Skip to content
MERVIN

Collectibles · Product Design · 2026

VIIV.

Collectors don’t need another spreadsheet. They need a shelf that remembers.

A home for a 1:64 diecast collection — catalogue, card scan, orders, spend and habits — designed and built end to end in seven weeks.

VIIV on a desktop and a phone: a grid of car photographs on a dark interface, and the phone home screen with status tiles

Real screens from the product’s demo mode (sample data).

My role
Product designer and owner — research, UX, UI, brand — with AI-assisted development (Claude Code)
Duration
8 Sep – 27 Oct 2026 · 355 commits
Industry
Collectibles
Platform
Responsive web app, installable on a phone's home screen
Tools
Claude Code · TanStack Start · React · Tailwind CSS · Supabase · Vercel · Razorpay · Gemini (card scan)

Overview

The product, plainly

What it is
VIIV (viiv.si) is a collection manager for 1:64 diecast cars: a shared catalogue of castings, a private collection with status, price and copy details, orders and pre-orders, spend and buying-habit insight, and a scanner that adds a car from a photo of its card. The name reads VI.IV — the Roman numerals for 6 and 4, the scale.
Why it exists
It began as a dashboard for my own list of cars and grew into a product when the limits of the spreadsheet became the brief: nothing connects the car you want, the one in the post and the one on the shelf.
Who it serves
Collectors of 1:64 cars who buy online and pre-order releases — brands in the catalogue include Hot Wheels, Mini GT, Matchbox, Tarmac Works and Kaido House — and who want to know what they own, what is arriving and what they have spent.
The problem
Collections are managed in spreadsheets and disconnected tools. The same car appears under different spellings, orders live outside the list, and entering a car by hand is slow enough that the list goes stale.
My contribution
Product definition, information architecture, interaction and visual design, the brand (wordmark, app icon, ViiV Orange), and the direction of an AI-assisted build across 355 commits. I designed every screen shown here.
  • 355

    commits in seven weeks (8 Sep – 27 Oct 2026)

  • 11 → 6

    statuses, with “Late” derived

  • 14

    fields that fill from one catalogue pick

  • 1,652

    castings in the shared catalogue when captured

Stage 1 of 5 · Empathise

Know the collector

Research and discovery

VIIV started with one collector and one spreadsheet — mine. This stage is therefore honest about what exists: first-hand experience, the limits of the spreadsheet, and a clear list of what I still haven’t asked other collectors.

What kind of evidence is in this stage

  • Research conducted
  • Competitive observation
  • Provisional persona
  • Proposed validation — not yet run

Reading the product backwards. Looking at what the finished product asks a collector to record shows what a collection turned out to be made of:

  • The car — make, model, series, colour, scale, brand, the box or blister it came in.
  • The copy you own — what you paid, the retail price (MRP), condition, rarity, photographs of your own car.
  • The order — seller, parcel, courier, tracking, the expected day.
  • The want — a car you are still searching for.
  • The habit — when and how much you buy.

The collecting loop

Conceptual diagram
Feeling
+−
Doing
  • Sees a new casting on a seller’s page
  • Checks whether they already own it
  • Places an order or a pre-order
  • Notes seller, price and date
  • Follows a parcel and a courier
  • Loses track of what is arriving
  • Reads the card
  • Types the details in
  • Browses the shelf
  • Wonders what the year has cost
Thinking
  • “Do I already have this one?”
  • “I’ll log it properly later.”
  • “What is coming this week?”
  • “Typing it all in is the chore.”
  • “How much have I spent?”
Opportunity
  • Look the casting up and see ownership before buying
  • Record it as it is placed, with an expected date
  • One shelf for what is on the way
  • Photograph the card; let the catalogue fill the rest
  • Spend, habits and a collection that looks like a collection
Provisional journey, assembled from my own collecting and from the features VIIV ended up needing. Not validated with other collectors.
Provisional personaA design artefact — not drawn from interviews

The release-day hunter

Buys online, pre-orders often

Follows releases closely and buys several cars a month from different sellers. Their collection is a stream of orders as much as a shelf.

Wants

  • Know instantly if a casting is already owned
  • See what is arriving and when
  • Add a car without typing everything

Struggles with

  • Orders scattered across messages and emails
  • Buying a duplicate by accident
  • Losing track of what has shipped
Provisional personaA design artefact — not drawn from interviews

The shelf curator

Buys less, cares about the collection as a whole

Thinks in series, brands and sets. Wants the collection to be browsable, complete and presentable.

