Case · Kitzoomer

Kitzoomer
a store I designed, priced and ran for five years.

Role · Owner & operatorScope · Product, pricing, art direction, opsSpan · 2021 — 2026

months trading, unbroken

¥1.54M

gross, five years

110

listings priced and shipped

441

paying members, in parallel

Most of my work carries a client's numbers. This one carries mine.

A product business I owned end to end. I set the prices, wrote the listings, made the card art, shipped the releases and read the panel every month — for five years, with nobody to sign any of it off. In 2026 I audited the whole thing line by line and changed the roadmap on what came back.

The catalogue is 3D character assets sold to VTubers, and that market has a shape worth knowing: every paid item competes with thousands of free ones, so the only things that hold a price are the ones the tool doesn't ship with. That isn't a taste call. It's the economics of the category — and it's what the audit below put a number on.

01 · Sixty-one months

Monthly gross sales for the Kitzoomer store, 61 consecutive months from June 2021 to June 2026, peaking at 56,600 JPY in August 2021 and averaging 15,508 over the last 24 months — Every month of the store, gross and unsmoothed. The peak year averaged 41,608 JPY a month; the plateau that followed is what the 2026 audit set out to explain. My reading of it is cadence, not season: it starts when the releases stop — which makes it a problem with a lever, not a trend.

02 · What the audit found

The best sellers were never the hairstyles.

Three original listings priced identically at 500 JPY compared by units sold: rigged ears and tail 179, long hair with kitty ears 147, plain brunette hair preset 95 — Three originals priced identically at 500 JPY, so units compare directly: the accessories sold 1.9× the plain preset. A plain hairstyle competes with thousands of free ones; rigged ears and a tail compete with almost nobody — and the demon set proves the same point at 2.4× the ticket. Cumulative per-product figures from the shop's panel.

So the next production run is accessories, not more hairstyles. The lever was never the price or the catalogue size — it was making the thing the tool doesn't ship with.

And one decision I chose to stage instead of take.

Most of the revenue sat in a line I had already decided to retire. Pulling it in one move would have cost around 20,000 JPY a month — real money for a shop this size — so I set a threshold instead: the new line has to sustain 10–15k a month on its own before the old one comes down. The decision was never whether. Only what it should cost.

03 · The listing system

Nine Kitzoomer product listing images shown small as a set, split into two successive card systems: a type-led one from 2021–22 and a badge-led one from late 2022 — Nine of the shop's own listing images at the size they're meant to be read as a set — what survives is the structure, not the artwork. Two successive systems: the first type-led, telling you what's in the folder; the second badge-led, showing you the variants instead. The redesign spends the contents block's space on the one thing the first system made a buyer guess at.

Five years of running the whole loop:
make it, price it, ship it, read the numbers, change your mind.

Free items were the entry product, not a loss — one of them holds 1,562 favourites and built a following of 2,576. Alongside the shop I ran a paid membership of 441, opened five months before the store took its first order — two ways of charging for the same craft, running at once. What this proves isn't scale: it's that every decision in the loop was mine to make, mine to measure, and mine to correct.

Shop figures read from the shop's own sales panel · JPY, gross · Membership is a public platform count, not a panel reading

All work Start a project