SecondFunnel is a collection of related applications for converting browser into shoppers. Additional information on the project can be found in the wiki.

See the super useful Setup Environment wiki.

The SecondFunnel project is broken into different folders for the application.

• ansible: Deployment and Provisioning of server code
• secondfunnel: Where settings and site-wide URLs are recorded
• apps: Applications and common code
  • api2: The base API used by various of our front-end services
  • assets: Common models and functions used across applications
  • dashboard: A dashboard to manage & review analytics for pages
  • imageservice: All management of images across servers and services
  • intentrank: An algorithm for ordering content in pages
  • light: The front-end pages
  • scrapy: A scraper to update product & content from websites & datafeeds
  • tracking: A web tracking pixel loader
  • utils: A whole whack of utilities
  • website: All thats left is a redirect to Squarespace
• scripts + fabfile: Python and Bash scripts for deployment and/or utility

The SecondFunnel project has a few primary components:

• API2: The backbone of the SecondFunnel project; provides an API for requesting and serving content.
• Light (Pages): The front-end javascript; manages how content is arranged on the discovery page or ad, services API queries to IntentRank and handles interactions between the user and the pages.
• IntentRank: Provides Pages with products and content (in the form of "Tiles").
• Scrapy: The system for keeping products and content continually updated.

The SecondFunnel API2 is a RESTful list of resources, for internal use only. Currently it is primarily used by the dashboard

IntentRank is the main recommendation system, and controls the output of the feed (ordering, and what tiles (content/products) are shown). It is the foundation for analyzing users and determining what the user is most interested in based on various parameters.

Pages is the front-end javascript of the SecondFunnel project that manages how content is displayed on the Light pages. In addition, it manages how users are able to interact with those pages; by clicking, scrolling, social media, etc. It API calls out to IntentRank to fetch tiles based on a myriad of factors dependent on how the user is currently interacting with the page. It makes use of the popular Masonry Library to render content in a cascading infinite scroll grid. As users interact with the page, it also records analytic data.

Scrapy is our system for keep pages updated. There are two types of updates: datafeed updates and page scraping updates. Datafeed updates are fast, but usually some of the data (pricing & availability) is 24 hours out of date and the pictures are limited and low-res. Page scraping updates are really slow, but provide great data and images.

Copyright 2012-2016, Willet Inc.