Sainbury's Viva 

- Embracing voice-UI to improve
the family food shopping experience -
 

Sainbury's Viva 

- Embracing voice-UI to improve the family food shopping experience - 

'Design an omni-channel experience which embraces voice UI for Sainsbury’s customers'

The User - Older UK Families

The average UK family spends approximately £58 a week on food. Having to shop with an entire family in mind can create a variety of difficulties that could be addressed with a voice based solution. This is especially true for older families with kids aged 12-18; with each member having differing food preferences and needs, and the whole family contributing to shopping habits in some way.


The Solution

Sainbury's Viva comes in the form of a voice system that runs on an Alexa microphone, with a supporting mobile app to provide a visual interface for users at key stages. It allows each member of the family to add items to the household shopping list individually through voice recognition.

The system offers safety parental controls, delivery options, and additional features that link members of the same family living in different households; allowing lists to be created, purchased and ordered for those who may not be able to shop themselves. The smart system takes into account dietary requirements, personal food preferences and age to intuitively accommodate each individual user; learning over time to make the ordering process more efficient and re-defining the traditional home grocery shopping experience.

'Design an omni-channel experience which embraces voice UI for Sainsbury’s customers'

The User - Older UK Families

The average UK family spends approximately £58 a week on food. Having to shop with an entire family in mind can create a variety of difficulties that could be addressed with a voice based solution. This is especially true for older families with kids aged 12-18; with each member having differing food preferences and needs, and the whole family contributing to shopping habits in some way.


The Solution

Sainbury's Viva comes in the form of a voice system that runs on an Alexa microphone, with a supporting mobile app to provide a visual interface for users at key stages. It allows each member of the family to add items to the household shopping list individually through voice recognition.

The system offers safety parental controls, delivery options, and additional features that link members of the same family living in different households; allowing lists to be created, purchased and ordered for those who may not be able to shop themselves. The smart system takes into account dietary requirements, personal food preferences and age to intuitively accommodate each individual user; learning over time to make the ordering process more efficient and re-defining the traditional home grocery shopping experience.

OUR PROCESS

THE PROJECT

This was a group project, comprising of a cross-cultural team of industrial and user experience designers. The live brief was provided by Sainsbury's, who we were in contact with through weekly Skype meetings and two visits to their main head office.

The Team

Screen Shot 2018-02-06 at 18.16.20

The Background  

In 1950 Sainsbury’s became one of the UK’s first self-service supermarkets, a transition that signalled the beginning of the move from traditional food counters to a more efficient way of shopping. Fast forward to the noughties and a trend started to emerge; rather than doing one large weekly or fortnightly shop, customers were doing fewer of these big shops and more top-up trips to small convenience stores; reflected by the fact that local stores make up over half of all Sainsbury’s stores across the country.

first sainsburys store

Photo via Flikr

Sainsbury’s, as a brand, continually looks to evolve and generate trust between itself and its customers. While online business is growing at a good rate, it is still based on the big weekly shop method and so Sainsbury’s are starting to think about how to provide ‘on demand’ food to its customers.

They believe the best way to do this is through voice UI, and can see a world where voice will challenge a current shopping experience that is very visually driven; see, click, buy. They view voice as the next big disrupter of the e-commerce industry after mobile and feel that an effective shopping experience, supported by voice, will really set them apart from their competitors.

THE BRIEF

Design an omni-channel experience which embraces voice UI for Sainsbury’s customers; this could include understanding the wider voice trend, the difference between customer segments and online or in-store shopping.

