Chapter 4 · Case Study: Claw Crushers

A second real claw arcade — and this one answers the questions Prizemate left open. It needed a city permit before it could open. Its machine has a name you can look up. Its token ladder turns out to stop rewarding bigger spend, and its prize wall has a deliberately bad middle tier.

Learning objectives

• Understand that a claw arcade is a regulated use, not just a shop — and see a real city permit file for one.
• Learn how to identify a commercial crane from a photograph now that you know where the names actually live.
• Read a token ladder and work out your true cost per play — including the tier that quietly gives you less.
• Understand a trade-up prize wall and why the middle bundle is the worst deal on it.
• Compare three ways into this business: independent outlet mall, independent ethnic-anchor mall, and franchise.

Why study a second store at all

A single example teaches you one shape and lets you mistake it for the rule. Prizemate — an independent claw shop in a Simon outlet mall — showed the format working with a cheap lease and an outlet-mall foot traffic. The obvious objection is: fine, but that is one store, in one kind of mall, in one city, and everything about it was reconstructed from Instagram posts.

Claw Crushers is a fair test of that objection. It is also small and independent, but nearly everything else is different. It sits in an ethnic-anchor neighbourhood mall rather than a premium outlet centre. It has been open long enough to accumulate 44 reviews, where Prizemate has essentially none. It needed an explicit government permit. And because a customer photographed the machines, the counters, and the machines' own labels, it lets us check claims that Prizemate's material could never settle.

Part 1 · The store

Claw Crushers trades as suite C-127 inside Pacific East Mall, 3288 Pierce Street, Richmond, California. It sits well inside the building: the store's own listing points you "towards the back of the mall, past the 99 Ranch market." That is a normal, sensible location for this format — deep in the building, on the path between a grocery anchor and the food court, where families already walk.

The mall itself matters to the economics. Pacific East Mall opened in 1998, is managed by GD Commercial Real Estate, and has just completed a full interior and exterior renovation. Its published amenities include free WiFi, outdoor dining, community events, a highway-visible digital billboard, and Tesla Superchargers. Mall hours are 9:00 AM to 10:00 PM, seven days a week.

A long mall day is worth pausing on. An arcade inside a centre that is open thirteen hours can trade whenever the mall is open, and its peak hours are the mall's peak hours. You are not paying for a storefront that is dark half the day.

Part 2 · An arcade needs a permit, not just a lease

This is the part of the Claw Crushers story you will not find in any guide that is written from the operator's side, because operators rarely mention it and landlords do not volunteer it.

In January 2025 the Richmond planning department issued a Notice of Public Hearing for project PLN25-003, "CLAW CRUSHERS ARCADE." It is an ordinary public document and it is unusually informative. The project is a request for an Administrative Use Permit:

Read that sentence twice. The city classified the business as commercial entertainment — not retail — and made it a use that required an explicit permit. The applicant was Claw Crushers, LLC. The property owner was Pacific Infinity Center, LLC, and the site's assessor's parcel number is 510-060-006.

The notice was heard by the Zoning Administrator on Tuesday 4 February 2025 at 10:00 AM, over Zoom, and mailed to every property owner within 300 feet of the site, as Richmond's municipal code requires. The staff planner was named in the document.

The environmental review is the cheerful part. The project was declared categorically exempt from CEQA — the California Environmental Quality Act — under Guidelines Section 15301, Existing Facilities, "involving no expansion or interior alterations to an existing facility." In plain terms: because they moved into a finished space and did not knock through a wall or change the building's footprint, they skipped California's environmental review entirely.

Part 3 · Identifying the machine

Chapter 3 ended on a frustration. The Prizemate cabinets carry no manufacturer's name anywhere visible — the marquee, the cabinet, the decals all belong to the shop, and the only remaining place to look was a serial plate inside the coin door that nobody had photographed. The closest match available at the time was a product by a factory that might or might not have built them, and I flagged it as not confirmed.

