UI / UX Design
John Glasgow Car Sales
A UX evaluation and concept redesign for an independent used-car dealership in Ballyclare: two responsive Framer prototypes that make it easier to trust a car and book a test drive.
Year :
2026
Industry :
Automotive retail
Client :
John Glasgow Car Sales
Project Duration :
5 hours

JOHN GLASGOW CAR SALES – EASY TO TRUST, EASY TO BOOK.
A concept redesign for an independent used-car dealer
John Glasgow runs a small used-car dealership in Ballyclare. Customers visit by appointment, Monday to Saturday, so the website has one job: help a cautious buyer trust the cars and book a time with John.
I evaluated the existing site, redesigned the key journey (browse, check a car, book a test drive) and built two responsive prototypes in Framer that work on phone, tablet and desktop. I designed with accessibility in mind from the start.
Try the prototypes:
Concept A: Bold – high-impact, with number-plate style prices
Concept B: Calm – quieter and more conventional, built around clarity
This is an unofficial concept, not the live John Glasgow Car Sales website, shared with John’s permission. Vehicles marked “sample” are illustrative.
Process: I managed the project in ClickUp. Each step (interviewing John, the heuristic evaluation, user walkthroughs, the accessibility audit, Lighthouse scores, the competitor review and the prototypes) was a task holding my notes and evidence. I also tested a ClickUp AI agent to run the competitor review. See the ClickUp board.
Role: UX/UI designer (solo)
Software: Figma, Framer, HTML/CSS, ClickUp, Claude
Skills: Heuristic evaluation, Accessibility design, Information architecture, UX & UI, Responsive design, Prototyping, UX writing.

RESEARCH & DISCOVERY
For a car dealer, the website is the shop window. Buyers decide quickly whether a seller feels honest, and a clear next step matters more than a long feature list.
Talking to John. His site was set up for him by UsedCarsNI and he had never changed it. It hadn’t brought in a single enquiry; his sales come through UsedCarsNI itself. He also told me most of his customers start their search on their phone. When I asked him to try his own site, he got stuck in the filters and noticed for the first time that his logo wasn’t loading.
Heuristic evaluation. I screen-recorded my own journey on my phone and reviewed it against Nielsen’s 10 usability heuristics, rating each finding from 0 (not a problem) to 4 (blocks the user or damages trust). There was an issue under every heuristic, and 8 of the 10 were rated major or worse. The most serious: the site and Google list different postcodes, so a buyer using satnav could end up in the wrong place.
Accessibility and performance. I mapped the findings to WCAG 2.2 AA and prioritised them into 16 issues, 4 of them critical. White text on the site’s blue was around 3.99:1, below the 4.5:1 minimum, and the filter controls had no accessible names, so a screen reader only announced “button”. Lighthouse (mobile) scored 45 for performance: the main content took 17.9 seconds to appear on a 5.6 MB page.
Three walkthroughs. I was on holiday in Spain, so I asked three friends to use the site, each with a different goal: browsing at their own pace, ready to buy, and comparing across platforms. The one ready to buy couldn’t book a test drive at all: the required date dropdown opened empty. The filter also needed two taps to apply and never showed what was selected. Comparing platforms showed listings that didn’t match, and John’s 5-star rating on other sites appeared nowhere on his own.
Competitor review. I compared the site with UsedCarsNI, Charles Hurst Usedirect and Maguire’s Garage. Patterns worth borrowing: filters with counts and a clear “X cars” result, visible phone and contact options, and specific reassurance rather than broad promises.
Goals
Make booking a test drive actually work, in about two minutes and with no account.
Show price, mileage and key specs up front so buyers can judge a car in seconds.
Make filters apply in one tap and show what’s selected.
Build trust with real, consistent details: logo, address, hours, phone and Google rating.
Design for the phone first, and meet accessibility basics: contrast, tap targets, labels and clear errors.


