1. Response to
Request for Clarification
Surpassing Pinterest’s
Visual Discovery
- From Curation to Algorithm -
Mark Lee
Nov 2012
Private & Confidential
3. Learning from History
Web 1.0 Discovery
Human Curation vs. Algorithmic Search
Curation Algorithm
1994 1998
Using Partner Search Algorithms
2000: Google; 2009 Bing
Will the Curation -> Algorithm
Historical Pattern Repeat Itself?
Will Algorithms
Eventually Dominate Visual Discovery?
Private & Confidential
4. Anglo Digital’s Visual Recommendation Engine
Uses Algorithms for Visual Discovery
Excerpt
from the 2010/2011
Application
We Can Use
Algorithms to Surpass
Pinterest
for Visual Discovery
Private & Confidential
5. Our Algorithms and Superior Functionality
Surpass Pinterest’s Value Proposition
Consumer Need Anglo Digital Pinterest
Visual Discovery Algorithm Curation
Superior TRI/ID Functionality Delivers a Superior Interactive Customer Experience
User Can
Digital Try Not Available
Digitally Try Product
User Can
Digital Restyle Restyle Selection to Not Available
Better Match Desire
User Gets
Digital Iterate Superior Visual Fit Not Available
Recommendations
Users Can Iteratively Try and Restyle Selections
to get Superior Visual Fit Recommendations
Private & Confidential
8. Curation versus Algorithm for Discovery
(Readings)
The primary difficulty of Web 1.0 was one of information overload.
Google’s solution to this problem was to improve search.
Yahoo’s solution to this problem was to improve curated directories
http://thoughtfaucet.com/strategy/insight/pattern-recognition-last-years-battle/ .
Yahoo!’s human approach seemed quaint and largely irrelevant and sure enough they
eventually discontinued that approach, hired mathematicians of their own and competed
with Google on an algorithmic basis.
http://www.choicestream.com/curated-vs-algorithmic-brand-safety/
Private & Confidential
9. The Visual Recommendation Engine 2010/2011
Slide
Empowers the Customers
Sell Better
TRI/ID* Tool
Customer Picture
Try Restyle
1st Rec
Anglo Visual
Recommendation 2nd+ Rec
Digital Engine
Data Driven Insights Restyling
Feedback
Databases
• Product (mfg cost, time) Iterate Loop
• Customer (prefs, contact) Until Customer
is Delighted
Improvement Input for Customer Process of
Add’l Recommendations *Try-Restyle-Improve/
Of Existing Products Iterate until Delighted