Wants

  • See the collection by series, brand or set
  • Spot what is missing
  • Show the collection with photographs

Struggles with

  • A flat list that cannot be grouped
  • Inconsistent spellings of the same casting
  • No photographs next to the data

Key insights

  1. 01A collection is three things at once: a catalogue of castings, the copies you actually own, and a pipeline of things on the way. (Retrospective — read back from how the product is structured.)
  2. 02A spreadsheet row can hold a car, but not the difference between a casting, the copy you bought and the parcel it is arriving in.
  3. 03Entry is the costly step. If adding a car is slow, a collection tool goes stale. (My working assumption; not yet tested with other collectors.)

Decisions made

  1. 01Treat the absence of research as a design constraint: make everything easy to correct (undo, edit anywhere) rather than assuming the first model is right.
  2. 02Use my own collection as the first real dataset, and be explicit that it is one perspective.

With a clear picture of what I knew, what I had assumed and what I hadn’t asked, the next question was which problem to solve first.

Next: Define

Stage 2 of 5 · Define

Frame the problem

Problem framing and strategy

Three problems, in the order they hurt: getting a car in, knowing what you own, and knowing what is coming. The product principles below are the ones the commit history shows me actually committing to.

What kind of evidence is in this stage

  • Retrospective analysis
  • Proposed validation — not yet run

How might we…

  • …let a collector add a car in seconds, without typing what the card already says?
  • …decide, consistently, that two entries are the same casting?
  • …show what is arriving, from whom, and how late it is?
  • …let someone use the collection without ever exposing what they paid?
  1. 11 Sep

    The phone is a first-class screen

    Commit: “make the phone a first-class screen.” A bottom bar followed on 12 Sep.

  2. 12 Sep

    Photograph the card instead of typing it

    The scan flow starts here and is later renamed VIIV Scan (8 Oct).

  3. 18 Sep

    One catalogue entry per car, one Car ID per car owned

    The split between the shared casting and your own copy.

  4. 23 Sep

    Eleven statuses become six

    One shared status module instead of a dozen copies.

  5. 15 Oct

    What you paid can stay out of sight

    Privacy as a switch, not a setting buried in a menu.

Key insights

  1. 01Identity is the hard design problem. “What counts as the same car?” decides duplicates, ownership marks and what the catalogue can share.
  2. 02Money is personal. A collector may want the numbers, or may not want them on screen. (Observed: a hide-investment switch exists from the start.)
  3. 03Late is not a stage of a car’s life; it is a condition of an ordered car whose date has passed. The data model should say so.

Decisions made

  1. 01A casting belongs to everyone; a copy belongs to you.
  2. 02Photograph, don’t type — the camera is the primary way in.
  3. 03The phone is a first-class screen, not a shrunken desktop.

With the problem framed and the principles written down, the question became structure: what are the parts, and how do they connect?

Next: Ideate

Stage 3 of 5 · Ideate

Architect the collection

Exploration and solution architecture

The key move was separating what is shared from what is yours. Everything else — navigation, the add-a-car flow, the status model — follows from that split.

The domain model, as the commits record it. The shape changed more than once before it settled:

  • 18 Sep — one catalogue entry per car, one Car ID per car owned.
  • 18 Sep — a mix number is a case release, not a sub-series.
  • 26 Sep — the number on the box decides what is one casting.
  • 30 Sep — a casting comes in colours; a box is a name somebody keeps; the catalogue can be read as castings or as boxes.

The product ecosystem

Drawn from the shipped product

Shared catalogue

Everyone reads it; owners contribute

  • Castings
  • Boxes and assortments
  • Brand marks
  • Photographs
  • Public casting pages

Your collection

Private

  • Copies you own
  • Six statuses
  • Price paid and MRP
  • Condition and rarity
  • Favourites
  • Wants (ISO)

The pipeline

  • Orders and parcels
  • Courier and tracking
  • Pre-orders
  • Calendar of what is coming
  • Sellers

Insight

  • KPI tiles
  • Top five by make, model, series
  • Spend by month
  • Buying-habit heat map
  • Duplicates

Ways in

  • Guest mode
  • Scan a card
  • Add in bulk (CSV)
  • Welcome wizard
  • Free, trial and paid plans
Redrawn from VIIV’s shipped structure. The catalogue is shared by every account; everything below it is private to the collector.