Secondary Researc

Having been given the brief and some background information about Sainsbury's, their customers and the wider shopping picture, we began with some secondary research to give us a clearer picture of the project topic. This research focused on four areas:

voice UI design

Voice UI:

  • History of voice
  • Basic design principles for voice UI
  • Challenges and opportunities
  • Scripting the language to be realistic
  • Human/machine interaction

Voice UI:

  • History of voice
  • Basic design principles for voice UI
  • Challenges and opportunities
  • Scripting the language to be realistic
  • Human/machine interaction
speech recognition

     Speech Recognition:

  • Siri, how it works, and how it learns about the users voice
  • Language and accents, and how they affect speech recognition 
  • Future developments/improvements

     Speech Recognition:

  • Siri, how it works, and how it learns about the users voice
  • Language and accents, and how they affect speech recognition 
  • Future developments/improvements
supermarket shopping experience

Supermarket Shopping Experience:

  • What is the current experience?
  • How does this differ for different people?
  • Comparison between best/worst experiences
  • Early thoughts on how this could be improved

Supermarket Shopping Experience:

  • What is the current experience?
  • How does this differ for different people?
  • Comparison between best/worst experiences
  • Early thoughts on how this could be improved
existing shopping products

Existing Smart Shopping Product:

  • Smart shopping cart
  • Amazon go
  • Ibeacon - customer location
  • Other existing products

We carried out this research to give us a deeper understanding of the current voice UI landscape and how it might change in the future, to allow us to identify and understand the current limitations of voice UI and wider voice technology. 

Defining the Target User

Some early research and brainstorming led us towards International Students as a potential target user group, as we felt they had some specific needs and painpoints that could really be improved with a voice based solution.

Deining the Target User

Some early research and brainstorming led us towards International Students as a potential target user group, as we felt they had some specific needs and painpoints that could really be improved with a voice based solution.

sainsbury’s trip 2 – 21:02:17

However during our first meeting with Sainsburys' they encouraged us to rethink this group, and to continue brainstorming different demographics using a creative exercise to answer three key questions:

  • What products/things do you associate with voice in particular?
  • What might be difficult or easy for different user groups to buy using voice?
  • Positive and negative aspects of shopping for different people? 

Understanding these aspects would then allow us to gauge how much opportunity and need there might be for a voice UI shopping system within each of these groups, and to begin to identify how voice would be able to benefit each of their shopping experiences. 

Meeting 4 – 22:02:17

We repeated this exercise with a variety of different demographics, leading us to a choice of three that we felt had a good range of needs and requirements when going shopping:

We repeated this exercise with a variety of different demographics, leading us to a choice of three that we felt had a good range of needs and requirements when going shopping:

  • International Students - Adjustment to shopping in a different country, different foods etc. 
  • FamiliesShopping for multiple people of different ages, requirements and preferences. 
  • Young Professionals - Need for quick and efficient shopping trips, on the go. 
  • International Students - Adjustment to shopping in a different country, different foods etc. 
  • FamiliesShopping for multiple people of different ages, requirements and preferences. 
  • Young Professionals - Need for quick and efficient shopping trips, on the go. 
choice of three users

Chosen target demographic - UK Families:

  • Lots of scope for an innovative design solution
  • Makes up a large proportion of Sainsbury’s customers
  • Differing users needs within this group - parents buying for themselves and their kids

Chosen target demographic - UK Families:

  • Lots of scope for an innovative design solution
  • Makes up a large proportion of Sainsbury’s customers
  • Differing users needs within this group - parents buying for themselves and their kids
chosen target user

The average UK family spends approximately £58 a week on food and drink; providing Sainsbury’s with a large potential customer base to tap into. Building a stronger customer experience for these average size families would go a long way towards supporting Sainsbury’s desire to be a trusted brand amongst its customers, and generate the brand loyalty it looks for.

Having to shop with an entire family in mind can create a variety of difficulties for this demographic that will be highlighted through user research; allowing a voice UI based design solution to provide these families with a uniquely great customer experience when shopping with Sainsbury’s.

The primary research will help to refine which type of families would benefit most; those with very young children or a slightly older demographic, along with at which point of the shopping experience voice could be most effective.

PRIMARY RESEARCH 

This took place in two different stores, superstore and convenience, to give us the chance to explore two very different shopping experiences; from both the customer and staff point of view - observations and interviews.

