This document presents research and design work for a system to help people with low socioeconomic status make healthier food choices at grocery stores. The system involves a device that attaches to shopping carts and recommends healthier alternatives when items are scanned. Research included interviews and observations to understand how people select foods. Concepts and prototypes were developed and tested with users. The goal is to facilitate healthier choices within budgets through in-store recommendations and self-tracking of shopping habits over time.

1. One of a CHInd Concept
Healthy Food I Can Afford
IDP Project 5
Team I
Scott Trepper
Haley MacLeod
Steve Layton
Denique Ferguson

2. Let’s go shopping
with Anna!

3. We want to help Anna find
healthier food within her budget.

21. Anna is becoming
empowered to make
healthier choices
for her family.

22. Future Strategies:
• Facilitate
additional selftracking of
shopping history
• Independent of
loyalty systems
• Coupon
Promotions

23. Questions?

24. Appendices

25. Predispositions
People struggle to evaluate how healthy different foods are
People with less disposable income buy less healthy foods.
Health recommendations alone will likely not motivate shoppers to alter their
buying patterns
Pragmatic factors such as taste, cost, and convenience primarily influence
shoppers.
We eat food, not nutrients.

26. Our Design Challenge
Help people of low socio-economic status make
healthier food decisions at grocery stores.

27. Research - Methods
Methods Used
- Interviews with nutritionists
- Interview with psychologist
- Interview with grocery store manager
- Observation of grocery shopping behaviour
- Observe target group members interacting
with concepts

28. Research - Results
Understandability of Nutritional Information
It is easier to be guided by dietary guidelines; people don’t really understand
nutrition labels - unless they’ve been taught how to read them.
-MyPlate is easier to understand than the food pyramid1
-The meaning of points can be difficult to communicate in a shopping
environment2
-“A flood of information is becoming available from a variety of sources
aimed at informing the consumer so that better decisions can be made
while shopping.”3 But much of it is confusing unless properly curated.

29. Research - Results
Selecting what Nutritional Factors to Display
The population is deficient in fiber, potassium, iron, and whole grains, but has
too much sodium, saturated and trans fats and cholesterol (which happens
easily because the body produces its own cholesterol).1
-Alternatives do not need to be given when healthy foods are selected2
-Nutritional hierarchies can be facilitated by the Delphi method (give
nutritionist 50 foods and ask to rate top 10 healthiest multiple times)3

30. Research - Results
Recommended Alternatives
Alternatives should be based on practical categories (frozen meal, snack food,
dessert, side dish, main course, etc) to facilitate buy-in
-Skewing recommendations based on food group deficiencies in
purchase history could aid overall diet1

31. Insights
Grades as a dietary guideline are useful because they give a ranged
“snapshot” of the nutritional value.
-A food’s specific health grade is better to show because it provides itemspecific information
A comprehensive comparison of sodium, fat, and fiber will aid users in learning
about appropriate nutritional ranges.
Randomized alternatives can minimize food company backlash - “Food =
Politics”1

32. Concepts & Concept Systems
• Educate people on healthier food choices in situ
• Help people find alternatives within the constraints of
budget, dietary needs/wants and availability
• Provide in situ information without relying on the user to
have expensive equipment such as smart phones
• Take advantage of reinforcement from an existing, familiar
healthy eating campaign: myPlate.

33. Concepts
Additional sketches following the Project 4
design leading up to the final prototype.

34. Concepts — Sketches

35. Concepts — Sketches

36. Concepts — Sketches

37. Concepts — Sketches

38. Concepts — Sketches

39. Concepts — Sketches

40. Concepts — Sketches

41. Concepts — Sketches

43. Usability Testing
Step 2: Ask the tester some basic demographic and background questions (age,
income, education level, ownership of technology) to establish that they are similar to
our personas, as well as to ensure they are comfortable talking, so it will be easier for
them to think out loud during the task walkthroughs.
Step 3: Show the main screen to the tester. Explain that this is a device that would sit
on your shopping cart handle, and that items can be scanned using the barcode
scanner. Ask for initial impressions, without interacting with the device.