Claw Crushers has the same problem, and then solves it in the one place that actually works.

The marquees all lie

Walk in and you see "CLAW CRUSHERS" on the marquee, "CLAW CRUSHERS" on the cabinet sides, "COLLECT THEM ALL" above the glass. None of it is the manufacturer. All of it is the shop's own branding, printed on panels that any factory will print to order.

But look at the control module — the little apron that carries the joystick and the button — and the name changes. Every station on the double-decker machine reads "Lucky Claw." That is not the shop's brand. It is the machine's.

What "Lucky Claw" actually is

Here is the trap, and it is the same trap as the "GOOD LUCK" marquee discussed in the previous chapter. "Lucky Claw" is not a manufacturer. It is a product name used by several unrelated factories. Searching it turns up at least four: Guangzhou EPARK, Guangzhou Dinibao ("Lucky Claw Pink"), Guangzhou Ama Amusement ("Popular Coin-Operated Lucky Claw"), and Lifun, whose near-identical product is called "Lucky Gift." There is also a genuine vintage machine, the UPL "Lucky Crane," from the 1980s, with no relation to any of them.

So "Lucky Claw" tells you the machine came from the Guangzhou cluster, and nothing more. It is a catalogue name, not an identity.

Where the identification actually comes from

The identification did not come from the cabinet at all. It came from putting the control module next to a manufacturer's product photograph. EPARK publishes its own picture of the "Lucky Claw" four-person crane, and the match against the store photo is unmistakable on the details that are hard to fake:

• A white cabinet with LED accent trim.
• A clear faceted "crystal" joystick ball — an unusual, very specific component.
• A large round illuminated action button.
• The wordmark "Lucky Claw" printed on the control-module shelf, in the same stylised lettering.

That is Model EP-G149, from Guangzhou EPARK Electronic Technology Co., Ltd. One honest discrepancy: EPARK's own photo shows violet accent lighting where the store's machine is blue. That is not a contradiction — accent colour is on the factory's own list of customisable options, along with cabinet, logo, and claw size. It is the same cabinet, ordered in a different colour.

The rest of the manufacturer's specification is what you would want before importing anything: a 1-year warranty, CE and RoHS certification, a 100 × 80 × 220 cm crate, and a stated production capacity of 500 sets a month. The factory describes itself as thirteen years old and reports that it will build to order — "customize cabinet, logo, color, and claw sizes (S to XXL)" — with a minimum order quantity of one set.

The price, and the number that corrects this guide

Here you have to be careful, because the sourcing pages on these marketplace sites put a lot of prices on the screen at once and most of them belong to something else.

The EPARK listing shows a volume ladder of US$600 per set at 1–19 sets, US$400 at 20–49, and US$300 at 50+. It is tempting to present that as "the price of a Lucky Claw." It is not. That block sits under a generic heading in the page furniture and is not tied to the Lucky Claw product, which carries no published price at all. The US$200–US$800 figures elsewhere on the same page belong to other products in the "also viewed" strip. Quoting any of them as this machine's price would be a mistake.

What can be said honestly is that a single commercial crane of this class sits in a broad band from roughly US$200 to US$800 per set ex-works, corroborated from a second direction by Lifun's comparable "Lucky Gift" cabinet and by TKY's stated minimum order of one set.

Where the US price really sits

That factory figure is not what you pay if you buy in America, and this is where a domestic dealer becomes genuinely useful rather than merely expensive. HighFiveJF, a reseller operating out of Rancho Cucamonga in California, publishes its prices on the open web, which almost nobody in this trade does:

Both are listed 110V and in stock in California, with local pickup available and warehouse visits by appointment. That last detail matters more than it reads: you can put your hands on the machine before you pay for it, which is not true of anything you import.

Read down that table and the earlier discrepancy resolves itself. The factory price is real, the reseller price is real, and the higher figure is real too — it is simply the wrong default. Plan against the middle rung and you are neither importing blind nor over-reserving.

