Where does Loblaws lose its customers?
Loblaws is one of Canada's largest grocery retailers — yet its desktop web experience was creating invisible barriers between shoppers and their carts. This study set out to find where, and why.
Industry
Grocery Retail
Platforms
Desktop Web
Timeline
April 2026
Meathods
Moderated Usability Testing, Affinity Mapping, Heuristic Evaluation
Team
Neha Thakur, Ariana Yang, Saadaf Mohsin, Salman Zafar
8
Participants across 8
remote sessions
4
Tasks spanning the full
shopping journey
75%
Completion rate
on checkout
12+
Distinct usability issues identified and prioritised
01 — THE PROBLEM
People could shop. They just couldn't check out.
Loblaws' desktop site looked functional on the surface — all 8 participants could browse and add items. But the further they moved toward checkout, the more the experience fell apart.
Carts reset after logging in. The $30 minimum order appeared only at the final step. Location prompts interrupted the flow repeatedly. Two participants couldn't complete checkout at all, not because they gave up, but because the system blocked them.
02 — WHAT WE TESTED
The closer to payment, the worse it got.
We focused on four tasks that covered the full journey: browsing and adding items, finding deals, reviewing the cart, and continuing to checkout. Sessions ran ~30 minutes each, moderated remotely via Microsoft Teams with a dedicated note-taker in each session.
Findings were synthesised through affinity mapping in FigJam, then rated by severity based on frequency and impact.
01
Browse & Add Items
Slow load times, cluttered cards, filter gaps
6.25/7
~5.2 min · 8/8 completed
02
Find a Deal
Inconsistent labels, "Weekly Features" hard to find
4.88/7
~3.2 min · 8/8 completed
03
Review Your Cart
$30 minimum hidden, recommendations obscured cart
5.94/7
~4.1 min · 8/8 completed
04
Continue to Checkout
Cart resets at login, unclear steps, location confusion
4.44/7
~7.25 min · 6/8 completed
03 — WHAT WE FOUND
Six friction points that broke the experience — and trust.
HIGH SEVERITY
The $30 minimum was a surprise at the worst moment
Disclosed only at cart review, after users had spent significant time browsing. Participants were forced to backtrack and add unwanted items just to proceed.
Multiple participants blocked from proceeding
HIGH SEVERITY
Logging in wiped the cart
Cart state was inconsistent across sessions. Items disappeared or merged with previous orders. The cart icon vanished in certain states, leaving users hunting for their order.
5 of 8 participants affected
HIGH SEVERITY
Nobody knew which step came next
Checkout had no visible progress indicator. "Save & Continue" had to be clicked twice with no explanation. Users were told to "complete steps to proceed" without knowing what those steps were.
6 of 8 participants affected
HIGH SEVERITY
Store selection confused everyone
The default location didn't match users' actual location. Pickup and delivery modes were visually indistinguishable. Repeated location prompts kept interrupting the flow mid-checkout.
7 of 8 participants affected
HIGH SEVERITY
The site felt like it was fighting the user
Pages took 5–7 seconds to reload during cart updates. Unexpected refreshes reset scroll position. One participant: "It's like doing the groceries twice."
Affected every session
MEDIUM SEVERITY
Recommendations hid the actual cart
Promotional items appeared at the top of the cart page — above actual cart contents. The "Items you might like" label masked what was essentially an ad carousel.
6 of 8 participants affected
"What am I missing? It says complete steps to proceed. But what steps?"
— Participant 8, during the checkout task
04 — WHAT WE RECOMMENDED
Fix what's blocking checkout first — then address the foundation.
01
Show the $30 minimum from the start
A persistent "spend $X more" indicator during browsing and in the cart. Disable checkout only when the threshold isn't met — with a clear, visible reason why.
02
Preserve the cart across login
Merge anonymous and authenticated carts. Keep the cart icon persistent at all times. Offer guest checkout to remove account creation as a blocker.
03
Make checkout steps visible
Add a numbered progress indicator. Show what's complete and what remains. Provide inline feedback when a step must be saved before the next one unlocks.
04
Auto-detect location, simplify pickup
Default to the nearest store on first visit. Give pickup and delivery visually distinct active states. Cut repeated location prompts to a single moment at the start of checkout.
05
Let the cart be a cart
Move promotional recommendations below the subtotal. Show actual cart items on first scroll. Fix the broken substitution selector and surface error messages when interactions fail.
06
The real fix is the technology stack
Slow loads, unexpected refreshes, unresponsive clicks — these are infrastructure problems. Design improvements will help at the margins. A platform rebuild is what resolves the root cause.
05 — WHAT I LEARNED
Running this study changed how I think about research.
Completion rate is the least useful metric on its own
Every participant completed tasks 1–3. But the time, hesitation, and frustration we observed told a completely different story. Completion is the floor, not the ceiling.
UX problems aren't always UX problems
The biggest pain points in this study weren't caused by bad design decisions, they were caused by a slow, unstable platform. Knowing when to escalate beyond design is a skill.
Affinity mapping is only as good as your notes
Sessions where the note-taker captured direct quotes versus paraphrases produced dramatically richer synthesis. The quality of observation shapes everything downstream.
Severity ratings need to be ruthless
Early drafts had too many "high severity" issues. Forcing triage, what would you fix first if you had one sprint, sharpened the recommendations into something actually actionable.
06 — CONCLUSION
Good enough isn't good enough, especially at the moment of purchase.
This study showed that Loblaws' desktop experience is functional but only barely, and only for persistent users. The friction wasn't random. It clustered around the highest-stakes moments: logging in, reaching the cart, moving through checkout. Those are exactly the moments a retailer can't afford to lose people.
​
Many of the issues we found weren't design mistakes, they were symptoms of an aging platform being stretched past its limits. That distinction matters. It means the fix isn't a new button colour or a reorganised nav. It's an infrastructure investment that unlocks every design improvement downstream.
​
If this study did one thing, it made that case with evidence, not assumptions.