How Accurate Is Your Domain Property Estimate? We Checked 1,689 Gold Coast Sales

You’ve probably looked up your home on Domain and seen an estimated value. A number, a range, and not much else. No explanation of how they got there. No indication of whether the estimate is reliable for your specific property, your suburb, or your price bracket.

We wanted to know: how often does Domain get it right?

So we compared Domain’s automated valuations against the actual sale prices of 1,689 properties sold across eight southern Gold Coast suburbs — Robina, Burleigh Waters, Varsity Lakes, Mudgeeraba, Reedy Creek, Merrimac, Worongary, and Carrara.

Here is what we found.

The headline numbers

Domain provides a valuation range — a low, a midpoint, and a high — for most properties. Of the 1,689 sales we analysed:

• 62.7% of properties sold within Domain’s estimated range
• 34.8% sold below the bottom of Domain’s range
• 2.5% sold above the top of Domain’s range

That means roughly one in three sellers received an estimate that was higher than every offer they ultimately received.

The median difference between Domain’s midpoint estimate and the actual sale price was $150,000.

Domain overvalues almost nine out of ten properties

This is not a random scatter. Of the 1,689 sales, 88.9% of Domain’s midpoint estimates were above the actual sale price. Only 10.2% were below.

Automated valuation models draw on recent comparable sales, property attributes, and market trends. But they cannot inspect a property. They do not know about the water damage behind the bathroom wall, the noise from the construction site next door, or the fact that the kitchen was renovated last year. The result is a systematic lean toward the optimistic side.

This matters because a seller who anchors their expectations to a number that is too high may reject reasonable offers, sit on the market longer, or choose the agent who promises the highest price rather than the one who will achieve the best result.

Your suburb changes everything

Domain treats all suburbs through the same algorithm, but accuracy varies enormously depending on where you are.

Suburb	Properties	Sale within range	Sold below range	Median error
Robina	195	97%	1%	5.7%
Burleigh Waters	225	93%	2%	4.5%
Mudgeeraba	201	88%	10%	5.7%
Carrara	221	71%	26%	8.0%
Varsity Lakes	230	51%	46%	11.8%
Reedy Creek	260	48%	50%	14.2%
Worongary	126	32%	65%	18.9%
Merrimac	231	19%	80%	30.5%

If you live in Robina or Burleigh Waters, Domain’s estimate fell within its own stated range 93–97% of the time in the period measured. That is genuinely useful.

If you live in Merrimac, the estimate was above the actual sale price 80% of the time, and the median error was 30.5%. That is not a valuation — it is a guess.

Domain does not tell you any of this. Every property gets the same presentation: a number, a range, and a confidence badge. There is no disclosure that the model performs four times better in one suburb than the suburb next door.

Price bracket matters more than you would think

We broke down accuracy by the actual sale price of each property.

Price bracket	Properties	Median error	Within 10%
$500,000–$750,000	40	34.4%	10%
$750,000–$1,000,000	220	24.5%	20%
$1,000,000–$1,250,000	315	12.7%	40%
$1,250,000–$1,500,000	428	10.5%	47%
$1,500,000–$2,000,000	348	8.3%	60%
$2,000,000–$3,000,000	143	7.9%	66%

Properties under $1,000,000 had a median error of 24–34%. Above $1,500,000, the median error dropped to around 8%.

Why? Higher-value properties tend to be more homogeneous (similar) within their market — established houses on standard lots in well-documented suburbs. The algorithm has more comparable sales to draw on, and the comparables are more similar to each other.

At the lower end, you get a wider mix: older homes needing renovation, smaller lots, unusual configurations, and properties where the condition of the home matters more than the land value. The algorithm struggles with these because it cannot see condition.

Bedrooms and land size affect reliability

Bedrooms	Properties	Median error	Within 10%
3 bedroom	470	10.5%	47%
4 bedroom	707	10.4%	48%
5 bedroom	257	14.0%	40%
6 bedroom	52	19.1%	33%

Three and four bedroom houses are the most common property type, and the algorithm has the most data to work with. Accuracy drops as properties become less standard.

The same pattern holds for land size. Properties on lots between 600 and 1,200 square metres had median errors of 10–11%. Acreage properties — lots over 5,000 square metres — had a median error of 23%. These are properties where the land itself varies significantly (slope, usable area, access, views), and an algorithm that only knows the lot size in square metres will miss the details that drive price.

The real problem is not accuracy — it is opacity

Automated valuations will always have error margins. No model can account for every variable that affects what a buyer will pay for a specific property on a specific day.

The issue is that Domain does not tell you how confident it is in your specific estimate. Every property gets the same treatment: a number, a range, and no methodology.

A seller in Robina with a standard four bedroom house on 700 square metres gets the same style of estimate as a seller in Merrimac with a six bedroom house on an acre — even though one estimate is likely to be within 5% of reality and the other could be off by 30%.

An automated estimate is a starting point. What determines how much weight it can carry is:

• How many comparable sales the estimate is based on
• How similar those comparables are to your property
• Which adjustments were made for differences in land size, bedrooms, condition, and location
• How much the comparables disagree with each other — a tight cluster of similar sales means more confidence; a wide spread means less
• Whether the model performs well in your suburb at your price point

That is what interpretable valuation looks like. Not a better number — a number you can actually evaluate.

What this means if you are selling

Domain’s estimate is a reasonable starting point for some properties in some suburbs. For a standard house in Robina or Burleigh Waters, it will often land within a useful range. For properties in Merrimac, Worongary, or at the lower end of the market, treat it as a very rough indicator.

Before you anchor your pricing strategy to any automated number:

1. Check whether your suburb and property type are in the algorithm’s sweet spot. The data above tells you.
2. Ask for the methodology. If someone quotes you a number and cannot explain how they got there, the number is not useful to you.
3. Look at the actual comparable sales. What did similar properties near you sell for in the last six months? What made them different from yours? This is the foundation of any credible valuation.

We publish our comparable sales analysis, our adjustments, and our confidence levels for every property on our platform — because a number without context is just a number.

---

This analysis covers 1,689 properties sold between 2024 and 2026 across Robina, Burleigh Waters, Varsity Lakes, Mudgeeraba, Reedy Creek, Merrimac, Worongary, and Carrara. Domain valuations were captured at listing time. Sale prices are as recorded on Domain. Fields Estate has no affiliation with Domain.

Fields Estate. Smarter with data.