Information architecture

Drawn from the shipped product
Status tilesStatus tilesTop 5Top 5Recently releasedRecently releasedHomeHomeIn HandIn HandIn TransitIn TransitOrderedOrderedOn HoldOn HoldPOPOISOISOMy CarsMy CarsFavouritesFavouritesSeriesSeriesSetSetBrandBrandAssortmentAssortmentMaker, seller, sizeMaker, seller, sizeCollectionCollectionBy parcelBy parcelBy purchaseBy purchaseMy OrdersMy OrdersCalendarCalendarPre OrdersPre OrdersDuplicatesDuplicatesHabitHabitSellersSellersCastingsCastingsBoxesBoxesCatalogueCatalogueVIIVVIIV
Hover a node to trace its path. Drawn from the app’s navigation.

Add a car

Drawn from the shipped product
Four ways in converge on one “Your copy” step. Hover a node to trace it.

After — six statuses

  • In Hand
  • In Transit
  • Ordered
  • On Hold
  • PO
  • ISO
  • “Late” is a badge worked out from the expected date, not a status
From the commit that made the change: every list, filter, tab and colour map had kept its own copy of whichever spellings it knew. Old data and old bookmarks still map onto the six.
  1. 8 Sep

    A dashboard for a spreadsheet

    First commit, “My Diecast Dashboard”.

  2. 9–10 Sep

    Accounts, per-user collections, guest mode, user-chosen handles

  3. 15–16 Sep

    The catalogue — added, removed, then returned as a shared page

    It was taken out of the navigation within a day, then came back as its own shared page once its job was clearer.

  4. 28 Sep

    Every casting gets a public page; a calendar of what is coming

    A stranger can read a casting without signing in; a collector can see their wants marked on a calendar.

  5. 7 Oct

    Tesoro becomes VIIV

    The working name was Tesoro; the product, wordmark and ViiV Orange arrive together.

  6. 13 Oct

    Blister view

    “A view shaped like the packaging” — the fourth way to look at your cars.

Key insights

  1. 01A catalogue entry (the casting) and a car you own (the copy) are different things with different owners. Treating them as one row is what makes spreadsheets fall over.
  2. 02Navigation should follow the collector’s questions — What do I have? What is coming? What have I spent? — not the database’s tables.
  3. 03Status is a small, closed vocabulary. Six words cover the life of a car; lateness is derived, never stored.

Decisions made

  1. 01Shared catalogue + private collection, with a sync between them.
  2. 02Add a car in two steps: which car, then your copy.
  3. 03Six statuses in one module; “Late” becomes a badge calculated from the expected date.

The architecture was settled enough to draw. Time to design the screens a collector actually touches.

Next: Design

Stage 4 of 5 · Design

Design the shelf

Wireframes, UI and prototyping

A dark, quiet interface where the cars are the colour and the orange is reserved for the next action. The same system runs from a wide desktop to a one-handed phone.

VIIV My Cars page in blister view, a grid of car photographs on a dark interface
My Cars in blister view — a view shaped like the packaging (13 Oct). Demo mode, sample data.

“I spent a LOT of time making sure it didn’t look like an AI-generated app. No random gradients. No generic dashboard.”

Mervin Kalaimani, on LinkedIn, about this product

Design tokens

Drawn from the shipped product
  • Inter · UI and display

    Know what you own.

Sampled from the running app. Click a swatch to copy its hex. Status colours are semantic (green In Hand, blue In Transit, and so on) and live in the product’s own module.
  • VIIV home on a phone with status tiles, recently added and top five
    Home
  • VIIV My Cars on a phone
    My Cars
  • VIIV Collection grouped by series on a phone
    Collection
Phone first: a bottom bar with Home, My Cars, My Orders and Habit, plus search. The same pages as desktop, re-composed rather than shrunk.
VIIV Add a car dialog with search, scan a card and add in bulk
Add a car: two steps — Which car, then Your copy. The three quick routes sit above the search result.
VIIV car details with a photograph, status chips and a Details, Purchase, Record tab bar
A car in three tabs. The photograph leads; status, chase and favourite sit beside it.
  • Tour step 1: Scan to add a car
  • Tour step 2: Bring your spreadsheet
  • Tour step 3: Track every shipment
  • Tour step 4: Know what you spend
  • Tour step 5: Catch duplicates
  • Tour step 6: See your collection
A six-step tour inside the product (27 Oct): scan, import, track, spend, duplicates, collection.
VIIV Buying habits page with a year heat map, top cars, top spend and top brands
Buying habits: a year heat map, streaks, top cars, brands and makes. Sample data.
VIIV Collection page grouping cars by series with counts and spend
The collection regrouped by series, set, brand, assortment, maker, seller or size.

