadsheet that formats, a presentation built from scratch every time.

A good decision basis has certain hallmarks, whether it is produced manually or automatically:

• It leads with the conclusion. The figure and the recommendation sit at the top; the documentation follows for those who want to dig.
• It shows its uncertainty. A valuation without a confidence statement invites over-interpretation.
• It is comparable. Same fields, same definitions, same order — so two cases can be held up against each other without translation.
• It is dated and sourced. A basis without a date goes stale silently; a basis with a date can be deliberately re-assessed.

This is where proptech for the property industry has its real justification. The value does not lie in giving access to registers that are already public, but in removing the manual link between the register and the report — and in making it reproducible, so the basis does not depend on who built the spreadsheet.

How to automate the exercise — without losing traceability

All of the above can be done by hand. It is just slow, and it is hard to repeat. The manual exercise — look it up in Plandata.dk, cross-check BBR, read the Land Registry, find the energy label, weigh the sources against one another and write it up — is precisely what a data-driven property platform is built to remove the friction from.

In Arcili, it is exactly this interplay that the Assistent (Assistant) binds together: it pulls across the public registers, keeps the definitions apart, shows which source a figure comes from, and assembles it into a basis you can use directly — without losing the traceability back to Plandata.dk, BBR, tingbogen, matriklen and the energy-label register. It does not replace the expertise of valuing a property; it removes the hours otherwise spent gathering and formatting, so that expertise can be applied to what actually requires a professional judgement.

If you want to see what your own data exercise looks like once the manual link is gone, you can explore Arcili or book a walkthrough, where we take a property or portfolio you already know as the starting point.