Research Questions

Following on from our secondary research we identified a series of research questions that we wanted to answer through our primary research: 

UK families current shopping habits with Sainsbury’s?

  • Their typical shopping experience, both in store and online?
  • The content of their shop and what influences this?
  • Their attitudes towards the shopping experience with Sainsbury’s

Voice UI and technology use among the target user group?

  • What are their experiences with voice?
  • When and where would voice UI be most effective for this demographic?
  • What problems might this user group face when using voice UI?

What are the primary shopping problems faced by UK families that need addressing?

  • When and where do they face these problems?
  • What are the main causes of these?
  • What are these issues, from a staff perspective?

Primary Research Locations

Megastore, Loughborough:
Larger store, likely to be busier with a heavily student customer base considering its location. People will likely be doing larger, more expensive shops rather than short top up trips. Staff likely to have less time available as they have a larger store to maintain.

Convenience Store, Nottingham:
Smaller, convenience store meaning the majority of customers will be getting small top up shops, or just buying items for lunch or dinner. Likely to be far quieter than the larger store with staff having less to maintain in terms of shop space.

convenience

Nottingham Convenience Store

Observations

The first stage of our primary research focused on observing customers critical shopping habits in store while they shop, allowing us to begin to understand the painpoints of the experience. To ensure participants were the correct demographic, they were asked some initial questions:

age?

How old are they?

kids?

Do they have kids?

kids age?

How old are their kids?

This process also enabled us to identify the different factors that influence a customers shopping experience, as well as factors that affect both their attitude and approach to shopping with Sainsbury’s

Interviews 

We then began user interviews, to understand our target users shopping experiences with Sainsbury’s, along with their attitudes and general habits surrounding this area, in greater depth.

This process involved speaking to both customers and employees of both stores, to give insights from two different perspectives and help us explore how the final voice system might impact Sainsbury’s employees as well as customers. The interview process focused on three key areas:

  • Their typical shopping habits - overview of how their family currently shops.
  • Use of voice based technology - current level of technology use and context of use.
  • Views on Sainsbury’s - how they perceive the brand and where it could improve.

Key Insights

We then, through affinity mapping, began to review, organise and analyse our research findings into key insights - to be taken forward to inform a creative voice based solution:

shopping important
  • They tie shopping into their own schedules
  • They aren’t always fully focused on shopping while they're doing it
  • Customers sometimes need assistance (from friends, other shoppers or staff)
  • They tie shopping into their own schedules
  • They aren’t always fully focused on shopping while they're doing it
  • Customers sometimes need assistance (from friends, other shoppers or staff)
whole family impact
  • Kids need distractions in store
  • Young kids will get involved in the shopping process 
  • Family members may all have different food preferences
  • Kids need distractions in store
  • Young kids will get involved in the shopping process 
  • Family members may all have different food preferences
efficient shopping experience
  • Knowing the store layout saves time
  • They prefer shopping in store
  • They know exactly what they want
  • They are constrained by time
  • They stick to a routine
  • Knowing the store layout saves time
  • They prefer shopping in store
  • They know exactly what they want
  • They are constrained by time
  • They stick to a routine
trust sainsburys
  • They are loyal to Sainsbury’s
  • Customers value staff and use them regularly when shopping
need for awareness
  • They’re health conscious
  • They’re price conscious
  • They value quality
  • They can be impulsive
  • They often check use by dates
not fully comfortable
  • They don’t really use voice day to day
  • They’re comfortable with their current resources (ie. Google)

How Might We's

If our insights represented challenges and problem areas for customers shopping experiences, we used how might we questions to reframe these and turn them into opportunities for innovation

The questions suggest that a solution is possible, allowing us to answer them in different ways and create lots of variation in potential ideas. It also ensures the solution is closely integrated with our initial research findings.

NARROWING THE FOCUS

We were then ready to take our research findings and insights forward to further develop the project; starting to identify some potential personas and where a voice system might fit within our users current lives - before selecting a concept. 

