What Each Part of a House Is Worth, and How Much That Changes by Region
A study of what actually moves an Australian house price, and of how differently the same feature pays in one part of the country versus another.
By Luke Metcalfe, Founder and Chief Data Scientist, Microburbs Research
The parts, not the total
This work deliberately does not produce a price for a house. Most price guidance resolves to a
single figure, which a reader can only accept or reject, and which is of no use at a negotiating
table. What is published here instead is the signed list: each part on which a particular house
differs from the homes around it, and what that part is worth. Every line can be checked, and every
line can be argued on its own.
For a house in Brisbane that sold this year, the list reads as follows. A fifth bedroom where
the surrounding homes have four adds $18,500. A third bathroom adds $16,000. No total is given for
those lines, because the sum would be a price, and a price is the thing this work declines to
assert. A sale ad that does not mention a pool or renovation is not treated as proof that the
feature is absent.
The composition is the entire product. A buyer holding a single number can only disagree with
it. A buyer holding the two lines above has two separate propositions, each measured on comparable
sales in that region and each defensible on its own evidence. That is the difference between an
opinion about a price and a position on one.
Abstract
Most price guidance treats a house as a single number. This work takes the house apart instead,
and puts a value on each part of it: the pool, the renovation, the land, the bedrooms, the bathrooms,
the garage. The values come from comparing about 620,000 Australian homes, and every value is
expressed as a share of what a house sells for locally, because the same feature is worth several
times more in one region than another. That regional variation is the main practical finding here,
and it is the reason a national average is a poor guide for any particular house.
Because a figure measured once proves little, each one is then measured again on years and on
places it was never fitted to. The values reappear year after year, and the parts of a house get
closer to the eventual sale price than the suburb figure alone in every half-year tested, in
5 of 6 states and 27 of 33
regions. The exception is named rather than averaged away.
Key findings
What changes
Share of the price
On a typical home
A whole-home renovation
7.9%
$67,000
A swimming pool
3.2%
$27,000
One more bedroom
2.3%
$19,000
One more bathroom
2.1%
$18,000
One more garage space
1.4%
$12,000
Solar panels
1.2%
$10,000
Open plan living
nothing
nothing
Air conditioning
nothing
nothing
Two of the features most often advertised, open plan living and air conditioning, carry no
measurable premium at all. They are worth stating precisely because they are so commonly used to
justify a price.
Land does not behave the way it is usually sold. On a block under 400 square
metres the next 100 square metres is worth about $31,000. On a block of 800 to
1,200 square metres the same 100 square metres is worth about $9,000. Extra
land pays hardest where there is least of it.
It varies enormously by region, and we did the work region by region
A single national figure for any of these would be misleading, because the same work pays very
differently depending on where the house is. A whole-home renovation is the clearest case.
What a renovation returns, by region
Region
What a renovation returns
Share of the price
Sydney - Parramatta
$185,000
+12.9%
Sydney - North Sydney and Hornsby
$180,000
+7.2%
Sydney - Sutherland
$122,000
+7.7%
Sydney - Baulkham Hills and Hawkesbury
$121,000
+5.4%
Sydney - Inner South West
$115,000
+7.8%
…
Perth - North West
$34,000
+3.2%
Sydney - Outer West and Blue Mountains
$33,000
+3.0%
Adelaide - North
$24,000
+4.3%
The gap between the top and the bottom of that table is more than seven times. A renovation in
Sydney - Parramatta returns roughly $185,000. The same work in
Adelaide - North returns roughly $24,000. Anyone applying a
national average to either would be badly wrong, in opposite directions.
What a pool is worth, by region
The same pattern holds for a pool. It is worth about $62,000 on the
Sunshine Coast and about $10,000 in
Melbourne - West. Climate, block size and how many neighbours already have one all pull
in the same direction.
