Product owners for online real estate services are facing new challenges in a market that is currently in a state of massive disruption. First time home buyers entering the market grew up with advanced technology, allowing everything to be at their fingertips. Buyers now expect to be able to browse homes from their phones and, for the home listing services, provide options that fit the specific criteria they’re looking for.
To create an intelligent feed that displays relevant homes to viewers, there are several steps that need to be taken. Five to be exact.
1. Capture Explicit User Preferences
This is the basis of doing any kind of personalized feed. The service needs to ask the user for criteria on what they consider to be acceptable in a home. Obvious criteria or filters include location, price, size, number of bedrooms/bathrooms, garage size, etc. Most real estate sites today have the ability to filter based on these options. Something to consider is to present the user with a “deal breaker / must have” toggle. Having this toggle, or some kind of sliding scale on how strong of a preference it is, will be fantastic input for your personalization engine later on.
2. Track Implicit User Behavior
As your users interact with the platform, be sure to track their actions and record those analytics. Some behaviors will indicate stronger preferences than others. For instance, sharing a home listing with a spouse is a much stronger indicator than viewing the first 3 photos of a home and moving on.
3. Compare to Other Users
As the models of user preferences are built up and evolve, you will find users who are similar to each other. Once you’ve identified these similar users, you can leverage their actions and the homes they’re viewing to add options for their counterparts.
4. Rinse, Repeat
The personalization engine evolves over time and requires constant input as well as tweaking. As more and more data is fed into the system, new insights will be gained and profiles will morph. This feedback loop of both positively reinforcing and negatively detracting behaviors will help the model tune itself and become better at recommending homes for each user to view.
5. Bringing It All Together
Once steps 1-4 are in place, the real estate platform will be well positioned to capture a market of users that expect to be able to self-serve. This platform will present the most relevant homes to each user, being sure not to show homes that are irrelevant and have non-starters (ie. only 1 bathroom when they actually need 3).
Users of an intelligent, highly personalized real estate feed are more likely to remain engaged, spread the word and ultimately find their dream home. There are many real estate companies out there, so providing your clients and prospects with the best user experience is crucial. To gain a broader understanding of the benefits of feed technology and how you can implement them, pre-register for our eBook or contact us with your questions!
Also published on Medium.