Persona's

Within our target demographic we knew there would be different types of family that might use our solution, each of these adopting different lifestyle habits and goals, with a variety of needs and preferences. Through our research we identified that there were four key family types who could be targeted with a voice based solution:

  • Singles and young couples - future users
  • Young families - with newly born children
  • Settled families - with young children
  • Older families - with older, grown up children

We began to build persona’s each of these, to unpack and clarify different aspects of their lives in relation to their background, shopping experiences, food preferences and technology use:

The final persona would be chosen later on in the process when a concept had been chosen, and it was clearer which group it would be most suitable for.

Further Supporting Research

We did some additional research to further establish a clear and defined design direction; to identify the limitations of designing a voice UI based service with respect to shopping:

1 – voice UI role

How voice UI plays a role in ‘conversational UI’

Such as facebook messenger, WeChat, Amazon Echo, Google home etc.

2 – voice interaction efficiency

How efficient is the interaction of voice with its users?

By looking at other major players in the market such as Amazon Alexa and Google home, focusing on the current major strengths and weaknesses

How efficient is the interaction of voice with its users?

By looking at other major players in the market such as Amazon Alexa and Google home, focusing on the current major strengths and weaknesses

3 – alexa

How does Amazon Alexa work?

Key advantages and disadvantages, applications that have been integrated with Alexa - Uber, Dominos, Lyft etc

Lifestyle Maps

We then looked at creating a lifestyle map; gathering images and exploring current relevant trends which might influence a families lifestyle choices, to gain an understanding of where voice might integrate into their lives. We focused on wearable technologies, smart homes, social media, clean and sustainable living and some deeper lifestyle trends. 

Screen Shot 2018-02-07 at 10.21.53

Concept Generation

We generated a variety of initial concepts in response to our how might we questions, and then pitched them to Sainsbury's to get feedback, helping us narrow our focus down to three potential directions:

personal trainer copy
anticipatory desig
family shopping

Of these three directions we felt that there was one that offered a lot of opportunity for innovation, along with being something that was very different to anything currently available to users - Sainsbury's shared our view and we selected it to develop further:

A system that allows family members to individually contribute to a household shopping list through voice recognition.

A system that allows family members to individually contribute to a household shopping list through voice recognition.

  • Voice recognition – User profiles for each family member
  • Smart system – Understands preferences
  • Sainsbury’s – Link to online shopping database

“Great idea - but it will only come to life if you show the depth at which the service would work and how clever it would be, in terms of knowing what the different family members have asked for. How does it map these together? Is it separate orders or one group one? Really fascinating area and nothing that clever on the market at the moment”

- Tom and Richard, Sainsbury's Contact

Having selected a concept to develop, we identified the persona family that would be most suitable to target - Older Families. This persona was selected because of three key factors:

  • Kids will be more aware of food and dietary issues/have more specific food preferences.
  • Kids will be early adopters of technology, with the adults still willing to try new things.
  • Kids have more responsibility and contribute more to family life, rather than relying on parents to buy the food they want.

Having selected a concept to develop, we identified the persona family that would be most suitable to target - Older Families. This persona was selected because of three key factors:

  • Kids will be more aware of food and dietary issues/have more specific food preferences.

  • Kids will be early adopters of technology, with the adults still willing to try new things.

  • Kids have more responsibility and contribute more to family life, rather than relying on parents to buy the food they want.
older family

DESIGN DEVELOPMENT 

Once the design direction had been chosen, we went about developing it further; to define its key features and how it will work, understand the user journeys and develop the conversation that will take place between the system and its users at various stages. 

Mapping the User Journey  

We began mapping out the user journey for the main functionality of the system; adding items to their shopping list and exploring different responses the system would give them at each stage of use. This helped us start thinking about the different conversations that might be had in a variety of contexts and between different users.

This process also allowed us to map how the service might work in conjunction with another family member living in a different out (ie. Grandparents) and where/how these two contexts of use might overlap.

