Selecting Comparable Sales for a Credible Valuation

A valuation is never stronger than the sales it rests on. You can have the sharpest model, the most carefully worked area schedule and the tidiest spreadsheet — but if the basis is a handful of poorly chosen comparables, the whole conclusion collapses the first time an auditor, a credit committee or a counterparty’s lawyer puts it to the test. And they do. A price per m² is not a number you present; it is a number you have to be able to defend, sale by sale.

The problem is that the selection itself rarely gets the attention. We spend hours fine-tuning adjustments for condition and location, yet make the decision about which sales even go into the basket in a couple of minutes — often intuitively, often shaped by whatever happens to be easy to find. This guide makes that selection explicit and methodical. We work through the four questions that decide whether your comparable sales support a conclusion or undermine it: how widely you search, how far back you go, what actually makes two properties comparable, and how many sales it takes before the price per m² is anything more than a guess.

What a comparable sale actually is — and what it is not

A reference property is not simply a nearby sale in the right price range. It is a sale where the price-bearing characteristics are similar enough that you can explain the price difference to the subject property with known, justifiable adjustments — not with a shrug.

That distinction matters, because most weak valuations do not fail on the maths. They fail because one or more of the references does not really belong in the basket:

• The sale between related parties — an intra-family transfer, an intra-group sale, a sale arising from a divorce or an estate. The price reflects a relationship, not a market.
• The forced sale — a foreclosure auction, a bankruptcy, an acute need for liquidity. Buyer and seller were not on equal footing, and the price is not a free expression of the market.
• The sale with hidden baggage — a major easement, a pending dispute, included chattels, or agreed financing that pulls the nominal price away from the property’s own value.

A good working principle is that a sale only belongs in the basket if you can describe it in a single sentence with no caveats: “detached single-family house from the 1970s, 140 m², in normal condition, sold on the open market between independent parties.” If you cannot say that sentence cleanly, the sale must either be deliberately adjusted or dropped.

What the sale data does not tell you on its own, you have to pull from the registers. The tingbogen (the Land Registry) shows title and can reveal related parties, easements or encumbrances that affect the price. BBR (the Buildings & Dwellings Register) tells you whether the areas the price was calculated on are even the same area concepts you use for the subject property. And a quick check in CVR (the Central Business Register) can show that buyer and seller share an ownership circle. None of this comes attached to a sale price automatically — you have to bring it in yourself.

Radius and time horizon: the two dials that decide representativeness

Once you are clear on what a valid sale is, the next two decisions are geography and time. They pull in opposite directions, and the craft lies in finding the balance for this particular subject and this particular market.

Radius: wide enough for volume, narrow enough for relevance

Geography is the strongest price factor we have, and at the same time the hardest to adjust away. Two identical houses on opposite sides of a municipal boundary, a lake or a railway line can carry markedly different prices without your being able to sensibly “correct” for the difference. So the starting point is always the narrowest possible radius that yields sufficient volume.

In a dense urban area with high turnover, a few hundred metres may be plenty — and there you should stay close, because micro-location (a quiet residential street versus a busy main road) means everything. In the countryside or in thinly traded segments you have to open the radius up just to find anything at all, and then the demands rise on how carefully you adjust for the differences the distance introduces. The point is not a single number, but a deliberate trade-off you can account for. We go into the balance in practice in our review of radius and time horizon for comparable sales.

Time horizon: fresh enough for the market, long enough for a data basis

The same dilemma applies to time. Fresh sales reflect the current market best, but the shorter the window, the fewer sales — and the greater the risk that a single outlier comes to dominate. A longer window gives volume and stability, but draws in older sales that were struck under different market conditions.

Rule of thumb: start with the narrowest window that gives you enough valid sales for a robust average. Only widen it if volume forces you to — and always stay on the terms of the existing market. A reference basket should describe the market as it has been traded, not the market as you think it is moving.

That last point matters: references document actual, completed sales — not a forecast. The moment you adjust for where the market “is heading”, you have left the method and entered guesswork.

What actually makes two properties comparable?

