Many people are shopping on Wayfair at the same time. How can we detect real time shopping patterns and send targeted notifications to users so they get useful social proof that helps them moves down the funnel?
What it does
Listens for real time events as people shop. Events like PDP views, add to cart and purchases. An event mining service pulls pattern from this data and sends notification for popular SKUs when people are taking action on that SKU like browsing, adding to cart etc.
How we built it
Used scribe events and built a web-service to pull pattern from event data. PDP makes an ajax request to this service to check if this SKU wins the popularity contest.
Challenges we ran into
Deciding storage for real-time data. Debated different ways to mine the information for best effectiveness.
Accomplishments that we're proud of
Our service works real-time in dev
What we learned
This is very useful feature. Keeping track of real-time data is hard.
What's next for Real-Time Social Proof
We need much better mining services, perhaps use machine learning to discover patterns.
Try browsing SKUs in dev on ssonawane's dev env.