The Singapore-founded company wants to become a ‘property trust platform’ and one-stop-shop for property hunters. A.I. will power its quest for success
When PropertyGuru introduced its home finance platform, PropertyGuru Finance, in March 2020, it felt like a culmination of its evolution from a property search platform to a one-stop-shop for property hunters. It had partnered with several banks to offer digital home financing services including instant in-principle approval, instant loan offers, and refinance checks, to enable buyers to consume home financing services conveniently, securely and quickly.
The development took the company one step closer to achieving its vision of becoming a ‘property trust platform’ rather than just a property search platform. Business Intelligence and Artificial Intelligence (A.I.) solutions had enabled the platform to draw in property seekers and agents and make the online property search a successful business.
However, several hurdles remain, including streamlining due diligence processes, cutting through silos across networks, resolving interoperability issues across platforms, and enabling solutions to speed up verification processes.
What other solutions could PropertyGuru build to provide end-to-end services through the online channel, for the home buying process? What emerging technologies could it tap on? Was its fully digital property pursuit a realisable goal, or an elusive dream?
Keeping an eye on the data
From its beginning in 2007 as a Singapore-based startup, PropertyGuru now attracts over 24 million property seekers every month with more than 2.7 million listings in five countries across Southeast Asia.
Its initial growth strategy was simple: acquire listings, build consumer traffic, and monetise the business model. By focusing on data that was generated by the platform on a daily basis, PropertyGuru was able to fine tune the user experience to increase consumer traffic. The strategy, according to the company’s Chief Technology Officer (CTO) Manav Kamboj, was “to attract more and more customers to the platform by building supplier volume”.
By catering to the needs of the three customer segments – property seekers, agents, and developers – the company was able to triple its revenue by 2011, doubling its traffic in the process. By studying its home market of Singapore, Kamboj notes that the average property seeker took an average of seven months to complete the journey of starting the search process to closing a transaction.
It also provided a treasure trove of data to improve the product.
“Our first focus in our data analytics journey was to understand consumer behaviour,” Kamboj explains. “What are our consumers looking for, what are the triggers of their purchase behaviour? We then analysed what we could do differently to help them find their property in a more efficient way.
“We implemented Google analytics; we then exported our analytics data into a data warehouse. Looking at that data and backing it with consumer research, we were able to build solutions to help make the search process more efficient. For example, a typical rental search process-cycle takes three to four weeks, whereas a purchase process takes six to eight months.
“We looked at how we could support the customers during this journey; we tried to help them shortlist the property faster by making the search results more targeted.”
While bringing in property seekers was important, there would be no transactions or traffic if agents did not benefit from listing their units on PropertyGuru. Through its AgentNet portal, to which all agent subscribers are given access, agents can secure information and customise analytics reports regarding an individual property to help them secure a sale or understand general market trends.
“To come up with our solutions, we looked at agent motivations,” notes Kamboj. “What can be an agent's next best action? What do his data tell me? Using simple and sometimes more advanced tools, we tried to suggest to them their next best action.
“The objective was to make them more successful. If the agent is not successful, the agent will not renew with us. So it makes sense for us to ensure that we provide agents with the right tools to succeed.”
Keeping A.I. on the data
That strategy worked a treat: By 2014, the PropertyGuru platform counted 30,000 agent subscribers in Singapore alone, and it was crunching about 200,000 data points per month. As data collected from users multiplied, PropertyGuru expanded its technical repertoire to introduce A.I.-based solutions. The first step towards making the platform A.I.-driven was to clean the data, ensuring that the database did not suffer from duplicates, illegal characters, and outliers.
When buying a house, a seeker would typically look for proximity to nearby amenities like schools, grocery stores, pharmacies, hospitals, clinics and public parks etc. Geocoding modules were built into the platform to help users search for such amenities within a specified distance around a selected property.
Another important consideration in property search for seekers was commute time to work or school. To allow users to determine a property's accessibility to facilities, the team also built a network module into the platform that could determine driving directions and trip duration using historical traffic information.
What now?
All this was done to achieve its vision of being a property trust platform, to which end PropertyGuru invested heavily in its Data Science capabilities. It established a Data Science team and hired top data scientists and practitioners. It shifted gears from focusing on basic business intelligence to more advanced Data Science techniques to meet growing customer needs. Basic predictive techniques such as regression analysis, done on samples of past data, had given some estimates of the future trends.
Along with the move to become a one-stop-shop, there were several key questions that now required careful consideration. What had been PropertyGuru's key value proposition? How had Business Intelligence and A.I. contributed to the company's growth and success so far? And how could A.I. help further accelerate the growth of the company and help it achieve its vision?
This is an adapted version of the SMU Case, "PropertyGuru: Driving AI Powered Real Estate”. To see the full case, please click on the following link: https://cmp.smu.edu.sg/case/4466”.
Follow us on Twitter (@sgsmuperspectiv) or like us on Facebook (https://www.facebook.com/PerspectivesAtSMU)
Last updated on 31 Aug 2020 .