CONCEPT & IDEATION
Rather than commit to one look too early, I explored three directions that share the same structure and journeys but test different ways of earning trust:
Bold: big type, a black and yellow palette and prices styled like number plates. Designed to be memorable and confident, with the price impossible to miss.
Calm: a lighter, more conventional layout with a blue accent, closer to what buyers expect from a dealer site.
Clear and trusted: an accessibility-first direction with a highly legible typeface, text-size and high-contrast controls, a three-step booking form with a review step, and an FAQ that answers common worries up front.
John wanted to keep his original logo, so all three concepts work around it. He also liked the number-plate price style, which I kept in the bold concept.
The site structure is built around three journeys: find a car → view details → enquire; find a car → book a test drive; and contact John directly.
DESIGN DEVELOPMENT
Lead with the booking path. Every page carries a “Book a test drive” button in the header, and the booking page shows the car being booked beside the form, so the buyer’s goal stays visible.
A short, honest booking form. The original form had 10 required fields, including Title, Confirm Email and Comments. Mine has 5: name, phone, preferred day, preferred time and consent. Email is optional, because John confirms by phone or message. The button says what it does: “Request test drive”.
Filters that just work. One tap applies a filter (All, Petrol, Diesel, Under £5,000). Each option shows how many cars it holds, the selected filter is highlighted, and a “Showing X of Y cars” line confirms the result.
Price and specs on the card. Each card shows the full price, year, mileage, gearbox and fuel, and is far shorter than the original, which took more than a phone screen per car. Two cars are real listings; the others are labelled as sample data.
Errors that help. An error summary at the top of the form plus a message beside each field, replacing the original generic “Please fill in all required fields” pop-up. Errors use an icon and text, never colour alone, and say how to fix the problem (“Enter a phone number John can call”).
Trust through real details. John’s own logo, working this time, his Google rating, address, phone number and “by appointment” hours, stated the same way on every page.
Accessible by default. I checked contrast ratios by calculation (for example, white on the calm concept’s blue is 7.14:1, against 3.99:1 on the original), kept buttons at least 44px tall, gave every input a visible label and made focus states clear.
Responsive from the start. Designed for desktop, tablet and phone: card grids reflow, the header stacks and the booking form becomes a single column.



OUTCOME & REFLECTION
This is a concept, so I can’t claim business results and I won’t invent any. What I can show is the design evidence: the critical issues from the audit are addressed in the prototypes, contrast ratios are calculated, tap targets are sized, and layouts were checked at three widths.
With John’s go-ahead, I would measure booking-form completion, calls from the phone button and time to a first booking request, and rerun Lighthouse against the original scores.
Limitations
Contrast was calculated, not yet tested on real devices or with assistive technology.
Most car images are placeholders; the real site has photos of every car.
The prototypes haven’t been tested with users yet. Next step: put them in front of the same three people and compare.
Reflection
While I wanted to be bold with the design, I also wanted to keep accessibility in mind throughout. John was happy to see different looks and feels, so I used the bolder concept to play with branding and design direction.
Realistically, a small dealership like John’s doesn’t need a website that screams innovative design, but a clear and easy experience is a must. A memorable look is a plus.
What I’d explore next is connecting the places John already advertises his cars, like UsedCarsNI, to the new website, so listings are updated automatically across every platform. That would also fix one of the biggest trust problems I found: prices, details and stock that don’t match from one site to the next.
How I used AI: I used Claude to give me a starting baseline for the heuristic evaluation and to summarise the walkthrough notes, then corrected and added to the findings myself. I tested a ClickUp AI agent for the competitor review, and used Claude to speed up prototyping in Framer. The design decisions and the relationship with John stayed with me.
More Projects
UI / UX Design
John Glasgow Car Sales
A UX evaluation and concept redesign for an independent used-car dealership in Ballyclare: two responsive Framer prototypes that make it easier to trust a car and book a test drive.
Year :
2026
Industry :
Automotive retail
Client :
John Glasgow Car Sales
Project Duration :
5 hours

JOHN GLASGOW CAR SALES – EASY TO TRUST, EASY TO BOOK.
A concept redesign for an independent used-car dealer
John Glasgow runs a small used-car dealership in Ballyclare. Customers visit by appointment, Monday to Saturday, so the website has one job: help a cautious buyer trust the cars and book a time with John.
I evaluated the existing site, redesigned the key journey (browse, check a car, book a test drive) and built two responsive prototypes in Framer that work on phone, tablet and desktop. I designed with accessibility in mind from the start.
Try the prototypes:
Concept A: Bold – high-impact, with number-plate style prices
Concept B: Calm – quieter and more conventional, built around clarity
This is an unofficial concept, not the live John Glasgow Car Sales website, shared with John’s permission. Vehicles marked “sample” are illustrative.
Process: I managed the project in ClickUp. Each step (interviewing John, the heuristic evaluation, user walkthroughs, the accessibility audit, Lighthouse scores, the competitor review and the prototypes) was a task holding my notes and evidence. I also tested a ClickUp AI agent to run the competitor review. See the ClickUp board.
Role: UX/UI designer (solo)
Software: Figma, Framer, HTML/CSS, ClickUp, Claude
Skills: Heuristic evaluation, Accessibility design, Information architecture, UX & UI, Responsive design, Prototyping, UX writing.

