Real-estate portals versus e-commerce websites
Website visitors and online customers are more and more expecting personalized content when visiting an online store. Indeed, with personalization the overall customer experience is improved, visitors find relevant items faster and easier and they will more likely come back.
This is why most e-commerce websites have implemented website personalization projects targeting not only visitors’ conversion (visitors turn to customers) but also customers retention (customers who will buy again).
Indeed customer acquisition comes with high cost (advertising, special offers…) and gaining the loyalty of the acquired customers has a positive P&L impact.
Online stores ignoring this trend will be left out of game.
What about real estate websites?
Compared to traditional e-commerce sites real estate websites are facing unique challenges:
- Products stay less time in catalog, usually the lifespan of a property for sale or rent is between 20 to 80 days, while products can stay several month or years in the e-commerce catalog
- Products are unique and users make a one-time purchase
- There are much less interactions between users and items
- The cold start problem which happens when a new user or a new item are introduced is the rule
The challenges above create specific requirements to tech solutions, for example an e-commerce recommendation engine will most probably underperform on real estate portal.
However, real-estate websites will boost their revenues with website personalization by:
- Increasing the number of leads: visitors find quickly properties matching their requirements and taste
- Automatically qualifying leads: identify ‘only curious’ from ‘potential buyers’
- Shortening the sales cycle: with qualified and targeted customers the sale’ closing is quicker
What are the solution enablers to fully benefit from real-estate portal personnalization ?
Smart search, NLP based
The first thing a visitor does when browsing a real-estate portal is to find properties matching with his requirements (surface, number of rooms, price…).
Offering a flexible full text search like: “New apartment with 2 bedrooms, balcony and garage near Liberty square” helps visitors to quickly find properties matching their requirements.
Compare to a search based on pre-defined filters a natural language search enables to discover new requirements and trends, not yet reflected in the list of filters.
For example since the Covid new search terms have appeared like “home-office”.
Natural language search allows to identify most demanding or trendy features like “garden”, “quiet”… It guides searching for new properties to add to the catalog or adjusting dynamically the pricing.
Then visitors are inviting to provide an explicit feed-back about the search results (a rating or just like/don’t like). Such ratings will feed the recommender engine and will unlock personalized suggestions.
Personalized recommendations
Based on visitors’ interactions and explicit feed-back, the recommender engine will be able to find properties that not only matches customer requirements but also customer taste and preferences.
Specific challenges shared by the Portal and the Solution provider
Data quality
The power of the data is most often untapped, however the data value depends largely of the quality of the data:
- Tracking meaningful visitor interactions (search, explicit and implicit feed-back) with timestamp
- Cleaning the listings and enriching the property description (categories, free text, pictures…)
Our experts will help you to conduct an exploratory data analysis (missing fields, duplicates…) and to put in place data collection strategies in compliance with GDPR.
Natural Language Processing (NLP) training
The NLP engine should be trained to the real-estate case (for example: residential/office, sale/rent…). The main work will be done by NEXiT, however data inputs and samples must be provided by the real-estate portal.
Recommender engine
NEXiT will select the best recommender algorithm(s) to match your business goals and to handle the real-estate specific challenges:
- Cold start
- Short listing lifespan
- Fewer customer interactions
- Unstructured and hidden data
At NEXiT we leverage AI to help businesses improving their products and operations. We deliver concrete benefits to our customers in the field of Recommender Systems (User privacy, Context Sensitive), Smart Search and forecasting solutions.
Tags: AI, Real-Estate, Website personnalization, Recommender systems, NLP