44. Usability Testing
Step 4: Have the tester complete 4 tasks, while describing their thoughts out loud.
Task 1: You pick up a bag of Cheetos off the shelf and scan it on your cart. What would you
expect to see on the screen?
Task 2: After you scanned the Cheetos you will see this screen. What information does this
tell you? (Follow up with questions about their understanding of specific components if
necessary). What do you think this screen allows you to do?
Task 3: If you were to click on the edamame beans on the right hand side, what would you
expect to happen? Go ahead and click on it now. What does this information tell you? What
do you think this screen allows you to do?
Task 4: If you were to click on the arrows at the bottom of the screen, what would you expect
to happen? Go ahead and click on one of the arrows now. What does this information tell
you?
Task 5: Now that you’ve seen the information for Cheetos, Edamame beans, and Pirate’s
Booty, which one (if any) would you choose to buy? Where could you find it in the store?

45. Usability Testing
Step 5: Ask questions of the tester to evaluate their understanding of how the system
was working, and get their opinion after having interacted with it.
·
Which of the three foods was the healthiest? How do you know?
·
Which of the three foods was the least expensive? How do you know?
·
What was your favourite thing about the system? What was your least favourite
thing?
·
Was there anything that was confusing to you about the system?
·
If you had a magic wand, what would this system look like?
Step 6: Answer any of the tester’s questions and thank them for their time.

46. Usability Testing - Conclusion
●
Purpose
We need to make clear why someone would want to “hold item here to scan” in the first
place.
●
Continuity
We need to include instructions for someone to scan their loyalty card in before they are
invited to start scanning in items.
●
Consistency
If green is used to indicate a positive characteristic, it shouldn’t be used decoratively e.g.
the Cheetos screen has the price and the “health points” of the item encircled in green,
when it is not a good thing compared to the alternatives.

47. Usability Testing- Conclusion
●
Clarity
○ We need to declare what the “health points” mean up front. We might want to have
an introductory screen that explains these things - an introduction that can either be
skipped, or maybe if the system is smart enough, plays the first time you scan your
card in, and can be played again if the user summons it.
○ We also need to indicate that the numbers by the alternatives are also “health points”.
We don’t necessarily have to label them, but if they have the same color and border
as the one for the scanned item, and that one is labelled, they may intuit the
relationship based on the gestalt principle of similarity.
○ The icons for the alternatives should be labeled.
○ We might need to think about using bigger text and a sharper contrast color scheme
to make the screens easier to read.
○ We might also want to consider not using dots to indicate which number is better in
the nutritional information, but a checkmark and possibly putting the number in green
as well (we’d have to choose a different color to decorate the alternatives.
○ We need to make it clear which items are clickable, and which are just information
only.

48. Strategies
-Facilitate self-tracking of shopping history
Allows users to track shopping habits over time, across multiple stores, so they can see their progress and take
a more proactive role in managing their diet.
-Become loyalty-card independent While the system can use data from loyalty systems to help
personalize recommendations, we also see it being adapted for use in stores that don’t have loyalty programs.
This would allow more people to have access to the system.
-Coupon promotion
Stores may also use the data collected to further personalize coupon promotions for their customers.
-Reduce health care costs
Helping people to make healthier food choices can contribute the reduction of diet-related health issues and
their costs on the health care system.

49. Golden Questions
1.
Why do we eat junk food when we know we feel better if we eat healthy? What factors contribute to someone making
an unhealthy food decision? Does our intervention need to recognize the user’s primary motivating factor? How would
an intervention that addressed someone’s stress eating differ from an intervention targeted at binge eaters.
2.
What current products exist that recommend substitutes based on taste? What strengths and weaknesses can we
identify in those systems? How have readers of the “Eat This, Not That” book series benefited? Does the series provide
any introductory warnings regarding how the information should be used?
3.
Most food applications are geared towards weight loss. How does this impact someone who wants to gain weight?
4.
Do people experience taste in the same way? Is my sweet the same as your sweet? How do we account for people
craving certain flavours or textures?
5.
How do we choose what food we eat? What is the difference between how we say we make food choices, and how we
actually make food choices?
6.
How do we think about food? What are different ways we can classify food, or evaluate how healthy it is?
7.
What is the relationship between food and social wellness? What influence do other people have over what hey eat?
Do people eat differently when they are alone than out with friends?
8.
What body data can we collect? Are there any introductory questions that our users would not feel comfortable
answering? Would our users feel uncomfortable if we asked them their weight?
9.
What are political issues surrounding food and class/community? Are there commonly accepted standards for what is
healthy and what is not? How does that vary from person to person?
10. What diet changes are cornerstone changes? For example, would people start eating smaller portions without explicitly
being recommended to do so if they are directed to more filling foods?