RESEARCH & DISCOVERY
For a car dealer, the website is the shop window. Buyers decide quickly whether a seller feels honest, and a clear next step matters more than a long feature list.
Talking to John. His site was set up for him by UsedCarsNI and he had never changed it. It hadn’t brought in a single enquiry; his sales come through UsedCarsNI itself. He also told me most of his customers start their search on their phone. When I asked him to try his own site, he got stuck in the filters and noticed for the first time that his logo wasn’t loading.
Heuristic evaluation. I screen-recorded my own journey on my phone and reviewed it against Nielsen’s 10 usability heuristics, rating each finding from 0 (not a problem) to 4 (blocks the user or damages trust). There was an issue under every heuristic, and 8 of the 10 were rated major or worse. The most serious: the site and Google list different postcodes, so a buyer using satnav could end up in the wrong place.
Accessibility and performance. I mapped the findings to WCAG 2.2 AA and prioritised them into 16 issues, 4 of them critical. White text on the site’s blue was around 3.99:1, below the 4.5:1 minimum, and the filter controls had no accessible names, so a screen reader only announced “button”. Lighthouse (mobile) scored 45 for performance: the main content took 17.9 seconds to appear on a 5.6 MB page.
Three walkthroughs. I was on holiday in Spain, so I asked three friends to use the site, each with a different goal: browsing at their own pace, ready to buy, and comparing across platforms. The one ready to buy couldn’t book a test drive at all: the required date dropdown opened empty. The filter also needed two taps to apply and never showed what was selected. Comparing platforms showed listings that didn’t match, and John’s 5-star rating on other sites appeared nowhere on his own.
Competitor review. I compared the site with UsedCarsNI, Charles Hurst Usedirect and Maguire’s Garage. Patterns worth borrowing: filters with counts and a clear “X cars” result, visible phone and contact options, and specific reassurance rather than broad promises.
Goals
Make booking a test drive actually work, in about two minutes and with no account.
Show price, mileage and key specs up front so buyers can judge a car in seconds.
Make filters apply in one tap and show what’s selected.
Build trust with real, consistent details: logo, address, hours, phone and Google rating.
Design for the phone first, and meet accessibility basics: contrast, tap targets, labels and clear errors.


CONCEPT & IDEATION
Rather than commit to one look too early, I explored three directions that share the same structure and journeys but test different ways of earning trust:
Bold: big type, a black and yellow palette and prices styled like number plates. Designed to be memorable and confident, with the price impossible to miss.
Calm: a lighter, more conventional layout with a blue accent, closer to what buyers expect from a dealer site.
Clear and trusted: an accessibility-first direction with a highly legible typeface, text-size and high-contrast controls, a three-step booking form with a review step, and an FAQ that answers common worries up front.
John wanted to keep his original logo, so all three concepts work around it. He also liked the number-plate price style, which I kept in the bold concept.
The site structure is built around three journeys: find a car → view details → enquire; find a car → book a test drive; and contact John directly.
DESIGN DEVELOPMENT
Lead with the booking path. Every page carries a “Book a test drive” button in the header, and the booking page shows the car being booked beside the form, so the buyer’s goal stays visible.
A short, honest booking form. The original form had 10 required fields, including Title, Confirm Email and Comments. Mine has 5: name, phone, preferred day, preferred time and consent. Email is optional, because John confirms by phone or message. The button says what it does: “Request test drive”.
Filters that just work. One tap applies a filter (All, Petrol, Diesel, Under £5,000). Each option shows how many cars it holds, the selected filter is highlighted, and a “Showing X of Y cars” line confirms the result.
Price and specs on the card. Each card shows the full price, year, mileage, gearbox and fuel, and is far shorter than the original, which took more than a phone screen per car. Two cars are real listings; the others are labelled as sample data.
Errors that help. An error summary at the top of the form plus a message beside each field, replacing the original generic “Please fill in all required fields” pop-up. Errors use an icon and text, never colour alone, and say how to fix the problem (“Enter a phone number John can call”).
Trust through real details. John’s own logo, working this time, his Google rating, address, phone number and “by appointment” hours, stated the same way on every page.
Accessible by default. I checked contrast ratios by calculation (for example, white on the calm concept’s blue is 7.14:1, against 3.99:1 on the original), kept buttons at least 44px tall, gave every input a visible label and made focus states clear.
Responsive from the start. Designed for desktop, tablet and phone: card grids reflow, the header stacks and the booking form becomes a single column.