Try VIIV in Guest Mode

Open the live demo

This product blocks embedding for security, so the demo opens in its own tab.

No account needed — choose Guest Mode on the sign-in screen. Sample cars, stored only in your browser.

Key insights

  1. 01Photographs carry the interface. With real car photos the grid needs almost no ornament.
  2. 02One page shape everywhere — heading, filters, content — means a collector learns the app once.
  3. 03Onboarding belongs inside the product: a six-step tour and a welcome wizard that only offers what the new collector picked.

Decisions made

  1. 01Dark by default, with light, OLED and auto appearances; orange only for action and identity.
  2. 02Four views of the same cars: blister, grid, compact grid, table.
  3. 03Car details in three tabs — Details, Purchase, Record — instead of one long page.

Designs this specific only mean something when someone tries them. What did testing look like — and what did it not?

Next: Test

Stage 5 of 5 · Test

Use it, break it, fix it

Validation, iteration and improvement

There has been no formal usability test of VIIV, and this stage says so. What there is: daily use on a real collection, a guest mode strangers can try, and a commit history that reads like a log of things used, broken and fixed.

What kind of evidence is in this stage

  • Research conducted
  • Retrospective analysis
  • Proposed validation — not yet run
  1. 10 Sep

    Keep a half-filled car through a phone going to standby

  2. 23 Sep

    Stop the add-a-car form wiping itself every few seconds

  3. 24 Sep

    Only a car in transit is asked for a courier

    A form should only ask what the status makes relevant.

  4. 26 Sep

    An import shows every field it will write, and can be taken back

  5. 29 Sep

    A metadata editor, because a spelling was permanent

  6. 29 Sep

    A duplicate has to be the same car — and says how close it is

  7. 13 Oct

    Uploading a card left the camera on

  8. 27 Oct

    A welcome wizard for accounts with no cars

Key insights

  1. 01Most of the fixes are about interruption: a phone going to standby mid-form, a form wiping itself, a camera left running. Real use finds the moments a happy-path design forgot.
  2. 02Correction is a feature. Undo, merge, a metadata editor and “take it back” on imports all arrived because a first version was wrong.

Decisions made

  1. 01Keep a half-filled car through a phone going to standby.
  2. 02Make imports reversible and show every field before it is written.
  3. 03Ship a guest mode so anyone can judge the product without signing up.

The work isn’t finished. Here is what I took from it.

What I learned

Reflection

What I took from it

01

What I learned

Identity is the real design problem in a collection tool. Deciding what counts as “the same car” shaped the catalogue, the duplicate check, public casting pages and even navigation. I also learned how much a product improves when you stay inside it: most of the best decisions here came from using it, hitting a wall, and fixing the wall.

And I learned the limit of that: one collector’s instincts are a start, not a study.

02

Key design decisions

  • A casting belongs to everyone; a copy belongs to you.
  • Photograph, don’t type.
  • Six statuses in one module; lateness derived, not stored.
  • Add a car in two steps, with four ways in.
  • Money can stay out of sight.
  • The phone is a first-class screen.

03

Product impact

Not measured. The product is in beta (v0.9) and I am not publishing usage numbers, because I have no verified ones to show. What I can state: it was designed and built in seven weeks, it runs publicly at viiv.si, and anyone can try it in Guest Mode.

04

Opportunities for future development

  • Interview and test with real collectors (plan under Test).
  • A structured review of existing collection tools.
  • Collector profiles that show a collection to others: handles and an account page exist; a public shelf does not yet.
  • Community beyond the shared catalogue — following, trading, wishlists visible to friends.
  • An accessibility audit of the dark theme and the scan flow.
“Good design isn’t just about creating interfaces. It’s about understanding people, connecting ideas and making complex experiences feel simple.”

Let’s talk

Have a challenge worth solving?

Whether you’re building a product, transforming an experience or looking for someone to help shape your design direction, let’s talk.