50. Golden Questions
11. How long do people typically spend grocery shopping? What patterns do they move in throughout the store? Do they
move systematically from aisle to aisle? Or do they jump around the store? Do they stop at every aisle? Do people feel
rushed when shopping, or do they take their time? Is grocery shopping an enjoyable experience, or a necessary evil?
12. Can tracking the body data of a shopper allow us to generalize about the progress of the rest of the household? What if a
friend accompanies me to shop, and she holds the cart handle?
13. What if friends shopping for two different households share a cart? How can we make recommendations for alternatives
specific to each? What if one household has a nut allergy and the other one doesn’t? Should the displays be built into
the cart, or available for people to clip on to the cart as needed? What happens if the toddler throws a tantrum? Can he
remove the display and toss it across the aisle? Hit himself in the head with it? What happens if someone spills a Slurpee
on it?
14.
If we don’t ask set-up questions, how will we avoid recommending chicken to a vegetarian who’s lacking protein? Can the
system pick up that a household is vegetarian by the shopping habits? If a significant portion of the low-SES population
are immigrants, will the system be sensitive to cultural food preferences? Religious ones? How will it know the difference
between avoiding allergens and avoiding religiously proscribed food?
15. How can we keep track of healthy foods that don’t have barcodes?
16. How many shopping trips does it take to learn someone’s grocery habits? Can the system tell when there’s a temporary
blip? When should a blip be ignored? What recommendations can we provide before the system has had time to learn
people’s habits?
17.
What happens if the system recommends a product that is recalled from public consumption? Could we be held liable for
any damages?

51. Golden Questions
18. Will people get confused if they were eating Cheetos and we’ve been recommending Pirate Booty, and then one day we
recommend something else instead of Pirate Booty? Will people think that Pirate Booty is always generally good, as
opposed to this item fits your needs right now?
19. What happens if someone scans an item that is extremely healthy? Will they still be recommended alternatives? Will
they then want to distinguish between healthier alternatives and “you might like this for a change” alternatives?
20.
If the system pre-arranges a list of alternatives to what you bought at your last shopping trip to help you choose
differently on your next trip, what happens if you pick up an item you never picked up before?
21.
Some stores offer discounts for group purchases (“Buy 10 participating items for $5 off). How will the system react if an
item fall within the correct price range only if the discount is applied? What happens if 10 participating items are added
but then one is removed?
22.
Grocery carts can collect germs and spread illness. Are our users in danger of coming into contact with more pathogens
by using our system? Will an employee need to clean the devices?
23.
How will the system handle unusual shopping behavior? For example, if someone is buying food for a party they are
hosting or to donate to a food bank, how does the recommendations they receive?
24.
How can we circumvent the endowment effect? If people pick up a product, have they already decided to buy it? How
can we engage their curiosity and encourage them to compare alternatives?
25.
What material considerations need to be made for a device that is intended to frequently go outside? What about bad
weather? If the risks of weather are too great, should the device be a “clip-on”? Would employees need to detach and
reattach the device at the start and end of each shopping? How is the device powered? How might grocery companies
react to having their employees take on this new responsibility.

52. Design Process - Determining Device Alterations

53. Design Process - Determining Device Alterations

54. Design Process - Determining Device Alterations

55. Design Process- Determining Device Alterations

56. Design Process- User Testing

57. Design Process- Documentation

58. Citations
• Krug, S. Rocket surgery made easy. The do-it yourself guide to
finding and fixing usability problems. New Riders, Canada,
2010.
• Lindeman, Alice. Personal interview. 26 Nov. 2013.
• Thiagarajah, Krisha. Personal interview. 26 Nov. 2013.
• Todd, Peter M., Yvonne Rogers, and Stephen J. Payne.
"Nudging The Trolley In The Supermarket." International
Journal of Mobile Human Computer Interaction 3.2 (2011):
20-34. Print.