Through creating this map we were also able to identify where the user might need a transition between the voice system and a supporting app based interface, as well as the key features that would be needed to provide an intuitive and efficient voice experience.

user journey map
17554755_10158446151895652_1781465312_n

Delivery of the System

Informed by our secondary research, we decided the system would run through Amazon Alexa, as an Alexa skill.

This would mean that it would need a single Alexa microphone in the house to function, rather than it being a separate product that requires each family member to have a high powered smart phone with voice capability.

Informed by our secondary research, we decided the system would run through Amazon Alexa, as an Alexa skill.

This would mean that it would need a single Alexa microphone in the house to function, rather than it being a separate product that requires each family member to have a high powered smart phone with voice capability.

alexa

Designing the Mobile App  

Having mapped out our user journey defined the systems key features and capabilities, we set about designing the supporting mobile app and its interaction points - finalising the sitemap and starting some basic wireframes. The sitemap helped us to understand the overall structure of the final app in greater detail:

sitemap

We then began developing the key app screens through low fidelity wireframe sketches. This process helped us begin to visualise  the basic layout of each screen, its overall structure, its main components and the navigational content that it would need to effectively facilitate the user journey we had mapped out. 

app wireframes

We reviewed these wireframes as a group and iterated until we were happy with each screen; eventually developing them into some final high-fidelity designs that effectively support the use of the voice system. 

Conversational Testing

As a testing method, and to inform the conversational content of our solution, we decided to record some example conversations that might occur between a user and our system when performing a variety of actions:

  1. The testing began with a ‘user’ selecting items they wanted to add to their list.
  2. They would then speak to the ‘system’ (us), adding these items and getting appropriate responses.

  3. Responses given were based on results from Sainsbury’s website when each item was searched for.
  4. The process was repeated for a variety of items and system functions.

The aim of this process was to identify any potential issues that might arise in the general functionality of the system, and to ensure we give examples of realistic and intelligent conversations that could occur in the home when the product is being used. This process was completed for four example functions: 

Final Group Presentation  2
Final Group Presentation  2
Final Group Presentation  2
Final Group Presentation  2

Multiple scenarios were explored, with the user getting a variety of responses and asking for each item in different ways. These interactions were recorded to highlight any problems or painpoints that might arise when using the system, and at which stage of use these problems might occur. Four examples of these tester conversations are shown below:

User requests Uncle Bens rice, and is then asked for his chosen flavour and quantity to identify the specific item they want.

User requests eggs so the system suggests the same ones as last time; learning from their previous orders.

 Conversation between the system and a younger user who asks for coca-cola, which their parent won't allow them to have

User directed to their app to confirm the order and pay, before choosing delivery and selecting their chosen time slot. 

This also helped to ensure that our content covered a huge range of different conversations that could happen between the system and user, allowing us to present the final concept, its functionality, and the users journey in the most effective and believable way possible. 

THE SOLUTION  

Sainsbury's Viva is a voice user interface system that utilises voice recognition; allowing multiple family members to individually contribute to a household shopping list, and giving families an entirely new home shopping experience.

How It Works

Sainbury's Viva comes in the form of an voice system that runs on an Alexa microphone, with a supporting mobile app to provide a visual interface for users at key stages. It allows each member of the family to add items to the household shopping list individually through voice recognition.

The system also offers safety precautions through parental control, delivery options, and additional features that link members of the same family living in different households; allowing lists to be created, purchased and ordered for those who may not be able to shop themselves. 

How It Works

Sainbury's Viva comes in the form of an voice system that runs on an Alexa microphone, with a supporting mobile app to provide a visual interface for users at key stages. It allows each member of the family to add items to the household shopping list individually through voice recognition.

