User Research

This page synthesises our research: methods, who took part, what we found, and how each finding changed the design. Every session has a de-identified record with a participant code in documents/research/interviews. Original recordings and transcripts stay in the team's private OneDrive.

1. Methods at a glance

Round	Method	Purpose	Run by	Participants	Prototype shown
R1	Semi-structured interviews (5 story-based questions: last trip, someone who changed what you bought, social media, healthy vs cheap, hardest part)	Understand how UQ students actually shop, and where social influence shows up	Aman 2, Anujitha 2, James 1, Kenan 2	7 (R1-P01–P07)	none
Lit	Literature review	Position the gap	All	[confirm: 8 unique papers? the protocol notes say 10]	–
AP	Think-aloud with the v1 app probe, then follow-up questions (in person or by screen share)	React to something concrete: friend recommendations, group goals, recipes, checklist	Aman 5, Kenan 3, Anujitha 1, James 1	10 sessions (AP-P01–P10); AP-P08 may be R1-P06 again	v1 app
HP	Think-aloud with the v2 NutriCart Hub physical Cart Unit (three screens: list, alerts and suggestions, compare)	Test the reframed household concept	Kenan	3 sessions (HP-P01–P03), all with people interviewed before	v2 Hub (physical)
SH	Shadowing in store (accompany a shop, then post-shop questions)	See behaviour rather than recalled behaviour	James (1 completed), others declined	1 (SH-P01, same person as R1-P05) + [n] declines	none
ST	Studio peer interviews (8 Sep) plus proposal Q&A, studio reviews and peer comments	Critique from staff and other teams	Team	Teams 16, 3 and two others (ST-2026-09-08); not target users	slides, v1, v2
R4	Group evaluation of v3 (households at their kitchen bench; pairs shopping)	Does a household actually coordinate through it?	everyone	see Evaluation Plan	v3

Counting rule. Each session has one code. Some people took part more than once (noted under Linked records in each record), so sessions ≠ people. Across R1, AP and HP we ran 20 sessions with UQ students, plus one shadowing session. Counts on the poster and in the Design Process Overview come only from the register in section 6.

Note: an earlier internal analysis said some app-probe sessions used the HTML Hub. The transcripts show all of Aman's five sessions used the v1 app.

2. Who we spoke to

UQ students aged roughly 18–25 who shop in person near St Lucia (IGA Hawken Drive, Chai Supermarket, Woolworths Toowong and Indooroopilly, Coles). They are budget-conscious, juggle study and work, and live in share houses, student accommodation or on their own. Most shop alone, even when they live with others.

3. What we found

Price structures the shop

"I ended up choosing the unhealthier option because of the price. Since I'm an international student, budget is tight." When asked how often: "99 out of 100 times." (R1-P02)

"Nine times out of ten I'm going to go for the product on sale." (R1-P06)

Plans give way to time, mood and routine

"I don't really depend on my list at all. I just depend on my mood." (R1-P01)

People rush between classes and grab what's closest, buy the same few things, and only compare when something is unfamiliar or the price changes.

"Healthy" means different things to different people

Participants judged food by protein, sugar, fat, sodium, freshness, whether it "wasn't deep fried", or simply didn't prioritise health. No single "healthy" score could serve them all.

Social influence is real, but it runs on trust and timing

"My housemates are quite good at finding cheap alternatives and more affordable options. Every time we share the milk, or toilet roll, or whatever, we try to look for the cheapest alternative." (R1-P07)

"Without social media I probably wouldn't know what to cook." (R1-P07)

Recipes from a trusted friend or partner, or a reel seen just before going out, led to purchases. Generic content mostly didn't.

Social comparison divides people

"It's a pretty good thing if two people have the same goal, but sometimes it could be stressful for one person because if he's getting less pay." (AP-P05)

"It's essentially telling me that I'm fat… if anything, it could cause body dysmorphia." (AP-P09), reacting to the group dinner goal

Coordination beats a feed

"I don't have to remember what my roommate texted me a few days ago." (HP-P02)

"I don't really care about Sarah's [a persona in the prototype] chicken burrito bowl, to be honest… I'd rather see like, oh, like a protein powder is on sale." (AP-P09)

Shared lists, housemate requests, knowing what's already at home and who's buying what were consistently useful. Friend feeds were not.