We have region-level answers for a renovation in 49 regions
and 12 smaller districts, for a pool in
27 regions, and for solar in
12. Where a region cannot support its own answer on its own
evidence, it inherits the next size up, whole city then state, and the reader is told which level
the figure came from rather than being handed a national number dressed up as a local one.
Why the suburb median misleads
Almost every price a buyer is quoted rests on a suburb figure. Yet the homes closest to a house
frequently sell well away from that suburb figure, in both directions. A suburb median treats the
best street and the worst street in the same postcode as the same place, and buyers already know
they are not. Pricing the street a house actually sits on, rather than the suburb around it, turns
out to matter more than any single feature in this paper.
Worked examples
Three sales that illustrate the findings. Each is a matter of public record.
A pool, Baldivis, Perth
29 Morwell Loop sold for $890,000. 19 Morwell Loop, the same street, both four bedrooms and two
bathrooms, sold for $859,000 five months later in 2025. The difference is $32,000.
A pool, Coburg North, Melbourne
19 Ronald Street sold for $725,000 and 33 Ronald Street sold for $700,000, both two bedrooms
and one bathroom, on the same day in 2026.
Solar, Carlisle, Perth
29C Star Street sold for $925,000 and 29B Star Street, next door, both three bedrooms and two
bathrooms, sold for $900,000 nine weeks earlier in 2025.
These are three examples from about 620,000 Australian homes compared to
arrive at the figures above, not the basis for them.
Implications for buyers and buyers agents
The practical use is at the negotiating table. When a sale ad does not show a feature that is
common in nearby sale ads, the difference can be named and evidenced with real
sales. A house whose advertising leans on features that carry no premium can be discussed on that
basis.
For a professional buyer, this is the part of the service clients already pay for separately.
The value at stake in a single negotiation is routinely larger than the cost of the research behind
it.
The regional table above is the caution as much as the finding. A buyer applying a national
renovation figure in Adelaide - North would overestimate by several times. Use the figure
for the region in question.
Does it hold up over time, and does it hold up everywhere?
A figure measured once, on one pile of sales, proves nothing about next year or about the other
side of the country. Both questions are answered the same way: measure it again on slices it was
never tuned to, and see whether the same answer comes back. Nothing below is refitted. The values
are the published ones; the slices only decide which sales get to judge them.
What is being measured on what, stated plainly before any result. The values
themselves come from the whole country: about 620,000 homes, every state, houses only. The
head-to-head test that follows is deliberately narrower. It runs on about 30,000 sales
in 254 suburbs, restricted two ways: to homes selling between $400,000 and
$800,000, which is the range our buyers actually shop in, and to suburbs where the houses are
genuinely alike, so that comparing one to another means something. Both restrictions are choices,
and both are stated because they change the number a great deal. Without them, on
about 97,000 sales at every price from $150,000 to $18 million across
656 suburbs and 87 regions, both answers are far
weaker: the suburb figure alone lands within 10% on 34.3% of homes and
pricing the parts takes it to 39.7%. Pricing the parts still wins there, by
5 homes in every hundred rather than
6. The narrow figures below are the ones worth quoting because
they describe the homes our readers actually buy, not because they flatter.
Over time
The first test re-measures a feature from scratch inside each year, using only that year's own
sales. A figure that is really there should keep turning up.
A swimming pool comes out at 4.1% in 2024,
3.1% in 2025 and 3.4% in
2026, on entirely different sales each time. The published figure is
3.2%. The 2024 reading rests on about 350 comparisons against
about 7,600 the following year, which is why its range on the chart is so much wider and
why it should be read as agreeing with the others rather than as a higher answer.
The renovation can be traced back further, because it is measured on homes that sold twice
rather than on pairs of neighbours. That gives 6 separate years to check.
One of these has genuinely moved, and it would be dishonest to smooth it. Each
figure above is measured against two benchmarks at once: what the home's own suburb did between
its two sales, and what homes that were never renovated did from the same starting point. On that
basis a renovated home resold 12.8% ahead in 2021, at the top
of the post-lockdown building rush, and 7.2% to
7.5% ahead across the three years since. It is ahead in
every year measured, which is the finding that matters, but the size of it was larger during the
boom. The published figure follows the recent years rather than the peak.
