Crimibox: 1,000 Emails From a $200 Ad, 100+ Orders in the First 4 Hours of Launch — Building the Unsolved-Murder Box

Dollar figures throughout this article are converted at an approximate rate of $1 = ¥150.

Much of physical-product failure comes down to sequence — building inventory before finding buyers. Crimibox, out of Ghent, Belgium, is a case that deliberately reversed that order. Founder Jimmy Cowe put $200 (about ¥30,000) into Facebook ads before a single box was built, and collected over 1,000 email addresses. Four hours after sending an announcement to that list, over 100 orders had come in.

Cowe’s day job is detective with the Belgian federal police. Working unsolved cases, he realized that “most cases get solved not at the crime scene, but at the detective’s desk”, cross-checking testimony, verifying alibis, piecing together motive. His professional instinct that this desk work itself was the interesting part became the substance of the product.

Crimibox by the numbers

The structure of the product

Crimibox is a physical box containing the investigation files for a fictional murder case: case files, crime scene photos, evidence bags. Buyers work through it using a dedicated smartphone app while conducting the investigation. The app includes a chatbot and a fictional police database, populated with “thousands of fictional people and license plates.” Real-world web surfaces like social media profiles, surveillance footage, and Google Maps are also used as part of the investigative stage.

The design brings the escape-room experience into the home while having players move back and forth between the physical box and digital investigation. The $35 unit price is cheap next to a single escape-room session, and standard next to a board game.

The turning point was the $200 spent before building anything

The turning point in this case sits before launch, not after. Cowe ran Facebook ads before finishing the product and collected over 1,000 email addresses for $200. At that point, the question of “will this sell” was already, essentially, answered.

He then took pre-orders against that pre-launch list. The result: 100+ orders in four hours, roughly $3,500 (about ¥520,000) at the $35 unit price. That he was able to launch on just $5,000 of personal capital is because these pre-orders carried the inventory-purchasing funds in ahead of time. The before-and-after contrast is clean: $200 in ad spend pulled in the prepaid capital to start a business without carrying inventory risk.

The biggest lesson Cowe cites himself is exactly this. “The most important thing we learned is to start marketing before you launch. Build the email list before launch.”

Why this sequence worked

The product itself pairs well with pre-orders. For the buyer, an unsolved-murder box is entertainment with no urgency, so it’s fine even if delivery is weeks out. The same move doesn’t work for daily necessities or food.

The pre-launch list did the heavier lifting: it fundamentally changed the efficiency of advertising. At interview time, email subscribers numbered 5,000 and Messenger chatbot subscribers numbered 800. What stands out is the gap in open rates: 25% for email versus 95% for Messenger. Every time a new case ships, there’s a channel that reaches nearly the entire subscriber base with near-certainty. Given that designing one case takes about a month, the fact that over 38% of existing customers buy the next one too is inseparable from this reach.

Unusually low paid acquisition costs complete the picture. Facebook ads average $4 (about ¥600) per purchase, and Instagram Stories run $0.08 (about ¥12) per site visit. With a $35 unit price and 43% gross margin, a $4 acquisition cost means the ad spend is recovered on the very first purchase. On top of that, the site gets 300–400 visitors a day with a 4% conversion rate, for a niche product, that’s 1 in 25 visitors buying.

Put differently, Crimibox’s revenue structure runs on a single line: acquire cheaply, sell repeatedly at a high open rate. The pre-launch list was both the startup capital and the origin point of that line.

What isn’t working, and caveats on the numbers

One caution up front, on reading the numbers. The article’s body is from an interview conducted in November 2018, and monthly revenue is described as “a bit over $8,000.” Meanwhile, Starter Story’s profile listing shows $85,000/month. These are figures from two different points in time, and how it grew from one to the other isn’t disclosed. There’s a similar discrepancy on startup cost, the body says “$5,000” while the profile listing says “$18.9K.” This article presents both rather than picking one.

Operational issues remain too. At interview time, manufacturing and inventory were still handled out of Cowe’s home garage, and he himself says “it’s only a matter of time before we’re forced to move into a warehouse.” As a physical product, the more it sells, the more it runs into physical constraints. The 43% gross margin also has room to improve to 58% at scale, which, read the other way, means costs are currently running heavy.

On the platform side, the business moved from Shopify to WordPress/WooCommerce, citing “WordPress felt like it offered more options”, but rebuilding what was built during the launch phase carried real cost.

And the biggest structural risk is content exhaustion. In a product where each case takes about a month to design and the revenue pillar is repeat purchases (38%+), if new releases stop, revenue stops too. Unlike software, a box you’ve already built doesn’t multiply on its own.

Conditions for reproducing this

What’s easier to reproduce is the process itself: measure demand with ads before building anything, build a list, and fund inventory through pre-orders. This flow works regardless of the product category, and can be tested for as little as around $200. The idea of accumulating subscribers in a high-open-rate channel (Messenger, in this case) also applies directly elsewhere.

What’s harder to reproduce is the content itself. Cowe is an active police detective with professional depth in constructing cases. Whether an amateur building the same product could hit a 38% repeat rate is a separate question. On top of that, 2017-era Facebook ad prices ($4/purchase) are difficult to replicate at current levels. If acquisition cost triples, a $35 product at 43% margin no longer breaks even on the first purchase.

In short, what should be taken from this case isn’t “unsolved-murder boxes sell”. It’s the judgment criterion of “figure out whether your product category lets you secure both demand and funding, via a list and pre-orders, before you build any inventory.”

Related reading

• Launching a food truck in four months — a domestic case of getting a physical business running on a small amount of capital.
• ¥7.5M/year running a rental space — useful for comparing the revenue structure of a side-project physical business.