New retail experiences

How machine learning can be used to personalise the discovery journey and create new retail experiences.

Transcript

1. AI and the future of retail Peter Petrovski Product Manager at THE ICONIC
2. Hi, I’m Peter Product Manager at THE ICONIC 10 years in tech industry ex Yahoo, Domain.com.au, Telstra Passionate about creating the future
4. Australia and NZ’s #1 14M+ visits (May 2018) Fashion and Sportswear Online Retailer + 88% NPS leading user experience 50% YoY Growth 70,000 Products 2,000,000+ App download 16M Unique people reached yearly across all channels
5. We turn shopping into a seamless and inspiring experience
6. DEMOGRAPHICS Limited product selection + shopping moments Catalog Evolution of commerce drivers 1890s - 1940s BRAND Rising product selection and shopping moments Departments stores 1940s - 1990s UTILITY Massive product selection and 24/7 shopping moments Ecommerce 1990s - 2015 DATA Artificial Intelligence Visual commerce Personalization Emerging tech & personalisation 2015 ... Amazon, eBay Macy’s, GAP, NIke Sears Roebuck, Montgomery Ward THE ICONIC, Amazon, Alibaba, Facebook...
7. Personalization Visual commerce Operations Placing customers at the center of the shopping experience, by understanding their tastes and preferences and presenting the right content at the right time. Empowering customers to shop by what they see. Making online shopping more convenient and engaging, breaking down sizing and look and feel barriers. Identifying order patterns to improve picking and packing accuracy, building automated processes to deliver to customers faster with more efficiency.
8. Personalization Visual commerce Operations Placing customers at the center of the shopping experience, by understanding their tastes and preferences and presenting the right content at the right time. Empowering customers to shop by what they see. Making online shopping more convenient and engaging, breaking down sizing and look and feel barriers. Identifying order patterns to improve picking and packing accuracy, building automated processes to deliver to customers faster with more efficiency.
9. Technology Market People Innovation drivers Innovation is the act of delivering breakthrough products and services that deliver value.
10. It all started with a machine that could recognize a pair of sneakers During a hackathon, some developers combined THE ICONIC product data with Alibaba Cloud image recognition technology….
11. But first, some context...
12. of the purchases are still happening offline ONLINE OFFLINE 90%
13. Size and Fit Discoverability Customers drop off due not finding what they need 40% customers feel uncertain when they’re choosing the right size 45% Look and feel don’t buy because they don’t have a sense of the garment’s tactility 31% Online shopping barriers
14. Fixing the Discoverability barrier
15. Snap to shop See a product you like in a physical store, take a photo and get it for a better price on THE ICONIC.
16. $$$ $$ Comparing prices Take a photo of a product or outfit for inspiration, find and buy them on THE ICONIC.
17. Social content Take any image from social media and upload into Snap to Shop, buy the products you love from Instagram.
18. demo
19. When it worked well, it was magic. But when it didn’t... During the user testing, users got frustrated when the results were not what they were expecting.
20. Machines need to be fed We made a decision to not store user photos. So we needed another way to feed the machine the data it needed to make better predictions, and produce better results.
21. Internal testing and refinement We made our employees the guinea pigs, creating an internal testing and reporting process. We used this local data to improve the learning models.
22. Very low discoverability Very high use rate
23. Technology moves fast, human needs change slowly.
24. Reinvent filtering
25. 70, 000 products from 1,000 brands
27. Category, subcategory Shape Material Occasion Gender Colour
28. Visual search A quick way to refine the catalog by finding visually similar products.
29. DEMO
30. Machine side Human side Fashion side What is obvious for the us is not obvious for the machine It’s hard to learn a new way to do things, so they need to be easier Offering similar products to the user is not good enough
31. The next step
32. Bringing innovation to the core of the business At THE ICONIC product design, business and technology work as a team to innovate at the intersection of desirability, feasibility and viability. DESIRABILITY FEASIBILITY VIABILITY
33. Working for innovation in multi disciplinary teams We work in cross-functional squads encouraging a collaborative culture and experimentation. DESIGN & USER EXPERIENCE DATA SCIENTIST PRODUCT SOFTWARE ENGINEERS BUSINESS STAKEHOLDERS
34. The design sprint methodology The sprint is a five-day process for answering critical business questions through design, prototyping, and testing ideas with customers.
35. Insights Expert interviews Define hypotheses Inspiration Sketches Storyboarding D1 D2 D3 D4 Prototyping User testing Validate questions
36. Customer Research Design iterations Development Development Design Sprint Testing Launch Design adjustments Human-centred process
37. Snap to shop Visual commerce as a alternative to traditional search We don’t know anything about the user and help the user to trigger the search Tool for searching Visual search Visual commerce integrated in the catalog search We know what the users is looking for and help them to refine the search. Next - More personalization - Closer to the user - Understanding preferences Personalisation Evolution of Visual Commerce at THE ICONIC
38. The future of visual commerce
39. AI COMMERCE VISUAL Shop using visual content by default Augmented reality shopping experiences See how it looks on you Personalised outfits based on your style Real time video and shopping Predict fashion trends Bridge offline and online The future of visual commerce
40. People first
41. Thank you