Usability
75Common tasks are reachable, but the primary task path has visible friction points that warrant attention.
UX Audit
Trust by DesignScentbird
Independent sample audit — not commissioned by or affiliated with Scentbird.
Independent UX audit · Trust by Design
Browse, checkout, account, mobile app · tailored audit deliverable.
- March 2026
- Browse, checkout, account, mobile app
- 12 findings · 5 tracks · tailored audit
Engagement date
Scope reviewed
Deliverable
Pricing and offer details verified as of March 2026.
36
Needs attention
/ 100Serious friction concentrated at the pricing card and checkout — the primary journey is workable, but recurring-price disclosure and card-switch recovery carry the highest risk.
- Findings
- 12
- High priority
- 7
- audit tracks
- 5
Skip the scroll. Open the full report as a PDF, or jump straight to the audit intake form.
Sample UX Audit
See exactly what a Trust by Design audit looks like.
This example shows the structure, scoring, findings, recommendations, priorities, and action plan included in a completed audit.
Methodology
How scoring works
Each audit is evaluated across six UX dimensions. Scores reflect the severity, frequency, and likely impact of the issues identified. The overall score gives teams a clear view of product friction and helps prioritize what to improve first.
The Trust by Design score is an expert evaluation score based on the issues identified during the audit. It is not an industry percentile or external benchmark.
Six dimensions, scored individually.
Scores evaluate the product experience, not the performance of the team that designed or built it.
A finding can reflect a defect, a business requirement, a technical limitation, a content decision, or a deliberate tradeoff. The audit separates the observed user impact from assumptions about why the experience exists.
- Confirmed
- Likely business or product constraint
- Assumption to confirm
- Confirmed implementation gap
- Required quality or compliance issue
Clarity
70Many screens carry the right meaning, but transition states and secondary actions could be clearer for first-time visitors.
AI Trust
Not applicableNot applicable — Scentbird's audited surfaces do not include an AI feature; this dimension is excluded from the score for this sample.
User Control
65Account-level actions like pause, cancel, and rating edits are reachable, but several are buried behind extra navigation.
Interaction Design
74Controls generally behave predictably, with a few places where the label does not match what the control does.
Errors & Recovery
60Failures during checkout and card switching do not currently tell the user what changed or what to do next.
Executive summary
12 findings · 5 audit tracksThree signals from the audit
A short read of where Scentbird's product is strong, where the friction clusters, and the two or three moves that compound across the rest of the report.
- Strong, recognizable brand voice carried consistently from the marketing site through the app shell.
- Browse and discovery read as the strongest area of the interface.
The trust the brand built is real equity; the gaps that follow sit adjacent to it, not underneath it.
- Recurring-price disclosure on the pricing card earns a Critical — first-time readers misread the offer.
- The next step after this toast is unclear, which makes recovery from this error harder than it should be.
- Mobile filter chips wrap past 375 px, so the last category is cut off-screen on iPhone SE viewports.
Two of those frictions sit at the highest-leverage moments of the funnel — fix cost is low, and the cost of letting them stay is the cost of lost checkout intent.
- A one-line subtitle on the pricing card, plus a rewrite of the checkout failure toast into an instruction, are the two moves worth shipping first.
- Coherent scent vocabulary and accessible focus rings compound across the rest of the funnel once those two land.
Together, those moves account for the highest-leverage gains in this audit; everything else is polish.
Top priorities
The issues worth addressing first
A short, ordered list of the highest-leverage moves — drawn from the full audit and ranked below the rows you'll find across the rest of the page.
- Critical/signup · pricing card
Recurring pricing not disclosed at the pricing card
The hero pricing card shows '$16 for your first fragrance' but never states that the subscription continues monthly until cancelled. Visitors encountering the pricing card for the first time are likely to misread the offer as a one-time order of a single 30 ml bottle.
- High/signup · step 2
Scent-quiz vocabulary drifts from product UI
The intake quiz asks users to pick from "Woody," "Fresh," "Floral," and three other categories; the catalogue filter on /scents uses the same words but mixes in "Earthy" and "Citrus." Subscribers reaching the catalogue cannot reliably find the scents the quiz recommended.
- High/account · "Your Rating" tab
'Your Rating' uses a 5-star scale; the rest of the app uses 0–10.