A third cabinet, and the point of the comparison

The reseller's four-player machine is not the store's machine, and the difference is instructive. It is a double-decker with the same play geometry — two stacked chambers, four control stations, four joysticks, four prize-out doors, white cabinet, clear faceted joystick balls — but its marquee reads "Panda Park," with its own panda mascot, pink and cyan trim, and rainbow graphics.

So three cabinets now sit side by side, all the same broad format, all wearing different names:

• Prizemate's — white-label, no identifiable OEM.
• Claw Crushers' — EPARK EP-G149, carrying two prize-out doors and four "Lucky Claw" station aprons.
• "Panda Park" — a reseller product with four prize doors, sold under a brand I could not trace to any manufacturer at all.

Note the structural difference, because it is a real design choice and not a detail: the store's EP-G149 gives two players per tier a shared prize door, while the "Panda Park" cabinet gives every station its own. That is worth thinking about before you buy. A shared door means fewer holes in the cabinet and less to maintain, but it also means the child standing at the left joystick watches a prize they won slide out of a door their sibling is standing next to.

And the naming pattern holds a third time. "Panda Park" is the seller's or factory's product name, not an identity — exactly like "Lucky Claw" and "GOOD LUCK" before it. The marquee tells you what someone decided to call the cabinet. It does not tell you who built it.

Part 4 · The token ladder, and the tier that rewards you less

Most arcades hide their pricing. This one puts it on the counter, in a plastic stand, where anyone can photograph it — which is how we have these numbers.

Five bundles. You pay a price, you get tokens, and at the upper end you also get bonus points, which are a separate currency for the prize wall in the next part.

Look at the bonus column, because it does something unusual. It climbs — 10%, 17%, 20%, 21% — and then falls back to 20% at the hundred-dollar tier.

That is not a rounding artefact. At 85 tokens for $70 you pay $0.824 a token; at 120 tokens for $100 you pay $0.833. The $100 bundle is slightly worse value per token than the $70 bundle. The gap is small — about one percent — but it is real, and it is precisely the kind of thing a customer grabbing "the big one" will never check.

The design is legible once you see it. The discount ladder ramps to a peak and then plateaus, so the best value in the store sits at $70. That tier maximises the amount of money a customer leaves behind while giving away the smallest possible discount. The store is not trying to sell you the biggest bundle; it is steering you to the sweet spot.

And the cheapest tier is the worst value in the store by a clear margin: at $20 you pay $0.909 a token against $0.824 at the top of the ladder, so the casual visitor subsidises the regular. A 10.4% spread between the best and worst entry point.

What a play actually costs

Tokens alone do not tell you the price of a game. You need the machine's per-play charge, and here the machine answers for itself: the control-panel display reads "Insert Token" above "1 Credit 1 play."

One token buys one credit, and one credit buys one play. So the cost of a play is the cost of a token:

Roughly 82 to 91 cents a go, depending on how much you buy at once. That is the whole pricing story of this store in one number, and it is a useful benchmark: it tells you what a small independent arcade actually charges, in a real market, in 2026 — as opposed to what a business plan assumes.

The two rules that do the real work

Two lines on the same sign do more for this business than the entire pricing table.

The first, in a red box: "ALL TOKENS MUST BE USED THE SAME DAY! thank you :)" A small yellow sticky note beside it clarifies the arithmetic: "BONUS TOKENS ALREADY INCLUDED."

Same-day expiry is the quiet engine of arcade margin. Tokens are pre-paid revenue, and any token a customer buys and does not spend is money kept with no prize given. In a venue where a family buys a $50 bundle and plays for forty minutes, the leftover tokens are not a liability you carry — they are pure margin. The rule converts breakage from an accident into a policy.

The second line reads "UPGRADE TIME! Trade today's plushies for cool upgrades! Only today's wins count!" — and that one is worth a whole section, because it is the entire second half of the business model.

