AI - Track browsing behaviour & Recommendations
Budget: $30 – $250 USD
Step 1 - Create a basic prototype of tracking each user's browsing behavior (by of end of April)
Step 2 - Create profiles for the user based on behavior (eg. from ads viewed, 90% looking for apartments, 65% in the 30-50M range, 60% were ready to move in, 80% 3 bedrooms, 100% within Colombo). Therefore user is someone looking for 3 bedroom completed apartments in Colombo within 30-50M range
Step 3 - Identify which properties were viewed by similar profiles
Step 4 - Suggest properties to the user based on the past behavior (current or past sessions) and properties viewed by similar profiles
Step 5 - Include user's browsing profile in to the Customer profile tab in Leads sheet ✓ Leads sheet - Customer List
Recommendation API
1) Create a Recommendation model as described here (https://cloud.google.com/retail/recommendations-ai/docs/create-models?&_ga=2.11457524.-135274068.1629379841#import-reqs, https://cloud.google.com/retail/recommendations-ai/docs/overview), including:
1.1) Uploading catalogue (Upload Ads through the API. After initial bulk upload, upload new ones and remove old ones once a day via API)
1.2) Setup Tag manager to record user ad views and 'show-tel' clicks plus form submissions (https://support.google.com/tagmanager/topic/7679108, https://support.google.com/tagmanager/answer/7679219)
1.3) Create Model ('Others you may Like')
1.4) Create placement for recommendations (https://cloud.google.com/retail/recommendations-ai/docs/manage-placements#create)
AWS ML
https://aws.amazon.com/sagemaker/?p=ft&c=ml&c=3
https://techblog.realtor.com/evolving-personalized-recommendations-using-match-score/
https://techblog.realtor.com/personalized-recommended-homes/
https://techblog.realtor.com/user-behavioral-profile-as-a-building-block-in-ml-feature-store/#more-602
https://techblog.realtor.com/predicting-home-buyer-stage/#more-576
Step 2 - Create profiles for the user based on behavior (eg. from ads viewed, 90% looking for apartments, 65% in the 30-50M range, 60% were ready to move in, 80% 3 bedrooms, 100% within Colombo). Therefore user is someone looking for 3 bedroom completed apartments in Colombo within 30-50M range
Step 3 - Identify which properties were viewed by similar profiles
Step 4 - Suggest properties to the user based on the past behavior (current or past sessions) and properties viewed by similar profiles
Step 5 - Include user's browsing profile in to the Customer profile tab in Leads sheet ✓ Leads sheet - Customer List
Recommendation API
1) Create a Recommendation model as described here (https://cloud.google.com/retail/recommendations-ai/docs/create-models?&_ga=2.11457524.-135274068.1629379841#import-reqs, https://cloud.google.com/retail/recommendations-ai/docs/overview), including:
1.1) Uploading catalogue (Upload Ads through the API. After initial bulk upload, upload new ones and remove old ones once a day via API)
1.2) Setup Tag manager to record user ad views and 'show-tel' clicks plus form submissions (https://support.google.com/tagmanager/topic/7679108, https://support.google.com/tagmanager/answer/7679219)
1.3) Create Model ('Others you may Like')
1.4) Create placement for recommendations (https://cloud.google.com/retail/recommendations-ai/docs/manage-placements#create)
AWS ML
https://aws.amazon.com/sagemaker/?p=ft&c=ml&c=3
https://techblog.realtor.com/evolving-personalized-recommendations-using-match-score/
https://techblog.realtor.com/personalized-recommended-homes/
https://techblog.realtor.com/user-behavioral-profile-as-a-building-block-in-ml-feature-store/#more-602
https://techblog.realtor.com/predicting-home-buyer-stage/#more-576