OUTCOME & REFLECTION
This is a concept, so I can’t claim business results and I won’t invent any. What I can show is the design evidence: the critical issues from the audit are addressed in the prototypes, contrast ratios are calculated, tap targets are sized, and layouts were checked at three widths.
With John’s go-ahead, I would measure booking-form completion, calls from the phone button and time to a first booking request, and rerun Lighthouse against the original scores.
Limitations
Contrast was calculated, not yet tested on real devices or with assistive technology.
Most car images are placeholders; the real site has photos of every car.
The prototypes haven’t been tested with users yet. Next step: put them in front of the same three people and compare.
Reflection
While I wanted to be bold with the design, I also wanted to keep accessibility in mind throughout. John was happy to see different looks and feels, so I used the bolder concept to play with branding and design direction.
Realistically, a small dealership like John’s doesn’t need a website that screams innovative design, but a clear and easy experience is a must. A memorable look is a plus.
What I’d explore next is connecting the places John already advertises his cars, like UsedCarsNI, to the new website, so listings are updated automatically across every platform. That would also fix one of the biggest trust problems I found: prices, details and stock that don’t match from one site to the next.
How I used AI: I used Claude to give me a starting baseline for the heuristic evaluation and to summarise the walkthrough notes, then corrected and added to the findings myself. I tested a ClickUp AI agent for the competitor review, and used Claude to speed up prototyping in Framer. The design decisions and the relationship with John stayed with me.
More Projects
UI / UX Design
John Glasgow Car Sales
A UX evaluation and concept redesign for an independent used-car dealership in Ballyclare: two responsive Framer prototypes that make it easier to trust a car and book a test drive.
Year :
2026
Industry :
Automotive retail
Client :
John Glasgow Car Sales
Project Duration :
5 hours

JOHN GLASGOW CAR SALES – EASY TO TRUST, EASY TO BOOK.
A concept redesign for an independent used-car dealer
John Glasgow runs a small used-car dealership in Ballyclare. Customers visit by appointment, Monday to Saturday, so the website has one job: help a cautious buyer trust the cars and book a time with John.
I evaluated the existing site, redesigned the key journey (browse, check a car, book a test drive) and built two responsive prototypes in Framer that work on phone, tablet and desktop. I designed with accessibility in mind from the start.
Try the prototypes:
Concept A: Bold – high-impact, with number-plate style prices
Concept B: Calm – quieter and more conventional, built around clarity
This is an unofficial concept, not the live John Glasgow Car Sales website, shared with John’s permission. Vehicles marked “sample” are illustrative.
Process: I managed the project in ClickUp. Each step (interviewing John, the heuristic evaluation, user walkthroughs, the accessibility audit, Lighthouse scores, the competitor review and the prototypes) was a task holding my notes and evidence. I also tested a ClickUp AI agent to run the competitor review. See the ClickUp board.
Role: UX/UI designer (solo)
Software: Figma, Framer, HTML/CSS, ClickUp, Claude
Skills: Heuristic evaluation, Accessibility design, Information architecture, UX & UI, Responsive design, Prototyping, UX writing.

RESEARCH & DISCOVERY
For a car dealer, the website is the shop window. Buyers decide quickly whether a seller feels honest, and a clear next step matters more than a long feature list.
Talking to John. His site was set up for him by UsedCarsNI and he had never changed it. It hadn’t brought in a single enquiry; his sales come through UsedCarsNI itself. He also told me most of his customers start their search on their phone. When I asked him to try his own site, he got stuck in the filters and noticed for the first time that his logo wasn’t loading.
Heuristic evaluation. I screen-recorded my own journey on my phone and reviewed it against Nielsen’s 10 usability heuristics, rating each finding from 0 (not a problem) to 4 (blocks the user or damages trust). There was an issue under every heuristic, and 8 of the 10 were rated major or worse. The most serious: the site and Google list different postcodes, so a buyer using satnav could end up in the wrong place.
Accessibility and performance. I mapped the findings to WCAG 2.2 AA and prioritised them into 16 issues, 4 of them critical. White text on the site’s blue was around 3.99:1, below the 4.5:1 minimum, and the filter controls had no accessible names, so a screen reader only announced “button”. Lighthouse (mobile) scored 45 for performance: the main content took 17.9 seconds to appear on a 5.6 MB page.
Three walkthroughs. I was on holiday in Spain, so I asked three friends to use the site, each with a different goal: browsing at their own pace, ready to buy, and comparing across platforms. The one ready to buy couldn’t book a test drive at all: the required date dropdown opened empty. The filter also needed two taps to apply and never showed what was selected. Comparing platforms showed listings that didn’t match, and John’s 5-star rating on other sites appeared nowhere on his own.
Competitor review. I compared the site with UsedCarsNI, Charles Hurst Usedirect and Maguire’s Garage. Patterns worth borrowing: filters with counts and a clear “X cars” result, visible phone and contact options, and specific reassurance rather than broad promises.
Goals
Make booking a test drive actually work, in about two minutes and with no account.
Show price, mileage and key specs up front so buyers can judge a car in seconds.
Make filters apply in one tap and show what’s selected.
Build trust with real, consistent details: logo, address, hours, phone and Google rating.
Design for the phone first, and meet accessibility basics: contrast, tap targets, labels and clear errors.


