Jamie

Your smart shopping assistant and long lost friend

 

Thesis summary

Jamie is a predictive commerce-based system that makes you a smarter shopper by connecting your daily activities and digital life. Using rich algorithm and data pulled from various platforms and networks, Jamie aggregates information about almost any product in seconds, allowing you to make better purchasing decisions. This thesis is scoped to a MVP project and is intended as a product proposal.

Problem

In recent years, there has been an ongoing friction between physical and digital retail. Jamie aims to bridge the gap between both spaces and improve the user experience as a result. Imagine a purchasing experience that fits within the context of your location, the events that you attend, and your current shopping desires. Jamie helps makes shopping an immersive activity that’s more organic and social, instead of feeling like calculated chores. In a seamless fashion, Jamie connects to your existing digital life to ensure that you make the best purchasing decisions possible through machine learning algorithms focused on discovering the most relevant, affordable, and accessible items.

Process

When I was working at Garage and Dynamite, I re-organized clothing product categories because there was a lot of repetitive categories and subtle differences that created too many categories and choices. Although re-organizing by physical characteristics versus styles helped reduce categories, there was still a paradox of choice and interacting with menus and categories is an impersonal experience.

Paradox of choice + personalization/customization
 — Connect with customer taste/aesthetic
 — Sometimes products can’t be described in words/search field (impersonal)
 — AI can help describe subtle differences using a visual conversation (personal)
 — Interact with the product to make faster purchase decisions
 — Frictionless browsing experience using intuition (less typing, searching)

 

Audit: Omni-channel in e-commerce

The audit and experience led me to thinking there must be a better way of finding relevant products. I started an audit about the rise of omni-channel shopping in the marketplace and marketspace, in order to explore and identify problems between the spaces. The primary driver of omni-channel is the complete integration of the shopping and brand experience, due to the rise of social networks and personalized retail. I was researching about omni-channel integrations because it involves environmental patterns and behaviours associated with different touchpoints, in order to have an immersive, human-centric mindset for identifying problems.

 

November 2015  —  I came across the Visual Filter AI by Sentient for Shoeme.ca and it appears to be their initial deploy.

November 2015  —  I came across the Visual Filter AI by Sentient for Shoeme.ca and it appears to be their initial deploy.

 
I was intrigued by the Visual Filter AI and noted down a sky blue possibility for a next project (this project)

I was intrigued by the Visual Filter AI and noted down a sky blue possibility for a next project (this project)

 
October 2016 (A year later): Shoeme.ca has 3 additional steps for size selection onboarding

October 2016 (A year later): Shoeme.ca has 3 additional steps for size selection onboarding

Visual essay and research report

I created a basis for the app as mobile-first: Products and camera act as the main visual conversation without menus and categories

I created a basis for the app as mobile-first: Products and camera act as the main visual conversation without menus and categories

Competitive analysis

 
Screenshots via thread.com . The homepage doesn’t include at least a sneak peek of what the product offers to establish trust.

Screenshots via thread.com. The homepage doesn’t include at least a sneak peek of what the product offers to establish trust.

 
 
Screenshots via thread.com . The onboarding process involves too many steps and requirements without actually seeing what the product can offer first.

Screenshots via thread.com. The onboarding process involves too many steps and requirements without actually seeing what the product can offer first.

 
 
Screenshots via snapfashion.co.uk.  The homepage consists of product offerings, while allowing shoppers to interact with the product and check out its features, such as “Snap Similar” using AI + photo segmentation, price drop alerts, and saving products — without forcing initial account creation.

Screenshots via snapfashion.co.uk. The homepage consists of product offerings, while allowing shoppers to interact with the product and check out its features, such as “Snap Similar” using AI + photo segmentation, price drop alerts, and saving products — without forcing initial account creation.

Photo recognition technology is being used by Amazon to scan and identify products to purchase. However it tends to lead to pricier merchants on the Canadian side.

Photo recognition technology is being used by Amazon to scan and identify products to purchase. However it tends to lead to pricier merchants on the Canadian side.

 

 

 

Mapping the customer journey using mobile micro-moments

I used contextual cues to map the customer journey and mobile micromoments: I-want-to-go, I-want-to-know, I-want-to-buy, and I-want-to-do.

Mapping the customer journey, touchpoints, and mobile micromoments

Mapping the customer journey, touchpoints, and mobile micromoments

 
Initial sketches of micromoments for user scenarios

Initial sketches of micromoments for user scenarios

 
 

Primary research and rough notes

 

Addressing the challenges

Working between the retail and e-commerce spaces, I noticed a lack of connection with people’s daily activities and digital life. I used correlational research methodologies by gathering and analyzing scholarly books, user research, and statistics to help confirm the demand for engagement, personalization, and self-service in e-commerce. Using correlational research is challenging because I have to determine the relationship between 2 variables from the same group of subjects — which helps determine a similarity but not a difference — so I cannot establish a direct cause and relation. I am also addressing the problem based on my personal experiences in working between the retail and e-commerce spaces in customer service which translated to this product — I was frustrated by the lack of personalized, social, and contextual selling to create a magical shopping experience, which resulted to this project.

Pact Analysis

La_Cindy_File3.jpg

People
∙ Busy, on-the-go, and stressed students
∙ Price conscious
∙ Spends about $2000 a year online
∙ Gen Z to Millennials (teens to age 35)
∙ Live for today
∙ Will select and continue to buy favourite brand at pre-recession prices

Activities
∙ Heavy social media users
∙ Actively online looking for inspiration
∙ Researches the best deals online and in-store
∙ Heavy online shoppers with data to pull from
∙ Early adopters
∙ Brand loyalty: Uses social media to tell friends about products and influence upcoming trends

Context
∙ Socially connected purchases
∙ Asks friends for feedback in purchasing decisions
∙ Browses individually for inspiration
∙ Shares inspiration using social media

Technology
∙ Current and proposed: mobile

Amazon’s Alexa and Google Home are starting the evolution of immersive shopping and smart assistants using connected devices and voices. Constraints include income and adoption for devices such as voice-activated speakers and virtual reality kits. What we need is a transitional app on smartphone devices with global adoption and inclusive reach, while reducing income barriers.

Scalable machine learning connects behaviour with environmental patterns and data from various social network profiles. By connecting to various data sources, we can recommend a personalized product catalogue which creates a sense of empathy and understanding — what humans can’t do efficiently on a global scale.

 

Solution

Scope

Jamie — your smart shopping assistant and long lost friend on iPhone, is scoped to MVP features including Visual Search, For You, Personalized Filtering, and Your Taste.

  • On-boarding of primary MVP feature: Visual Search
    • The on-boarding is low friction to encourage people to try out the product and then have opportunities to incrementally sync sign-ins from other platforms using in-app contextual cues to sync.
  • Dismissive "skip" sign-up flow in which user doesn't need to commit to signing up and have the option to explore the app
    • People still get the product results to try out the product

    • “Jamie” Would Like to Access the Camera — gives access to location

    • Once they’re done with the results, they get directed to the For You home feed.

  • For You is connected to your location via allowing camera access which is an iOS default. On the feed, the results are connected to your recent searches, location, and cookies. The results shown here are Trending in Toronto and Drake: Summer Sixteen Tour in 2 weeks.

  • Personalized Filtering and Your Taste contains documentation of your preferred products and past orders, as Jamie adjusts to your taste through adaptive filtering to create a personalized catalogue.

 
 

Jamie is my thesis project created for the YSDN Travelling Graduation Showcase in May 2017. The app is also named after my professor, Jamie!