Two rating scales coexist: the /account history page uses 0–10 stars ("You rated: 7.5/10") while /scents shows the canonical 5-star community rating. Users who leave a private 8/10 expect to see it move the public 5-star average and feel misled when it does not.
- High/checkout · payment step
Card-switch failure surfaces an opaque toast
When a user attempts to switch from a saved card to a new one, the failure toast reads "Could not update your payment method." It does not name a cause. The current payment-failure toast reads as a status, not an instruction. Users pausing here are likely to reach for support rather than recover on their own.
- High/checkout · review step
Order total visibly mismatches expected total
Across promo-apply behaviour, the Review step occasionally shows a total that does not match the sum of the line items, and shoppers notice.
- High/scents · filter sheet
Filter sheet traps focus but has no visible close cue
When the filter sheet opens over /scents, focus moves into the sheet but the close (×) button carries no border on focus. The close affordance lacks a visible focus indicator, making the keyboard traversal of the sheet visually ambiguous.
Findings · by track
12 findings · 5 tracksTwelve prioritized issues across five tracks
Each finding includes the observed user impact, context status, working ownership, a practical fix within known constraints, and a preferred solution if those constraints change. Severity is our confidence-weighted signal of where shipping matters most, drawn from the rendered surface and the copy that surrounds it.
SeverityCritical× 1High× 6Medium× 4Low× 1
Impact · Estimated effort · ConfidenceLowMediumHigh
Effort is a working estimate based on the visible experience. Development should confirm it against the codebase, dependencies, and release process.
Trust & onboarding
3 findingsFirst-session signals that determine whether a visitor believes they're signing up for a $16/month subscription vs. a $9.95 one-time purchase. The current copy makes this harder than it needs to be.
Where · /signup · pricing card
Recurring pricing not disclosed at the pricing card
CriticalImpact: HighEstimated effort, to confirm: LowConfidence: High
The hero pricing card shows '$16 for your first fragrance' but never states that the subscription continues monthly until cancelled. Visitors encountering the pricing card for the first time are likely to misread the offer as a one-time order of a single 30 ml bottle.
Where · /signup · step 2
Scent-quiz vocabulary drifts from product UI
HighImpact: MediumEstimated effort, to confirm: MediumConfidence: Medium
Validate with: content audit across quiz, catalogue, and account
The intake quiz asks users to pick from "Woody," "Fresh," "Floral," and three other categories; the catalogue filter on /scents uses the same words but mixes in "Earthy" and "Citrus." Subscribers reaching the catalogue cannot reliably find the scents the quiz recommended.
Where · /signup · success state
No next-action cue after first quiz completion
MediumImpact: MediumEstimated effort, to confirm: LowConfidence: Medium
Validate with: user testing on quiz completers
After the quiz finishes, the page jumps straight to checkout. There is no thank-you panel, no preview of the recommended scents, and no CTA for "skip the line — go to catalogue."
Information architecture & copy
3 findingsTerminology varies across account, billing, and cancellation surfaces; legal copy crowds the primary real estate.
Where · /account · "Your Rating" tab
'Your Rating' uses a 5-star scale; the rest of the app uses 0–10.
HighImpact: MediumEstimated effort, to confirm: LowConfidence: High
Two rating scales coexist: the /account history page uses 0–10 stars ("You rated: 7.5/10") while /scents shows the canonical 5-star community rating. Users who leave a private 8/10 expect to see it move the public 5-star average and feel misled when it does not.
Where · /account · "Subscription" tab
Subscription status is buried under a sub-tab
MediumImpact: MediumEstimated effort, to confirm: LowConfidence: High
Three clicks deep — /account → Subscription → "Plan details" — to discover whether you are on the 1-per-month plan, the 2-per-month plan, or paused. The plan-details state is buried under three clicks, and the discovery path is not obvious on entry.
Where · /legal · refund policy
Refund policy split across three pages
LowImpact: LowEstimated effort, to confirm: LowConfidence: Medium
Validate with: legal review
Refund language is split between /legal/refunds, /legal/subscriptions, and the footer "Help" link. /legal/refunds covers opened bottles; the other two diverge on shipping damages.