The second test is harder. Take real sales the figures were never fitted to, and put two guesses
side by side on the same homes: what a house sells for in that suburb and nothing else, against the
same suburb figure adjusted for every part this paper prices. Then score each half-year separately.
The built-up answer wins in 5 of 5 half-years, by between
5 and
6 homes in every hundred. There is no half-year in
which pricing the parts of the house made the answer worse.
Over geography
The same discipline, applied to place instead of time. First the values themselves, re-measured
inside each state.
State
A typical renovation
A pool
Solar
Canberra
+5.6%
not enough
+1.2%
New South Wales
+8.3%
+2.4%
+0.8%
Queensland
+7.3%
+3.6%
+1.4%
South Australia
+6.2%
+1.9%
+0.8%
Tasmania
+9.2%
not enough
not enough
Victoria
+9.3%
+2.9%
+1.3%
Western Australia
+6.0%
+3.5%
+1.1%
Every state that has enough sales to answer gives a positive figure for all three, and the
ordering is sensible rather than random: a pool pays most in Queensland and Western Australia, and
least in South Australia. The sizes differ, which is exactly the argument for doing the work region
by region rather than quoting one national number.
Then the same head-to-head test, scored state by state.
It does not work everywhere, and here is where it does not. In
South Australia the suburb figure on its own is already the
better guess: it lands within 10% of the sale price on 70% of homes, and
adjusting for the parts of the house pulls that down to 67%. That is the only
state where this happens, and the reason is visible in the same number: at 70%
the suburb figure there is the strongest in the country, because those suburbs are unusually
uniform. Where the houses in a suburb are all much the same, knowing the suburb is already most
of the answer, and adjusting for differences that barely exist can only add noise.
Below the state level the picture is the same. Of the 33 regions with enough
sales to be scored on their own, the built-up answer is closer in 27 of them.
The regions where it helps most are the ones where houses differ from each other: the outer and
regional markets where a renovated four-bedroom sits next door to an original three-bedroom on the
same street. The handful where it does not help are the uniform ones, and they are named above
rather than averaged away.
That is the honest summary of reliability. Pricing the parts of a house improves on the suburb
figure in 5 of 6 states, in 27 of
33 regions and in every half-year tested, on about 30,000 sales across
254 suburbs that none of the figures had seen. It helps most where houses differ from
their neighbours and least, occasionally negatively, where they do not.
Limitations
Three are worth stating plainly.
Some things cannot yet be answered. Sloping blocks are one. Australian agents
advertise a level block thousands of times and a sloping block barely at all, so the wording of a
listing cannot measure a slope. Bush frontage is another: it is concentrated in a handful of places,
so a national comparison has too few matched cases to be firm. Both are real and both are unanswered
here rather than guessed at.
A renovation is a description, not a measurement. The signal is whether a
listing describes a home as renovated. That says whether work was done, not how much, and not how
well.
This is not an appraisal. It is evidence about what parts of a house are worth
in a given region. It is not a valuation of any particular property and should not be relied on for
lending, tax or legal purposes.
Conclusion
A single number invites the reader either to believe it or to dismiss it. A list of parts, each
with a value and the real sales behind it, can be checked, argued with, and used. The regional
variation is what makes it useful in practice: the same renovation is worth more than seven times as
much in one Australian region as in another, and knowing which one you are standing in is worth more
than any national average.
The values survive the two tests that matter. Measured again on each year's own sales they keep
returning the same answer, and applied to sales they had never seen they get closer to the price in
every half-year, in 5 of 6 states and 27 of
33 regions. Where they do not help, in the most uniform suburbs, that is stated here
rather than buried. A figure a reader can check in their own market is worth more than one they are
asked to take on trust.
By Luke Metcalfe, Founder and Chief Data Scientist, Microburbs Research