In store, less is more

"If every second aisle my cart is buzzing with a housemate's opinion, that's going to feel like noise, not help, and I'll start ignoring it entirely." (HP-P02)

"Sometimes I forget my budget so seeing what's remaining will really help." (HP-P01)

Effort is the barrier

"Adding the ingredients and suggestions to the home unit will be a pain to do for me and my housemates." (HP-P01)

"If there was a sale on avocados will it let me add them or have to add to home dock to work?" (HP-P02)

Being watched changes shopping

Shadowing had low uptake. People cited privacy, feeling self-conscious and time. The method itself was evidence: in-store support has to work without an observer, and our evaluation had to lower the burden on participants.

4. From findings to design

Finding	Design response	Where in v3
Price structures the shop	Budget is the most prominent element in store	Cart Unit header: money left, progress bar
"Healthy" differs by person	Compare shows facts in the order the shopper picks; no verdict, no score	Compare: Price / Protein / Less sugar
Comparison can hurt (pay, body image)	Group health goals left out of v3; leaderboard-style comparison never built	– (left out)
Trust and coordination beat feeds	Household requests with notes instead of a friend feed	Dock list: "for Jay · the cheap one please"
Most people shop alone, but others shape the shop	Solo shopping works fully; others take part through the Dock	G2 requests, G3 "Jay is shopping now", G4 Ask home
Alerts become noise	No push alerts from housemates; one deliberate question at a time, answered with a tap	G4 Ask home
Off-list buys happen	Add items in store	Cart Unit: "Not on the list?"
People forget what's at home	"Already at home" list fed automatically by each shop	Dock pantry, badges on the list
Data entry is "a pain"	Quick-add chips, a meal → missing-ingredients button, prices optional	Dock quick-add, meal board
Another app is unwanted	A dedicated household object: Dock in the kitchen, Cart Unit on the basket	Physical housings [in progress]
Shopping together means splitting up in store	Split the list across two Cart Units	G5 Split Shop
Shared costs cause friction	A neutral cost split after the shop	G6 return summary

5. Limitations

• Samples are small and qualitative; we report patterns, not percentages.
• Some early prototype questions were leading, or explained a feature's benefit before the participant had reacted.
• Not everyone saw the same prototype version, so we don't compare versions as if they were conditions.
• AP and HP sessions were one participant at a time, which can't show household dynamics that unfold over days (Grudin). R4 is designed to address this.
• People who agreed to take part may be more interested in food and shopping than those who declined.
• Studio peer feedback is critique, not user evidence, and is labelled as such.

6. Evidence register

Every finding on the poster or in the overview traces back to a row here. Codes only, never names. Fill the last column as findings are used.

Code	Study	Run by	Prototype	Record	Used for
R1-P01	R1	Aman	–	record
R1-P02	R1	Aman	–	record
R1-P03	R1	Anujitha	–	record
R1-P04	R1	Anujitha	–	record
R1-P05	R1	James	– (also SH-P01)	record
R1-P06	R1	Kenan	–	record
R1-P07	R1	Kenan	–	record
AP-P01	AP	Aman	v1 app	record
AP-P02	AP	Aman	v1 app	record
AP-P03	AP	Aman	v1 app	record
AP-P04	AP	Aman	v1 app	record
AP-P05	AP	Aman	v1 app	record
AP-P06	AP	Anujitha	v1 app	record
AP-P07	AP	James	v1 app	record
AP-P08	AP	Kenan	v1 app	record
AP-P09	AP	Kenan	v1 app	record
AP-P10	AP	Kenan	v1 app	record
HP-P01	HP	Kenan	v2 Hub (physical)	record
HP-P02	HP	Kenan	v2 Hub (physical)	record
HP-P03	HP	Kenan	v2 Hub (physical)	record
SH-P01	SH	James	– (in store)	record
ST-2026-09-08	ST	Team	v1 app (peers)	record
H1-P1 …	R4 household		v3
S1-P1 …	R4 shop pair		v3

References

• Di Cosola & Neff (2022). [full citation]
• Grudin, J. (1994). Groupware and social dynamics: Eight challenges for developers. Communications of the ACM, 37(1), 92–105.
• Houde, S., & Hill, C. (1997). What do prototypes prototype? In M. Helander, T. Landauer & P. Prabhu (Eds.), Handbook of Human-Computer Interaction (2nd ed., pp. 367–381). Elsevier.
• Marshall et al. (2022). [full citation]
• Strecker et al. (2024). ShoppingCoach. [full citation]
• Wang et al. (2024). CSCW. [full citation]
• [remaining papers from our review, full citations]