Days on Market (DoM) and Discounting are important metrics used by real estate professionals to gauge market demand. Over the last 3 months, HtAG data analysts made good progress enriching our dataset aiming to incorporate these 2 metrics into all HtAG reports and dashboards.
The dev team are now sprinting to implement this feature on the site, targeting an early December deadline. In this post, we are releasing the preliminary Q3 DoM/Discounting report for 3,342 suburbs to the public. The table below summarises these 2 metrics for house, townhouse, unit and land sales in Q3.
|ID||Suburb||State||Postcode||DoM||Discount %||Property Type||No of Sales||Confidence|
Whereas DoM is available for all suburbs in the table, in some instances Discounting cannot be calculated due to data sparsity in listed/sold price. As we are constantly adding new data to our dataset, the expectation is that Discount % will soon be available for all suburbs on the list.
So what are the DoM and Discounting metrics and why are they important? Alex Hill from Over and Above explains Days on Market thoroughly in this short youtube video.
Now that we understand why DoM is an important metric for both buyers and sellers, let’s deep dive into the statistical analysis of DoM and Discounting data. By the end of this analysis, you’ll realise that these 2 metrics are closely related.
DoM (Days on Market)
DoM is the difference in days between the date of property first being listed for sale and the actual day of sale. Shorter DoM indicates strong demand for the property i.e. the smaller the value the better (or worse if you are a buyer). As can be seen on the histogram below, majority of suburbs have DoM in the 25-75 day range. Based on that we can conclude that:
- DoM of 25 days or less is representative of seller’s markets experiencing strong demand from buyers
- DoM of 25-75 days is representative of balanced markets where demand generally meets supply
- DoM of 75 days or more is representative of a buyer’s market
It is important to note that the DoM distribution for units and houses is slightly different. DoM for land sales has a distribution unique among all other segments. More on this below.
DoM is calculated as an average of all transactions that HtAG crawler indexes in the current calendar quarter per suburb/locality. This calculation is based on properties listed and reported as sold online, so off-market sales are excluded. Because the majority of properties are listed online, omission of off-market data has no influence on the accuracy of this metric, providing a large enough sample size is used. Where the sample size is below a threshold established by our data analyst, the Confidence metric is adjusted to be reported as Low.
As the name suggests, Discounting is a percentage by which the vendor reduced the advertised price by. This metric is calculated off the back of the same transactions as DoM, except that instead of day difference it is represented as a percentage difference in price. Here is the formula:
Discounting = (Advertised Price – Sale Price) * 100 / Advertised Price
Counterintuitively, negative percentage means that the property was sold for more than it was advertised for. Negative percentages are rare and are indicative of very strong demand in the market. The histogram below suggests that a discount in the range of 2-5% is very common among the vendors in the suburbs on our list.
Similar to DoM, the Discounting metric is calculated as an average discount of all known on-market sales in the suburb/locality. Rows with sparse sample data are tagged with low confidence.
Is DoM a tie-in for Discounting?
Now that we’ve quantified DoM and Discounting, let’s explore whether these 2 demand metrics are correlated. Correlation analysis is used to establish whether any 2 (or more) variables are interrelated. A correlation index of 0.00 implies there is no relationship, 1.00 on the other hand signifies that the variables tie-in perfectly. In our case, an index of 1.00 would mean that every additional day on market results in a fixed percentage increase in Discounting.
The image and table below show the correlation matrix for DoM, Discounting and No of Sales. Note that in this instance No of Sales is not the total sales recorded for the suburb, but is rather the sample size used to calculate DoM and Discounting.
|DoM||Discounting||No of Sales|
|No of Sales||-0.04||-0.03||1.00|
We can see that DoM and Discounting have an index of 0.31, which indicates a moderate correlation between the 2 variables. This is to be expected, as the more days a property spends on the market, the more likely the vendor is to offer a discount. However, a relatively modest value of 0.31 means that the vendors more often than not decide to maintain the advertised price or offer marginal discounts, even when a property spends a significant number of days on the market.
Negative correlation -0.04 and -0.03 of DoM and Discounting in relation to the Number of Sales (sample size used) is insignificant, which indirectly confirms that the sample size has no impact on calculation of DoM and Discounting metrics.
Does land exhibit the same distribution as houses and units?
Let’s now explore the correlation between DoM and Discounting for 3 of the property types in our dataset: houses, units and land. We intentionally excluded townhouses from this analysis as they account for a very small portion of sales recorded.
The chart below shows both the distribution (histogram) for each of the metrics as well as correlation plot per each property type. Let’s first look at the distributions of DoM, Discounting, Sales data for all suburbs on our list.
- Upper left quadrant (DoM): Although the DoM distribution is similar for units and houses, land sales are more skewed to the right, meaning that land listings tend to stay on the market longer. Also units seem to spend less number of days on the market than houses, albeit in just a tiny fraction of suburbs.
- Center quadrant (Discounting): Houses and units exhibit a similar pattern, except a small bump for units in the negative region. It seems some suburbs are more likely to have units sold for more than the advertised price. Land sales offer significantly lower discounts than house and unit sales and are more skewed to the left, overlapping the negative region more than the other 2 property types.
- Bottom right quadrant (No of Sales): This chart indicates that the sample size for units and houses is similar for all of the suburbs on our list. There are a lot more land sales recorded than those of houses and units.
Let’s now deep dive into the correlation analysis for each property type. The 3 bottom left quadrants plot the correlation chart for the 3 variables in our report: DoM, Discounting and Number of Sales. We will focus on the middle(vertical) left quadrant which plots the DoM to Discounting relationship.
You will notice the blue(houses), orange(units) and green(land) lines plotting linear regression for DoM and Discounting correlation. We can see that DoM and Discounting correlation is more pronounced for houses than for units i.e. the blue line has a blunter angle. For reference a 1.00 correlation would look like a perfect 45-degree diagonal line.
As expected, because the land sales offer smaller discounts and spend more time on the market than other property types, the green(land) regression line appears slightly below than that of houses and units.
Aggregated data for capital cities
To conclude the analysis, we have prepared a summary of the 2 metrics aggregated at the city level per each property type. Perth units spent the longest time on the market, while Canberra houses sold the quickest in 2020 Q3. Not surprisingly Perth unit vendors also offered the highest discount, whereas Canberra townhouses were sold with the smallest average discount recorded.
|City||Property Type||DoM||Discount %|
I hope you enjoyed reading this short but insightful analysis. I certainly had a lot of fun deriving meaning from the new raw data HtAG collected over the last 3 months. Whereas the data paints one story, there are always exceptions and deviations from the norm. I invite you to share your insights via the comment form below.
The analysis left me with some questions. Can you help me anwer them?
- Why does land take longer to sell?
- What’s special about Canberra townhouses?
- Why do more suburbs have negative Discounting for units than for houses i.e. the left bump on the Discounting histogram for units?
Please leave your answer via the comment form below.