CONCEPT & IDEATION
Rather than commit to one look too early, I explored three directions that share the same structure and journeys but test different ways of earning trust:
Bold: big type, a black and yellow palette and prices styled like number plates. Designed to be memorable and confident, with the price impossible to miss.
Calm: a lighter, more conventional layout with a blue accent, closer to what buyers expect from a dealer site.
Clear and trusted: an accessibility-first direction with a highly legible typeface, text-size and high-contrast controls, a three-step booking form with a review step, and an FAQ that answers common worries up front.
John wanted to keep his original logo, so all three concepts work around it. He also liked the number-plate price style, which I kept in the bold concept.
The site structure is built around three journeys: find a car → view details → enquire; find a car → book a test drive; and contact John directly.
DESIGN DEVELOPMENT
Lead with the booking path. Every page carries a “Book a test drive” button in the header, and the booking page shows the car being booked beside the form, so the buyer’s goal stays visible.
A short, honest booking form. The original form had 10 required fields, including Title, Confirm Email and Comments. Mine has 5: name, phone, preferred day, preferred time and consent. Email is optional, because John confirms by phone or message. The button says what it does: “Request test drive”.
Filters that just work. One tap applies a filter (All, Petrol, Diesel, Under £5,000). Each option shows how many cars it holds, the selected filter is highlighted, and a “Showing X of Y cars” line confirms the result.
Price and specs on the card. Each card shows the full price, year, mileage, gearbox and fuel, and is far shorter than the original, which took more than a phone screen per car. Two cars are real listings; the others are labelled as sample data.
Errors that help. An error summary at the top of the form plus a message beside each field, replacing the original generic “Please fill in all required fields” pop-up. Errors use an icon and text, never colour alone, and say how to fix the problem (“Enter a phone number John can call”).
Trust through real details. John’s own logo, working this time, his Google rating, address, phone number and “by appointment” hours, stated the same way on every page.
Accessible by default. I checked contrast ratios by calculation (for example, white on the calm concept’s blue is 7.14:1, against 3.99:1 on the original), kept buttons at least 44px tall, gave every input a visible label and made focus states clear.
Responsive from the start. Designed for desktop, tablet and phone: card grids reflow, the header stacks and the booking form becomes a single column.



OUTCOME & REFLECTION
This is a concept, so I can’t claim business results and I won’t invent any. What I can show is the design evidence: the critical issues from the audit are addressed in the prototypes, contrast ratios are calculated, tap targets are sized, and layouts were checked at three widths.
With John’s go-ahead, I would measure booking-form completion, calls from the phone button and time to a first booking request, and rerun Lighthouse against the original scores.
Limitations
Contrast was calculated, not yet tested on real devices or with assistive technology.
Most car images are placeholders; the real site has photos of every car.
The prototypes haven’t been tested with users yet. Next step: put them in front of the same three people and compare.
Reflection
While I wanted to be bold with the design, I also wanted to keep accessibility in mind throughout. John was happy to see different looks and feels, so I used the bolder concept to play with branding and design direction.
Realistically, a small dealership like John’s doesn’t need a website that screams innovative design, but a clear and easy experience is a must. A memorable look is a plus.
What I’d explore next is connecting the places John already advertises his cars, like UsedCarsNI, to the new website, so listings are updated automatically across every platform. That would also fix one of the biggest trust problems I found: prices, details and stock that don’t match from one site to the next.
How I used AI: I used Claude to give me a starting baseline for the heuristic evaluation and to summarise the walkthrough notes, then corrected and added to the findings myself. I tested a ClickUp AI agent for the competitor review, and used Claude to speed up prototyping in Framer. The design decisions and the relationship with John stayed with me.












