How Nordstrom Utilizes Human Intelligence to Blend Brick-and-mortar with Online Commerce

Erin Shellman PhD - Data Scientist - @erinshellman
David Von Lehman - Developer - @davidvlsea

How Nordstrom Utilizes Human Intelligence to Blend Brick-and-mortar with Online Commerce

Erin Shellman PhD - Data Scientist - @erinshellman
David Von Lehman - Developer - @davidvlsea

The Nordstrom Data Lab Mission

Delighting customers through data-driven products.


Nordstrom Data Lab Timeline

Multidisciplinary Team

Our Powers Combined

The Nordstrom Data Ecosystem

There has been a shift in our thinking. A part of our resources are now more human curated...

our engineers evolve the algorithm, and humans help us see if a suggested change is really an improvement.

- Scott Huffman
Engineering Director at Google

EASE

Emulate

Automate

Scale


Evaluate

Three Applications of EASE

  1. Our Stylists Suggest
  2. Recommendo API
  3. My Color Palette

Stylists at Nordstrom

Our Stylists Suggest

Our Stylists Suggest is a recommendation engine that analyzes transactions in Nordstrom stores that were facilitated by a personal stylist.

How to Emulate a stylist?


Stylist-based Market Basket Analysis

How to Emulate a stylist?


Stylist-based Market Basket Analysis

How to Emulate a stylist?


Stylist-based Market Basket Analysis

How to Emulate a stylist?


Stylist-based Market Basket Analysis

How to Emulate a stylist?


Stylist-based Market Basket Analysis

Evaluate


Test internally with a light-weight prototype


Compare to incumbent

Automate


Evaluate


Introducing Recommendo

How'd It Go?


  • 2 software engineers, 2 data scientists, 4.5 months, and a corporate card

  • Live in production as of Nov 24, 3 billion+ recommendations served

  • Full self-service infrastructure in the cloud

  • Continous delivery/deployment from day 1

  • Next up - more personalized, size + brand

  • If you build it, they will come

Preview Tool

Cloud Monitoring

Lessons Learned


  • Quick prototyping is the fastest way to internal advocacy.

  • We don't always need a complicated solution.

  • Play to your differentiating strengths.

  • Bias towards impact.

  • Cloud == Speed

An Exploration of Color

Pixels as Data

Photo Studio Expertise

Color Extraction

Color Extraction

Color Extraction

Color Extraction

X11 Color Space

Color Scoring

Color Patterns

Colors of the Season

Swimwear Trends by City

Top Products by Color Score

Precise Color Match

Pantone Bump

Customer Color Fingerprints

Empower our Customers with Data


What are the colors I buy?

What are the colors I should try?

Wrap Up

  • It Takes a Village
  • Experience >> Data
  • KISS
  • Ship It!
  • EASE!!

Acknowledgements

Sara Hogenson, Paul Payne and the Nordstrom Innovation Lab


Scott Jones, Prashanth Nair, Abid Saifee and the Advanced Analytics, Personalization, and Mobile Teams, @nordysanda

Thank You!

Questions? Ask ds@nordstrom.com


This Presentation
nordstrom.github.io/stratanyc


Nordstrom Technology is Hiring
techcareers.nordstrom.com


Nordstrom Shoebox
www.nordstromshoebox.com

Attributions