Conversion & checkout
3 findingsThe four screens that close the order — address, shipping, payment, review. The audit found real friction here, concentrated around payment-method switching and error recovery.
Where · /checkout · payment step
Card-switch failure surfaces an opaque toast
HighImpact: HighEstimated effort, to confirm: LowConfidence: Medium
Validate with: support ticket review and prototype test
When a user attempts to switch from a saved card to a new one, the failure toast reads "Could not update your payment method." It does not name a cause. The current payment-failure toast reads as a status, not an instruction. Users pausing here are likely to reach for support rather than recover on their own.
Where · /checkout · review step
Order total visibly mismatches expected total
HighImpact: HighEstimated effort, to confirm: MediumConfidence: Medium
Validate with: QA repro across every promo combination
Across promo-apply behaviour, the Review step occasionally shows a total that does not match the sum of the line items, and shoppers notice.
Where · /checkout · shipping step
Shipping options use marketing copy, not labels
MediumImpact: MediumEstimated effort, to confirm: LowConfidence: High
"Free ground shipping (5–7 days)" and "Express in 2 days" — neither label is hover-discoverable as the actual radio. The current labelling does not expose the option as a true radio control; the surrounding paragraph carries the meaning visually, and assistive technology users may not pick up the option the same way mouse users do.
Accessibility
2 findingsObservations consistent with a brand-led interface, focused on the visual polish of the rendered surface. Recommendations below propose additions for keyboard reach, contrast, and assistive technology.
Where · /scents · filter sheet
Filter sheet traps focus but has no visible close cue
HighImpact: HighEstimated effort, to confirm: LowConfidence: High
When the filter sheet opens over /scents, focus moves into the sheet but the close (×) button carries no border on focus. The close affordance lacks a visible focus indicator, making the keyboard traversal of the sheet visually ambiguous.
Where · /account · icon-only buttons
Icon-only "edit" buttons have no accessible name
MediumImpact: MediumEstimated effort, to confirm: LowConfidence: High
Several icon-only buttons on /account lack accessible names for assistive technology.
Mobile responsiveness
1 findingsThe responsive story on iPhone SE-sized layouts (375 px) has gaps. The same responsive behavior may affect related mobile surfaces. Confirm during implementation review.
Where · /scents · mobile sheet
Filter chip stack overflows at 375 px
HighImpact: HighEstimated effort, to confirm: LowConfidence: High
On a 375 px viewport, the filter chips wrap to a third row whose last chip is cut off on the right edge.
Priority roadmap
Now, Next, and Later — sequenced against confidence
Three horizons sized by confidence and effort. Now is the smallest ship; Next is a sustained slice; Later is the long tail that compounds if not addressed.
Now
Ship blockers · highest confidence impact
Disclose recurring pricing on the signup card
Critical
Critical: add the "Then $16/month. Pause or cancel anytime." subtitle to the hero pricing card.
Replace opaque payment-failure toast
High
High: rewrite the "Could not update your payment method" failure as an inline banner with named cause + support escape.
Patch filter sheet focus trap
High
High: add a visible focus ring on the close button and mirror across all overlay closing affordances.
Next
High-confidence wins · sustained slice of work
Unify rating scales (5-star everywhere)
High
Translate /account private ratings to the public 5-star scale and add a tooltip explaining the relationship.
Unify scent vocabulary across quiz and catalogue
High
Pick six mutually-exclusive buckets; reuse them everywhere the scent category is shown.
Animated totals line on promo apply
Medium
Animate the totals line when promo codes apply and break the total into per-line items.
Surface subscription status on /account landing
Medium
Add a status chip — payment, plan, next ship — straight on the account landing page.
Later
Cumulative polish · longest tail wins
Insert post-quiz confirmation panel
Medium
Add a one-step panel between the quiz and checkout that summarises matches and offers two CTAs.
Lift icon-only buttons to accessible names
Medium
Add aria-label to every icon-only edit button on /account.
Consolidate refund policy into one page
Low
Move /legal/refunds-subscriptions and footer "Help" copy under a single /legal/refunds page.
Switch mobile filter chips to a snap-scroller
High
Replace the wrapped chip row with a horizontally-scrollable snap row + scroll cue.
Next step
See what your product is missing.
Get the same structured review for your website, app, AI product, prototype, or Figma design.