How AI is transforming the 200-Year-Old Brooks Brothers
A Customer Success Story
ARTIFICIAL INTELLIGENCE FOR RETAIL Leading the Algorithmic Retailing Revolution
AI and Machine Learning Solutions optimizes processes and turns Data into Cash
ORS A.I. Platform Automate Decisions for maximizing performance ROI < 6 MONTHS
• Manage • Operate Source data • Accelerate • • Sensors and devices Analytics Profit • Enterprise systems Dashboard • Web based systems Data
Comprehensive, fully-integrated - straight-through processing ORS in a Nutshell
Optimizing Business through Artificial Intelligence since 1996
100+ Mathematicians & IT developers Applying Artificial Intelligence software for optimizing New York business processes TEAM Munich, end-to-end (E2E) Milan Lugano, MISSION OFFICES Tirana Alba
One Platform, Fortune 2000 companies in cross-process and - manufacturing cross-industry PRODUCTS CUSTOMERS - retail - services - financial industry www.ors.ai
5/7/19 CONFIDENTIAL 3 ORS Global Clientele
5/7/19 CONFIDENTIAL 4 MODULAR, LAYERED , MICRO-SERVICES AND SCALABLE ARCHITECTURE
Intuitive User interfaces that can be used also for complex analysis
Calculation Engines and specific data structure functional to calculation that has to be done. ▪ Engine programming languages: R, Phyton, C, C++, C, Java ENGINES RELATIONAL DATA DATA GRAPH DB CUBE MART ▪ Database: SQL and/or Oracle DB and/or Mongo DB
Data Lake ▪ Data import of various data sources ▪ Hadoop Ecosystem with a plethora of Tools
SPI (ORS Single Point of Integration): Rest API, Web services, data streams, csv/text files. To orchestrate data flow and to ensure data quality
Data sources: ▪ any existing customer data sources (SAP, Oracle, Excel, etc.) DATA DATA DATA NEW AND EXTERNAL DATA▪ plus external data sources (internet, sensors, weather, etc.) SOURCE 1 SOURCE 2 SOURCE N SOURCES…
5/7/19 CONFIDENTIAL 5 6 Two centuries of clothing innovation
Although today many people consider Brooks Brothers a very traditional clothier, the Company has introduced many clothing novelties to the American market throughout our history as a leader in the American menswear industry:
• Ready-to-wear was introduced in 1849 • In 1896, John E. Brooks, the grandson of Henry Sands Brooks, applied button- down collars to dress shirts after having seen them on English polo players • English foulard ties were introduced by Francis G. Lloyd in the 1890s before he was made president of the corporation • The Ivy League sack suit was introduced in 1895 • The pink dress shirt became a sensation to go with charcoal-gray suits • Harris Tweed was introduced in 1900 • The Shetland sweater was introduced in 1904 • The Polo coat was introduced in about 1910 • Madras was introduced from India via Brooks Brothers to the public in 1902 • In 1957, Brooks Brothers became the first American retailer to manufacture argyle socks for men • The first lightweight summer suits made of cotton corduroy and seersucker were introduced by Brooks during the early 1930s • In 1953, the store pioneered the manufacture of wash-and-wear shirts using a blend of Dacron, polyester, and cotton that was invented by Ruth R. Benerito, which they called "Brooksweave“ • Non-iron 100% cotton dress shirt was introduced in 1998
7 Addresses Current market challenges & opportunities
• Ability to target niche needs (UNTUCKit) Intense competition from • New distribution models (Bonobos) new brands born online • Aggressive pricing models due to lower capital costs
Intense competition from • Amazon invests significantly in decision scientists in order to study competing internet companies like brands, customer behaviors and choices and uses that information to offer its Amazon and Rakuten own private labels.
Great difference of tastes • Need to analyze different behaviors between customers of different generations and propensity to buy • Various ethnicities that have really different preferences, size and propensity to between Gen X and buy Millennials • Stock in store and warehouse inventory must be available for all the channels, Big growth of Online and with distribution models that are continuously changing. Buy online, pick up in mixed sales store, shipping from store, online orders, shipping from store B when a size is not available in store A and shipping to home or pick up in store Need to be able to react • Need for compressed production schedule (from 3-4 months to 3-4 weeks). Dual “in season” to fluctuating supply chain. demand for fashion • Need to manage carry-over products with optimized inventories per channel products
8 How to react to new market needs
Is it possible to cover all these needs with one or more products already existing on the market?
