As an agent, do you often find it difficult to track the value of your client’s homes? Perhaps sometimes, you feel it’s a hassle to always put your clients on hold so you can refer to past transaction records.
In any sales conversation, time is essential and convincing your clients how much they can potentially sell their house for quickly can often make or break the deal for you.
While pinning down the exact value of a property is difficult, our new X-Value tool leverages our deep bank of data to help you get a figure quickly. SRX, a subsidiary of 99 Group, was the first to introduce this concept to the industry since 2014. Since then, it’s become a reliable point of reference of a property’s value after certified appraisals.
What is X-Value?
X-Value automates property valuation calculations, based on different machine learning models which observe real estate data patterns through time, location and property attributes. In simple terms, it’s an artificial intelligence with access to a huge amount of historical and real-time real estate data.
Through complex calculations, it is able to provide a near-accurate value of a specific property in Singapore based on its postal address – serving as a quick, convenient and reliable point of reference for prospective buyers and sellers.
Why X-Value?
Short of paying for a certified appraisal and the time that would take, X-Value is a quick method to have an inkling of a property’s value on the fly without any additional charge.
This is critical for an agent like yourself, especially when you’re engaging with a client on the phone while on the move. Instead of making your client wait, you can quickly access X-Value on your mobile device and strike a confident conversation with your client, while preparing to deep-dive into further details.
How accurate is the X-Value?
Using various sources of real estate data across time, location and property attributes, our machine learning models are configured to remain relevant to current property price dynamics.
You can be assured the value you’re seeing is updated and true at the time of generation.
It is important to note that X-Value is a price estimate based on historical transactions, location, time and property attributes. We know that there are many variables when it entails predicting a price, or a point of reference, for a particular asset class.
In simple terms, we know the following makes sense:
- Higher floor (non-landed) => higher cost
- Bigger floor area => higher cost
- Newer house => higher cost
- Nearer to amenities (eg. MRT) => higher cost
Here’s an illustration to reflect this:
So when there are many variables, the probability of two properties being identical, or identically priced, is usually low.
Over the years of training its machine learning models, SRX has been able to compile its X-Value performance to be as accurate as possible. In other words, as much as possible, we aspire for X-Value to inch closer to the final transacted price for the same property.
From the histogram comparing X-Value vs the prices of past transactions of Resale HDB and Condominiums, it is clear that the majority of transactions are centred around 0% difference. This means that the difference between X-Value and the closing price is almost the same. For Resale HDB, the median percentage difference is only 2.8%. For Condominiums, it is even better at 2.1%.
How is X-Value constantly enhanced?
One method used to enhance X-Value is via backtesting. Backtesting assesses how well a strategy or model is when tested with unseen data (ie. no recent, comparable transactions).
The X-value is constantly backtested with historical data at every training and model-tuning step to ensure that it maintains or improves from its previous best performance.
In backtesting, we assess X-value predictions of properties against their true historical values (while ensuring that we do not have target leakage).
For example, if a particular property in our historical record was transacted for $500,000, an X-Value that perfectly predicts the price as $500,000 will yield a difference of 0%. If the X-Value over-values its prediction by $100,000 (i.e. at $600,000), this will yield a difference of +20%. Conversely, if it under-values it by $100,000 (i.e. at $400,000), this is a difference of -20%.
For all properties that we backtested, we recorded the X-Value percentage difference with the true value and plotted them as a histogram (as shown above).
Can an agent increase or decrease the asking price based on X-Value?
With the price deviation shared earlier, an agent can potentially use that to derive a target price range above and below X-Value.
–
As we continue to train and improve the X-Value machine learning models, additional datasets (like HDB Minimum Occupation Period) will be introduced to better bring the deviations lower.
For an agent, it will give you greater trust and confidence when you communicate the potential value of a property to your prospective homebuyers and sellers.
So do sign-up and track your clients’ properties X-Value for free today. Being on the cutting-edge of knowing a property’s potential value might just give you the advantage you need to satisfy a client’s request and seal that all-important deal.
The post X-Value: An agent’s free tool to track a client’s property value appeared first on 99.co.
