An algorithm that’s about to disrupt the planning, real estate and property development industries

George Giannakodakis
February 26, 2017

Algorithms and artificial intelligence are changing the way we do business. Companies like Google, Uber, WhatsApp and Alibaba clearly show that algorithms can disrupt an entire industry. So, what about the real estate and property development industry? Why haven’t we seen anything yet? One reason is that the feasibility process required when assessing a property for subdivision (STCC –‘Subject To Council Consent’), to determine upside redevelopment value, is a complex and daunting task. This ‘due diligence’ process can take a considerable amount of time and money, a luxury that most people don’t have before deciding to purchase.

Now let’s take stock of what is required in a typical residential property development decision making process before buying a property. Or you can jump to the solution below!

To begin with an obvious first question: Can I subdivide the property? How many homes can I, or should I build? That often requires an understanding of complex Council rules and a reading of inch thick Planning Schemes to figure out setbacks, percentage open space, plot ratios and the like.

The next question you might ask yourself is what style of dwelling should I build? Detached or semi-detached house? Townhouses or group dwellings? What does the market in this area demand? Is there a market optimum style and size?

A smart developer will understand that the process of sub-division itself increases land value; the location of the property (beach, leafy suburb, near a school or good public transport) influences the amount of upside or improved value. Developers also know that the act of building the dwelling is merely a vehicle to deliver this improved end land value, albeit there are some margins to be made out of the building process.

Then come the builder’s spreadsheets. Local estimator’s rates, pre and post-building costs, professional and council fees just to name a few of the hundreds of variables at play. The hard part is selecting a building standard level (low or high end?), to extract the most profit out of your land. It’s a tricky balancing act whether you are an experienced builder or first time developer.

Buying the property for the right price is also challenging because most automated valuation services are only a guide and reflect the seller’s ideal price. While there are varying levels of sophistication at this end, for instance third party sites from CoreLogic and banks, to House Canary at the algorithm end, the development process demands that a buyer’s emotion is parked because it is the end profit that counts.

The end profit highly depends on how much you can sell a ‘house and land’ package for. At this point you might engage the services of a local real estate agent to find comparable products and sales in the same street or suburb. But that can be misleading because the product you are selling will almost always be different in size, shape, and quality to local sales examples.

Finally, you might consider setting a minimum end profit target, say 20%, but only if all the ducks are lined up. After all of that you might decide to pull out. But what about all the other development scenarios? What if 3 townhouses as opposed to 2 double-storey detached homes deliver a higher profit outcome? That might require you to start the entire process all over again…

PointData is an Australian based start-up company that has spent the last 3 years developing an Algorithm that does all the above and more. Originally developed for Governments to be able to test planning policy scenarios and metropolitan utility capacity (see ‘Urban Policy Tool’ at www.rdatool.com.au ), the algorithm is currently being commercially developed and trialled for use across all Australian cities. Its self-learning intelligence coordinates several million data points, and formulas via an iterative process using innovative technology. The algorithm instantly cuts through ‘Big Data’ and the fog of planning rules to assess development potential and then repeats the process from start to finish until all development options are exhausted.

Most property analysis is based on historic data and linear forecasting. And most investors apply indicators such as rental yield and median house price movements to help them with their purchase, development, and sales decisions. PointData have coined a new term ‘Metro-Dynamics’ to describe the science of the ever-changing demographic and investment patterns at a suburb and metropolitan wide level; that have a fundamental impact on property values. Machine learning methods are applied to observe these patterns dynamically over time. Investment in a new park, rezoning, street trees, new infrastructure such as a tram, or simply redirected market interest, can change property demand in an area, which is often tempered by the rate of housing product supply.

Now imagine all that power reflected in a simple to follow feasibility report emailed directly to your account within minutes. Or better still the highest profit option being viewed on your phone within seconds of walking into a property that is for sale.

As Australian populations increase, so does the demand for residential development opportunities to fulfil the growing appetite for middle suburban and inner city living next to main streets and good public transport. PointData aims to have its algorithm undertake the due diligence on all 2.6 million development opportunities across 5 Australian cities by the end of 2017. Adelaide and Melbourne are well underway. For more information go to www.pointdata.com.au.