• Demand planning, big data analytics, advance statistics ! Partially covered with existing products, extremely expensive • Production Planning ! Partially covered from generic products of different suppliers (long, medium, short) not optimized and extremely expensive • Logistic Optimization and Requirements/Demand management ! Covered from ORS by way off a product previously implemented at Luxottica (RAISE) • BAGA ! Partially covered from existing products with basic sourcing logic but not science based • Retail planning ! Covered with few solutions but inflexible, expensive and not optimal
If we had chosen this path, we would have had higher one time and ongoing costs with integration challenges, so the decision was to create something innovative, flexible and totally connected.
9 ERP vs. ORS Platform – “Application Partitioning”
ERP and OMS roles AI Decision Sciences Platform (ORS)
• Application of Artificial Intelligence and machine learning techniques to support all company decisions, both strategic and tactical. • Managing all the transactions and all the • Simulation and operations decision support execution activities, back office and in store • Planning of all corporate activities: retail, • Finance, bills, invoices, intercompany operations, logistics, omnichannel sales transactions, warehouses, etc. • Big data retail analytics that analyze all available data (sales, customer data, stores, online navigation, weather, etc.) • Application that was “born global” to be used worldwide
Clear separation of roles between ERP and platform
10 Future Brooks Brothers Platform Vision
Prod. Plan Allocation Mer. Plan Dmnd. Plan ERP OMS In Store E-Com. CRM/Loyalty HRIS Analytics MPS Xstore (Southwick, LIC, Garland) (V16+) Mad Mobile BB NA GMPS (Global View) Vinyl
Xstore (V16+)
BB Japan Vinyl Demand Manhattan Success REAL Cloud Active Planning OMS Xstore Factors & ReMAP Demandware Epsilon / ORS BDRA & & SAP FMS & (V16+) Agility / RAISE (ORS) (SFCC) BIRST GMPS ORS Loyalty BB Korea ADP Not (ORS) Inventory Vinyl Applicable (ORS) Optimization (or local payroll) (N/A) Xstore (V16+) BB FE Vinyl
Xstore (V16+)
BB EMEA Vinyl
11 ORS RETa.i.L Platform
Merchandise Planning (REMAP) Assortment Planning Global Capacity Planning Supply Chain Management (REMAP) (GMPS) (RAISE)
Demand Planning Retail Allocation (REAL)
Assisted Smart Financial Planning
Web Order Management (WOM) Buy Anything, Get it Anywhere (BAGA) Retail Analytics (BDRA) Master Production Distribution Scheduling Planning (MPS) (Local RAISE) 12 ORS – Featured Solutions
Buy Anything, Retail Get it Anywhere Allocation
Retail Analytics
13 Buy Anything, Get it Anywhere
14 BAGA – key features
Omni-channel fulfilment
Configurable, adaptive algorithms
Store & Online Demand
Massive real-time data analytics
15 BAGA – snap shot of results
Since 2017 launch, BAGA has had a significantly strong positive impact on the NA business:
• Significant increase in sales
• Higher sell through – e.g. Women’s sell through increased significantly
• Increased conversion
• Enhanced customer experience due to better product selection and sizing availability
Notes: As of February 2019
16 Retail Analytics
17 Retail Analytics – key features (1 of 2)
Shops true What price? performance? ? ? ? Best value for Which promotion? customers? ? ? ? When sell-off? Best sellers? ?
18 Retail Analytics – key features (2 of 2)
Business Intelligence Store efficiency
Price sensitivity Best\Worst Seller
Promotion Analysis Cluster Analysis
19 Price Elasticity Overview
What is Price Elasticity? A statistical calculation that produces a coefficient that shows the correlation between price and demand volume. The behavior of a highly elastic product shows that as you decrease price, the demand volume increases at an increasing rate*.
LOW ELASTICITY MED ELASTICITY HIGH ELASTICITY 0.0-0.7 0.71-0.99 1+
Features
• Supports decision making for target ticket pricing
• Planning in-season promotional activity
Benefits
• Current process does not leverage statistical analysis (sensitivity algorithm) of historical Brooks Brothers data
• Forecasted view of demand and margin impact for various changes in discounts
• Understanding cross price elasticity impacts
*NOTE: Negative Elasticity shows that demand will increase if the price increases, very common in high end luxury products.
20 Price Elasticity Example: Dashboard
21 Retail Allocation
22 Retail Allocation – key features
AI / Machine Learning Sales Forecasting
Multichannel Store-by-store
Customer centric Highly automated
23 Retail Allocation – key benefits
Improved Reduced Reduced Customer Satisfaction Store Inventories Operational Costs
24 Retail Allocation – Demand planning key benefits: Stock out in stores Lost Sales ($’s)
25 Retail Allocation – Key benefits: Inventory Lost Sales ($’s) Units Units Units
26 Our Vision For The Future - An AI Powered World Our Vision For The Future – A World Powered by AI #RIC19