Artificial intelligence has quickly become an integral part of property valuation around the globe. Nowadays most digital property portals have adopted the technology to provide users quick and seemingly “accurate” measurements of how much their property is worth.
However, it’s an important factor for prospective buyers, developers, investors, appraisers, tax assessors and other real estate market stakeholders to be sure that these predictions are accurate. But, with time and money not always on-side, these soft computing techniques with higher data handling capabilities have become an optimum choice in valuing property.
But how accurate are these algorithms, and should we really be relying on them?
Traditionally, calculating the overall worth of your investment was done through the sales comparison method, rental method, land and building method or the profit method by certified individuals. But these methods took time and money, and when it was discovered that computerised algorithms could produce similar(ish) values for half the time and money, it seemed like a no-brainer. These values, calculated by automated valuation models (or AVMs), are driven from online property records to pricing trends to recent sales of similar properties.
The trouble is, how accurate the data really is.
Let’s take a look at the buying story of Mark and Jane. Mark and Jane are a middle-aged couple living in Sydney, who are hoping to relocate to Perth. They’re looking for a family home in Kensington to raise their 2 children in, with a bit of character that doesn’t require much work, and is on a fairly large block. After 2 years of looking, their dream home pops up. A 1920s character, on a large 800 m2 block, big enough to fit their whole family – perfect! Without properly being able to look at the property due to living over east, Mark and Jane do their research on a number of online valuation sites, and talk to the Agent.
The selling agent instructs them that the price guide is roughly around $1.4M, however through doing their own research on these online valuation sites, all of which state that the home is worth between $1.2-$1.25M, right within their budget. Little do they know though, that two comparable properties on the same street sold off-market within the last 2 months, for a price of $1.37 & 1.42M.
However, Mark and Jane were confident in their own research based on results they received from the AVMs, and proceeded to submit an offer on the house at $1.25M, subject to finance. 24 hours pass and the selling agent calls to say they’ve received a cash offer, and it’s time to submit their best and final offer, which for Jane and Mark was $1.25M. They’ve missed out on the home.
Here’s where they went wrong. The couple who were selling the home had lived there for over 40 years, thus there was no up to date accurate data on the property. Fully renovated throughout, they had added an additional 2 bedrooms, had refurbished two bathrooms and the kitchen, and relandscaped to include a pool. The Selling Agent had also thankfully done their research about the two additional off-market sales on the street. All of these factors of which ultimately affected the final sale price, and would have never been picked up on in the online valuation. Why? Because as mentioned earlier, these algorithms are designed to pick up on online property records only.
Without knowing what improvements have been made to the property, or the circumstances of the last sale, the difference in actual value of the property can be hundreds of thousands of dollars.
Beware of making this mistake when buying/selling. Don’t be like Jane and Mark.
Contact us today for an up to date opinion on the current market.