Part 5 · The trade-up prize wall

The wall behind the counter is a row of large plush on shelves, with handwritten price cards beneath them. They do not sell for cash. They sell for points, and points are earned two ways: as a bonus at the higher token tiers, and by trading in the small plush you already won.

Here is the whole menu, transcribed from the wall signs:

Two things fall out of this table, and only one of them is intentional.

The first is clean. Points per plushie is 2.0 at every tier — six plushies is twelve points, eight is sixteen, twelve is twenty-four. A "point" is always worth exactly half a plushie, in every bundle. That is a well-built internal accounting unit, and it tells you the operator thinks of the trade-in as a single currency rather than a set of ad-hoc deals.

The second is the interesting one. Look down the cost-per-plush column: $4.17, then $4.38, then $3.75. It is not monotonic. The eight-plushie bundle is the worst deal on the wall, seventeen percent worse per plushie than the twelve-bundle, and worse even than the smallest bundle.

Whether that is a deliberate trap or simply an arithmetic slip, the effect is the same: a customer who takes "the middle one, that looks reasonable" pays the highest price per plush in the store. The twelve-plushie bundle is the value play at $3.75 a plush — and it is also the one that requires you to have won the most.

The prize mix is the product

The wall itself is a well-chosen licensed-character set: Pokémon (Eevee), Sanrio (Hello Kitty, Kuromi, Hangyodon), Disney (Stitch, Lotso) and Marvel (Iron Man), in two sizes — jumbo on the top shelf, medium below. Those are brands a child already wants, which means the wall does not have to sell itself.

Inside the machines, the same logic runs in miniature. Prizes are themed per machine — one machine all-Pikachu, another all-pink, another all-white, another all-orange — under a "COLLECT THEM ALL" banner. Machines are individually numbered, a small detail that matters more than it looks: numbering exists so staff can track which unit is paying, and so a machine that is not performing can be identified.

Part 6 · Two machine generations, and the double-decker question

The store runs at least two different cabinet generations side by side, and the difference between them is instructive.

The double-decker

The machine photographed at length is a genuine double-decker: two glass chambers stacked, one above the other, in a single tower. Count what is actually on it:

• Two glass prize chambers.
• Four play stations — two control modules per tier, one left and one right.
• Two prize-out doors, one centred in each tier.
• Four joysticks, four buttons, four credit displays.

That format — two chambers, four play positions, two chutes — is the same one shown on the Prizemate cabinets in the previous chapter. It is a high-throughput shape: four people play at once, and a family of four can stand at one machine.

The tiered prize ladder, built into the cabinet

The double-decker does something quietly clever. The upper chamber holds the larger prizes and the lower chamber holds the small ones, sitting on a bed of loose white packing material. Lower tier, smaller prize.

That is the trade-up wall's logic expressed physically. A child does not need to read a price card to understand that the big penguins are upstairs and the keyring charms are down below, and that the way up is to keep playing. The machine teaches the value ladder through the glass.

The packing material is a small operational trick worth copying: loose fill stops small prizes from packing into a dense unwinnable mound at the bottom of the chamber, and it makes every prize sit somewhere a claw can plausibly reach.

The newer uprights

The second generation is a single-chamber, single-player upright. One chamber, one joystick, one button, one prize-out door, roughly six feet tall. It carries the same shop branding, and the same firmware message — its display also reads "1 Credit 1 play" — but the cabinet is a newer design, with blue and orange accent lighting and a mascot set (tiger, robot, dinosaur) instead of the double-decker's graphics.

The detail that marks it as a newer generation is inside the prize chamber: a flat-panel monitor mounted in the ceiling, above the claw, where it plays attract animation to the player looking in. That is not in the EP-G149 specification. These are different machines from a different factory, and the store is running old and new cabinets in the same room — which is exactly what you would expect from an operator who bought an initial kit and added to it later.

