Presentation on theme: "Putting Consumers first using Data Science uncover valuable insights into what customers want, where they want it, and how much they’re willing to pay."— Presentation transcript:
Putting Consumers first using Data Science uncover valuable insights into what customers want, where they want it, and how much they’re willing to pay for it
What we do – the ‘elevator pitch’ dunnhumby is the world’s leading customer science company. We analyse data and apply insights from more than 400 million customers across the globe to create better customer experiences and build loyalty. Our insights and strategic process help clients create competitive advantage and enjoy sustained growth.
WHY We’re passionate about loyalty and partnering for future growth HOW We place customers at the center of every decision and personalize their experience WHAT Our capabilities build loyalty and emotional connections with customers that create a lasting competitive advantage Our mission
Macy’s Custom Books leveraged data to deliver each customer with more relevant content A single Custom Book campaign could produce over 500,000 unique versions of the book, reflecting the shopping behavior of each customer Custom Books were personalized by featuring the most relevant combination of product categories for each household Content/category selection driven by customer shopping data Offers, promotions, and creative execution were the same between Custom and Traditional books 15
16 “effective reach” is a critical concept…you don’t need to talk to everyone about everything to drive category sales 36-MENS SPORTSWEAR X COLL 79-TRADITIONAL SPORTSWEAR Category Lift per HH Custom Book vs. Base Book 16-WOMENS SHOES 85-HOUSEWARES 3-PETITE SPORTSWEAR 7-HOSIERY 81-TEXTILES 33-COLOR AND TREATMENT 38-FRAGRANCES 26-KIDS 6-FASHION JEWELRY 5-MENS COLLECTIONS 25-MENS FURNISHINGS 8-HANDBAGS 35-FINE JEWELRY 12-CASUAL SPORTSWEAR 28-WOMEN'S SPORTSWEAR 84-LUGGAGE 13-DRESS ACCESSORIES 31-MENS TAILORED CLOTHING 86-FURNITURE 22-INNERWEAR 87-MATTRESSES 82-TABLETOP 10-JUNIOR SPORTSWEAR 68-DEC HSWRS/TABLE LINENS/DECOR 19-SILVER 27-BOYS 2-20 80-NEO COLLECTIONS 20-COATS Sales for most categories in the Macy’s Custom Book matched or beat the “Traditional Business as Usual” Book in spite of the category only appearing in books for 30% of households (on average)
Our experience shows the longer term benefits of relevant content 5 Relevancy is about continually engaging customers with the right content over time Impact of relevant content on Lift (37 weeks) 17
Kroger’s MyMagazine brings personalized custom publishing to the grocery space Targeted editorial content Increased emotional connection Customer shopping data drives content targeting Balanced coupon offering 8 ranked retention offers 8 ranked acquisition offers
And delivers seamless experience online Open rates significantly higher than industry average Personalized version outperformed previous campaign 2:1 Customers reaction was overwhelmingly positive “Finally! Thank you for making the magazine and the coupons digital.” “Neat. I like this being tailored to my shopping habits.”
? Several examples of why dunnhumby has been successful using Exadata Exadata performed 8x better than the next-closest competitor for our workload! Speed in data manipulation facilitates more data science and therefore more relevancy Concurrency was key SQL – A simple, yet powerful way to manipulate the data for analytical readiness Sharing and packaging SQL is easy 70% of our analysis done right in the database using SQL Oracle R enables deeper modeling without moving the data How Exadata makes this possible
? In addition to ETL staging and processing, we perform all full and incremental backups to the ZFS Appliance Roughly 14Tb per hour Full backups are then offloaded to tape Using the ZFS Appliance saves Exadata resources Backups using the ZFS Appliance
? Where are we headed with the Big Data Appliance Near-line storage for historical data Low cost distributed processing for analytical modeling Platform for machine learning Just the beginning Use Cases for the BDA
? dunnhumby is excited about the future Utilizing the ZFS Appliance for TEMP tablespaces Approximate DISTINCT COUNT functionality – a dunnhumby addition In-Memory Columnar Database R&D on creating RAM disk for TEMP tablespaces What about tomorrow?