Location and price get you close to the right sales. But comparability itself is decided by whether the properties’ price-bearing characteristics match — and whether the differences that do exist can be adjusted for credibly.

The characteristics that genuinely move the price are typically:

• Property type and use. A terraced house is not a detached house, an owner-occupied flat is not a villa apartment, and commercial property is an entirely different game. Never mix types in the same basket without a very deliberate reason.
• Area and area concept. This is where many stumble. BBR works with several area figures — living area, footprint, total building area — and if you calculate the price per m² on one and your reference on another, you are comparing apples with pears. Make sure it is the same area concept all the way through.
• Age, condition and energy. A newly renovated house and a run-down house of the same size on the same street are not the same sale. Condition is the hardest to adjust objectively and the one where the most errors hide.
• Site conditions and building rights. Two plots of the same size are not worth the same if one holds building rights the other does not. What may be built follows from the municipal plan and the local plan in Plandata.dk (the national planning register) — and differences in plot ratio or use can justify large jumps in price. If land value or development potential is part of the picture, you should have read the planning conditions; our guide to reading a local plan correctly covers building rights, plot ratio and use.

Comparability is therefore a scale of degree, not a binary yes/no. A sale can be a strong reference on five characteristics and diverge on one — and then the question is whether the divergence can be adjusted with a justified rate, or whether it is too large for the sale to be usable. We unpack that judgement — which differences can be adjusted away, and which disqualify a sale entirely — in the article on what actually makes two properties comparable.

How many sales does it take?

The most common question — and the one with the most wrong answers. There is no single magic number, but there is a logic.

Too few sales, and you are exposed to chance: one atypical sale can pull the average several percent in either direction, and you have no way of seeing whether it expresses the market or noise. Too many sales — if you have had to stretch radius and time to reach them — and you have probably diluted the basket with references that are no longer genuinely comparable.

The usable guideline is not about the number alone, but about the spread:

• When your valid references cluster tightly around a price per m², relatively few sales can support a credible conclusion.
• When they spread widely, more sales do not help you — they tell you the segment is heterogeneous, and that you must either tighten the selection criteria or accept a wider band of uncertainty and say so out loud.

In other words: the number is not the goal, but a means to bring the spread down to a level where the median actually means something. A valuation built on four tightly matched sales with low spread is more credible than one built on fifteen scattered sales where half only nominally resemble the subject. How to concretely assess whether the basis holds — and communicate the uncertainty honestly — we cover in how many comparable sales support a price per m².

The documentation is part of the valuation

A credible valuation is not only the right price per m² — it is a price per m² an outsider can follow you to. When the bank, the auditor or the counterparty looks over your shoulder, it is not the conclusion they attack first; it is the basis. Traceability is therefore not decoration, but part of the product itself.

Concretely, this means that for each reference you should be able to present:

• Which sales you have included — and why exactly those belong to the segment.
• Which you have excluded — and on what grounds (forced sales, related parties, too large a divergence).
• Which adjustments you have applied — and with what justification, so that condition and location adjustments do not look arbitrary.
• Radius and time window — so it is explicit how widely and how far back the basis reaches.

Anyone who cannot show their exclusions has no method, really — only a result. And a result without a method does not survive the first critical read-through. The same systematic approach applies, incidentally, when you check a site before a purchase, as we describe in the site due diligence checklist.

From manual curation to a basis that holds

The entire exercise above — defining radius and time window, sorting out forced and related-party sales, ensuring the same area concept, assessing the spread and documenting the exclusions — is fully something a skilled professional can do by hand. It just becomes slow and error-prone when it has to be done properly every time, across many subjects.

That is exactly where the Ejendomshandler (Comparable sales) module comes in. It gathers a large, comparable data basis, lets you adjust radius and time horizon, and gives you a price per m² where you can see the basis behind it — which sales support it, and which fall outside. It does not remove the professional judgement; it removes the manual legwork, so you can spend your judgement where it makes a difference: on the exclusions and the adjustments. The valuation remains yours — but the basis becomes verifiable from the first click.

If you want to see how it looks on your own properties, start in Arcili or book a walkthrough, and we will show you the selection in practice.