Reading a cabinet's own signals

These machines do their marketing on their own front panels:

• A "CLAW • WIN • REPEAT" band on the control apron, and the same words repeated along the chamber's interior edges.
• A "COLLECT THEM ALL" banner across the top of the glass.
• A glowing orange "WINNER" medallion on the lower front of the cabinet.
• Circular speaker grilles with amber LED halos flanking each tier.

That "WINNER" medallion is doing more than decorating. It announces wins to the room. The social proof that a big claw shop gets from a scoreboard, a small store gets from the front of the cabinet — everyone in the aisle sees when someone wins, and the machine's own lighting tells them to try.

Where the manufacturer marks actually are — and are not

We found one genuine part-level identifier: the coin door carries "JS 133A" stamped on the acceptor. That is the model of the coin comparator, a component — not of the machine. It is a nice illustration of the chapter's central point: the only legible name on a claw machine may belong to a parts vendor twelve levels down the bill of materials.

Everything else confirmed the rule from Chapter 3. There is no serial plate, no UL label, no model number and no manufacturer's nameplate visible anywhere on these machines. The claw assembly does carry a CE mark, and the gripper is the standard three-prong design — two fingers forward, one back, rubber-tipped, on a gantry trolley with a coiled cable conduit. But that tells you the machine complies with European norms and roughly which decade the design is from. It does not tell you who built it.

Part 7 · Three ways in

With two independent stores documented and a franchise system publishing numbers, this guide can finally lay out the actual choice in front of someone who wants to open one of these.

The franchise column is the one that has been unavailable until now. Fantasy Claw Arcade publishes a total cost to open starting at US$292,000, inclusive of the franchise fee; a franchise fee of $30,000 multi-unit or $40,000 single; and a royalty plus marketing of 6% multi-unit or 7% single, plus 2%, taken on net sales. It asks for at least $100,000 in liquid capital. And it reports, in its 2026 franchise disclosure document, a top store at $1.1 million and five operating locations, with two more announced for Ohio and Hawaii.

What the two independent columns have in common is that neither published its capital cost, which is normal for a small private operator. That is the real trade: the independent route keeps the margin and the equity, and pays for it in uncertainty. The franchise route buys a known cost and a known playbook, and pays for it in fees on every dollar of net sales, forever.

The franchise also confirms something this guide has asserted from the start, from a direction nobody expected. Fantasy reports having served more than 100,000 customers and generating more than 10 million social views, and describes its format as built for "malls, lifestyle centres and entertainment districts." A company raising money on that story is telling you where it believes this business works.

Common mistakes

• Signing a lease before checking the use. An arcade is commercial entertainment to a planning department. Confirm it is permitted at that address before you sign anything.
• Trying to identify a machine by its marquee. The marquee is the operator's. Look at the control module, and then go find the manufacturer's own product photograph.
• Assuming a catalogue name is a brand. "Lucky Claw" is used by at least four unrelated factories. So was "GOOD LUCK."
• Quoting a marketplace price without checking which product it belongs to. Those listing pages show many products and many prices, and the block nearest the top is often not the product you are reading about.
• Comparing stores on the price of a token. Compare cost per play — read the machine's display for tokens per play.
• Buying the biggest bundle because it is the biggest. Here it is worse value per token than the tier below it.
• Copying a prize wall price at random. The middle bundle on this one is the worst deal on the wall.
• Letting tokens roll over forever. Same-day expiry is what makes the pre-paid token a margin product rather than a liability.
• Planning against a vendor's payback claim. "2 to 3 months" is sales copy with no methodology behind it.

