Deep Data at Macy's - Searching Hierarchichal Documents for eCommerce Merchandising: Presnted by Peter Gazaryan, Macys.com & Eugene Steinberg, Grid Dynamics Inc.
3 likes3,911 views
This document summarizes a presentation given at a conference about search capabilities at Macys.com. It discusses the history of Macys and the evolution of search at Macys.com, including the implementation of a new Solr-based search engine in 2013. It provides details on features for customers, such as natural relevance, filtering, and personalization, as well as merchandiser-facing features like rule-driven results, dynamic product grouping, and tiered natural scoring. The presentation emphasizes the importance of merchandising and precision for ecommerce search.
Deep Data at Macy's - Searching Hierarchichal Documents for eCommerce Merchandising: Presnted by Peter Gazaryan, Macys.com & Eugene Steinberg, Grid Dynamics Inc.
- 1. O C T O B E R 1 3 - 1 6 , 2 0 1 6 • A U S T I N , T X
- 2. Deep Data at Macys Searching Hierarchical Documents for eCommerce Merchandising Denis Kamotsky, Macys.com Eugene Steinberg, Grid Dynamics Peter Gazaryan, Macys.com
- 4. 4 02 The Macys Story v 1858 Entrepreneur R.H. Macy opens R.H. Macy & Company, a small dry goods store. v 1902 R.H. Macy & Co. moves uptown to Herald Square and shortens its name to Macy’s. v 1924 Macy’s employees march from 145th Street to 34th Street to celebrate Thanksgiving, which sparks an annual tradition. v 1976 Macy’s sponsors the first annual Macy’s Fireworks, now a 4th of July tradition. v 1994 Federated Department Stores, the largest operator of department stores, acquires Macy’s. v 1998 Macys.com is launched and operates out of New York and San Francisco. v 2006 Macy’s expands to over 800 locations across the U.S
- 5. 5 03 The Macys.com Story v 1997 MCOM operates out of San Francisco, California and Brooklyn, New York. v 1998 MCOM is officially launched. v 2001 New York offices moves from Brooklyn to 1440 Broadway in Manhattan. v 2001 Macy’s By Mail catalogue business shuts down, making macys.com the sole provider. v 2010 We reach one billion dollars in annual sales volume. v 2013 Macys.com launches Keyword Search running on Apache Solr. v 2013 We reach two billion dollars in annual sales volume.
- 6. 6 01 History of Search Engine at Macys.com 2015 Dec 2012 Aug Dec 2013 Aug Dec 2014 Aug Dec Merchandising Management Tool becomes available to the Users 3/2013 Solr-‐Based Keyword Search Engine goes live on macys.com 4/2013 Solr-‐Based Type-‐Ahead autocomplete funcIonality goes live on macys.com 9/2013 Legacy Keyword Search Engine is reIred 10/2013 Management of the Category Browse funcIonality becomes available in the Saturn Tool 3/2014 Category Browse is fully migrated to Solr on macys.com 8/2014 Intra-‐day SKU availability updates go live 5/2014 Dynamic grouping and ungrouping of Product CollecIons is live on macys.com 1/2015 Solr Re-‐PlaQorm effort begins 12/2011
- 7. 7 01 Solr Delivers Result session conversion increase attributed to migration to new Solr-based Search engine28% customers clicking on the ungrouped collection convert higher6% macys.com traffic is Solr-Powered Keyword Search and Category Browse queries 8% increased conversion in type-ahead autocomplete sessions 35%
- 8. 8 01 Types of Retail Boutique Mall Big Box Specialty Department
- 9. 9 01 Types of E-Retail Inventory volume Curation
- 11. 11 01 Great Expectations Mary the Shopper Alice the Merchandizer
- 12. 12 01 Dreaded “Relevancy Tuning” One size doesn’t fit all, even if you can stretch it
- 13. 13 01 Customer facing features Quality • Natural relevance • Data quality • Concept search Refinement • Range, hierarchical facets • Sorting, grouping • Guided navigation Usability • Input methods • Presentation and productivity • Device form factors Targeting • Geography, demographics • Personalization • Collaborative shopping
- 15. 15 01 Structured Catalog and the Quest of Precise Filtration: Part 1
- 16. 16 01 Structured Catalog and the Quest of Precise Filtration: Part 2
- 17. 17 01 Structured Catalog and the Quest of Precise Filtration: Part 3
- 18. 18 01 Concept Search and Controlled Precision Reduction
- 19. 19 01 Concept Structured Search Under the Hood
- 20. 20 01 Merchandiser facing features Curation • Rule-driven results • Product categorization • Date and time-based campaigns Bias • Coarse natural scoring • Metric-driven boosting • Context-based placement Comprehensiveness • Omnichannel data • Real-time availability • Accurate pricing Referring • Cross-sells, up-sells, recommendations • Predictive search • “Did you mean”, “do not carry” suggestions
- 21. 21 01 Dynamic Grouping and Ungrouping
- 22. 22 01 Dynamic Grouping and Ungrouping: Under The Hood catalog = SELECT * FROM sku JOIN product JOIN collec=on SEARCH "DKNY" FROM catalog GROUP BY collec=on.id WHEN ALL(product.brand="DKNY") SEARCH "white cup" FROM catalog GROUP BY collec=on.id WHEN COUNT(product.id)>collec=on.threshold SEARCH ”dinnerware cup" FROM catalog GROUP BY collec=on.id WHEN NOT EXISTS(product.id)
- 24. 24 01 Rule Driven Query Rewriting
- 25. 25 01 Make The Magic of Macys! Inspiring Story o Great traditions o Versatility of the business model o Merchandising as a key differentiator Genuine Innovation o High Precision Concept Search o Controlled Precision Reduction o Tiered Natural Scoring o … and much more Challenging projects ahead o Omnichannel integration o All flavors of scalability on omnichannel data Technical challenges hazard
- 26. Thank you! Q&A