Turns out you can’t use algorithms in property investment
Theoretically you should be able to use algorithms in property, right? After all, house prices, location, demand and property details are all quantifiable information that can be processed and should be able to be used to make predictions about the future.
Algorithms, by and large, are pretty good at making predictions based on solid mathematical data, but surprisingly hit and miss when it comes to human behaviour.
Let’s take some examples – PageRank is, arguably, the most used algorithm in the world today. It’s the basis of how Google ranks pages and websites in their search engine. The test of its usefulness and popularity is in the fact you’ve probably already used it numerous times just today as well as being one of the world’s most valuable companies.
On the opposing end of the scale would be the algorithms that social media sites and the likes of amazon use to recommend your products. Sure, sometimes you’ll get some stuff that’s fairly accurate, however, it’s far more likely you’re going to be recommended cat beds despite hating cats.
That leads us neatly into the once successful market of US house flipping companies that tried to completely digitise their buying process.
Machines vs human behaviour
In an analysis for Wired, they took a look at the activity of some of the top house flipping companies that had centred their activity around Phoenix, Arizona, due to the identikit nature of housing in the city.
Most of the houses there are conventional cookie-cutter homes which makes them much easier to price and predict with previous valuations and sale-price data. The more similar the properties, the easier it is to use data usefully.
For 3 years, it worked fairly well, however, things changed during the pandemic. As described in the analysis “forecasts proved inaccurate in 2021’s gyrating housing market. In the second quarter Zillow actually was able to sell homes for 5.8 percent more than it expected. In the third quarter, though, Zillow sold homes for 5 to 7 percent less than it forecast.”
The algorithms actually didn’t see the surge in demand post-lockdowns and didn’t see a collapse in demand in certain areas due to wide scale changes in working practices, with many switching to work from home.
As it turns out then, computers are quite bad at predicting where people will want to live and where demand will shift to.
For investors
So as current or potential property investors what are we to take away from this?
Human behaviour and demand are both really difficult to quantify into computer code, and in most cases successful property investors either use their instinct and experience or defer it to experts and advisors.
We know that when we advise clients that they’re coming to us for expert advice, experience and guidance and that’s why we have so many successful investors on our books that come back to us time and again.
Property, certainly in the UK, is one of the most reliable and profitable investments around but, like any investment, there are always risks. The sensible move is to mitigate that risk by either educating yourself or seeking expert advice, and we know from our own experience that our most successful investors follow this mantra.