The system also offers safety precautions through parental control, delivery options, and additional features that link members of the same family living in different households; allowing lists to be created, purchased and ordered for those who may not be able to shop themselves. 

project summary

The smart system takes into account age, dietary requirements and personal food preferences to intuitively accommodate each individual user; learning over time to make the ordering process more efficient and re-defining the traditional home grocery shopping experience.

final screens arty

Scenario of Use

setting up user profiles

1. Setting up User Profiles 

The user sets up their mobile app; stating key information such as their name, and age and any food allergies or requirements they might have.

Adding items

3. Adding Items 

Each family member can then add food items to the household shopping list through voice, which is recognised by the system to identify who it is

confirmation and payment

5. Review and Payment 

In the app the user can check their list, make any last minute changes, and review items added by other family members; before confirming the order and then paying.

activating

2. Voice Activation 

Speaking to the system for the first time, the user introduces themselves, allowing it to create a link between their voice and their profile. 

confirming order

4. Order Confirmation

Once the parents are happy with their list, they can confirm their order and are then directed to their mobile app to review it and pay

delivery options

6. Collection and Delivery 

The system also offers the user a choice between home delivery and collection, with customisable options offered to them over time.

Key Features






User Recognition:

Each user sets up their individual profile on the mobile app, entering their name, age and any allergies they may have. When they first use the system, and introduce themselves to Alexa, the system recognises their name and creates a link between their voice and their app profile. This allows the system to filter items according to their profile, and learn about each users individual preferences over time.

KF user recognition

Voice to App Transition:

We wanted the transition between the voice system and the app to be as seamless and intuitive as possible, therefore the voice system will actively direct the user to the app when they need to use it. When the user wants to confirm the order and pay, the system will say “Ok, just check your app and confirm your order”, creating a really smooth transition between the voice UI and mobile interface.

KF – voice to app transition

Adding Items:

Users can add items in two ways. They can search for items using voice; speaking to the system and being asked questions so the system can find what they want as quickly as possible. In the app, they can see items bought regularly or in recent orders and quick add them directly to their list, so they don’t have to search.

When adding items using voice, the system uses the information in the users profile to filter the options given, taking into account dietary requirements and food preferences; suggesting gluten-free items to those who are gluten intolerant for example. The system will also have detailed information on prices and nutritional information, allowing it to identify the cheapest options when the user requests it.

Parental Control:

Even though all family members are able to add items to the shopping list using voice, the parents will always retain full control over whats paid for. They can review new items, added by their kids, when the system recognises it as something they might need permission for; such as fizzy drinks or alcohol. This will come through as a simple mobile notification, asking them to accept or decline the item in question.

KF parental control

They are also able to see another family members list and merge it with their own. This would enable users to make orders and organise delivery for elderly family members who may not be able to shop independently.

Order Preferences:

After the order is confirmed the family can decide between collection or delivery. For delivery, the system would initially ask the user for their preferred delivery time, and give alternative options if this isn’t available. Over time, it would learn which time slot the family has their food delivered in most often, and will start to suggest their ‘normal’ slot each time they order.

KF – order preferences

Collection would be organised from their nearest store, and they’ll also have the ability to use the app to request another family member, a husband for example, to collect the order for them. This would be done through the mobile app, and would come through as a simple notification they can accept or decline.

Future Developments

We felt that it would be good for us to consider how the system might be developed over time to add different functionalities, and what future applications it might have that could bring additional benefits to Sainsbury’s customers in relation to their shopping habits

FD – weekly meal plans
FD – last minute changes
FD – ingredient suggestions

Weekly Meal Plans:

Could the system allow the family to create weekly meal plans, with the system adding the items for these meals accordingly?

Last Minute Changes:

Could the families order be changed
even after its been confirmed, allowing
them to add items even at the last minute?

Recipe Requests:

Could they request full meals, with the system adding all the ingredients rather
than having to add each one individually?

Project Date: 7th February 2017 - 26th May 2017
Project Team: Group Project - Nikitha Suresh, Yiquan Zhao, Huitan Nui, Xiaoying Wang, Tom Braybrooke
Client: Sainsbury's