Summary

• Claw Crushers is an independent claw arcade inside a renovated Asian-anchor neighbourhood mall, open 13 hours a day, targeting local families.
• It required an Administrative Use Permit from the City of Richmond to operate as a commercial entertainment facility, and was exempted from environmental review because it made no structural alterations.
• The machine is an EPARK EP-G149, identified not from the cabinet but by matching its control module against the manufacturer's own product photograph.
• The machine price spans three rungs: $200–$800 ex-works, $1,299–$1,899 from a US reseller, and $2,900–$3,500 as a top-of-market fitted estimate. Plan against the middle rung.
• The token ladder ramps to a peak at $70 and then falls back; the $100 bundle is worse value per token than the one below it.
• One token is one play, so the real price is 82–91 cents per play.
• Tokens expire the same day, which converts unspent tokens from a liability into margin.
• The trade-up wall converts small wins into bigger plush, and its middle bundle is the worst deal on it.
• The store runs two machine generations, and the double-decker teaches the prize ladder physically — bigger prizes upstairs.

Checklist

• ☐ Confirmed with the city that an arcade is a permitted use at your chosen address.
• ☐ Asked whether it needs a permit or a public hearing, and put that on your opening timeline.
• ☐ Chosen between outlet-mall traffic and neighbourhood-mall repeat custom, deliberately.
• ☐ Requested ex-works, freight, and landed cost separately for any machine quotation.
• ☐ Compared a factory quotation against a US reseller's landed price before deciding which route to take.
• ☐ Measured the machine yourself — reseller spec sheets do contain copied-and-wrong dimensions.
• ☐ Worked out your own cost per play, not just your cost per token.
• ☐ Designed a token ladder that does not punish your biggest spenders by accident.
• ☐ Decided your token expiry policy and written it where customers can see it.
• ☐ Planned a trade-up or redemption path with a value ladder that actually rises.
• ☐ Set your claw difficulty deliberately, and considered what you promise in return.
• ☐ Numbered your machines so you know which ones are earning.

Looking ahead

Two stores, two malls, two machine generations, and one franchise system with published numbers. Chapter 5 takes everything this volume has shown you — the economics, the traps, and both worked examples — and turns it into a concrete, week-by-week launch plan.

Sources

• City of Richmond, Notice of Public Hearing, project PLN25-003 "Claw Crushers Arcade" — ci.richmond.ca.us/Archive.aspx?ADID=14502 (Richmond Community Development Dept., Planning Division)
• Pacific East Mall — store directory listing for Claw Crushers; leasing page (GD Commercial Real Estate) and mall amenities — visitpacificeast.com
• Claw Crushers location, hours, phone and review count — MapQuest, Roadtrippers, Yahoo Local listings; Yelp business page (44 reviews)
• Token bundle pricing, same-day token rule, trade-up wall pricing, machine identification and per-play cost — customer photographs of the store (provided to this project, October 2026)
• Machine identification and specification — Guangzhou EPARK Electronic Technology Co., Ltd., "Lucky Claw Arcade Crane Game Machine for 4 Persons," model EP-G149 (eparkgames.en.made-in-china.com); EPARK lucky-claw application page (eparkgames.com)
• Competing uses of the "Lucky Claw" product name — Guangzhou Dinibao Animation Technology; Guangzhou Ama Amusement Equipment (made-in-china.com listings)
• Comparable cabinet specification and minimum order quantity — Lifun, "Lucky Gift LED Claw Crane"; TKY claw machine manufacturer (lifunarcadegame.com, tkyclawmachine.com)
• Vintage machine of similar name, unrelated — UPL "Lucky Crane" (Arcade Museum forums)
• US reseller pricing and spec sheets, and the "Panda Park" cabinet — HighFiveJF, Rancho Cucamonga CA (highfivejf.com)
• Franchise cost, fees and performance — Fantasy Claw Arcade franchising page and 2026 FDD Item 19 reference (fantasyclawarcade.com/franchising); PR Newswire release, "Fantasy Claw Arcade Opens at AREA15," 4 May 2026
• Store events and community programming — Funcheap SF, "Claw Crushers Halloween Spooktacular," 25 October 2025

Chapter 4 · Case Study: Claw Crushers · Claw Machine Business Guide