Case study · 2026. Me as a founder – checking if my physical product audience actually exists. Gore Dog. Live · testing with ads.
TL;DR. Before: idea / concept. 2 weeks. After: live, working experiment. In two weeks, I created a working concept for a product that doesn't exist yet – and I'm currently testing it with ads. Will it take off? The data will tell! PLN 15 – my target: max cost per item added to the cart.
TL;DR. Before: idea / concept. 2 weeks. After: live, working experiment. In two weeks, I created a working concept for a product that doesn't exist yet – and I'm currently testing it with ads. Will it take off? The data will tell! PLN 15 – my target: max cost per item added to the cart.
AI in this project. How was AI used in this project? 01 Figma – iterating on the layout. 02 FigJam – iterating on the concept. 03 Claude Code – vibecoding, duh! 04 Meta Ads – bringing in the users. 05 PostHog – measuring user behaviour. 06 ChatGPT – generating the product images.
AI in this project. How was AI used in this project? 01 Figma – iterating on the layout. 02 FigJam – iterating on the concept. 03 Claude Code – vibecoding, duh! 04 Meta Ads – bringing in the users. 05 PostHog – measuring user behaviour. 06 ChatGPT – generating the product images.
Chapter 01: Problem. Subcultures went mainstream. Dogs became children. Do the two meet?
Chapter 01: Problem. Subcultures went mainstream. Dogs became children. Do the two meet?
The problem. Is there an audience for the new goth/metal style – but in the context of dogs? Axis – collar with two rows of studs.
The problem. Is there an audience for the new goth/metal style – but in the context of dogs? Axis – collar with two rows of studs.
Context. Two trends meet in the city. 01 Subcultures went mainstream: I don't think I need to convince anyone that the fashion styles of certain subcultures have broken free from those subcultures and fully entered the mainstream. That's actually one of the topics of my master's thesis in psychology, which I'm writing right now ;) 02 The dog as a child: people in cities have increasingly begun to adopt specific styles. At the same time, a very significant trend in cities is treating one's dog like a child. The Polish word for it: Psiecko – pies (dog) + dziecko (child).
Context. Two trends meet in the city. 01 Subcultures went mainstream: I don't think I need to convince anyone that the fashion styles of certain subcultures have broken free from those subcultures and fully entered the mainstream. That's actually one of the topics of my master's thesis in psychology, which I'm writing right now ;) 02 The dog as a child: people in cities have increasingly begun to adopt specific styles. At the same time, a very significant trend in cities is treating one's dog like a child. The Polish word for it: Psiecko – pies (dog) + dziecko (child).
The real question. But do your style preferences extend to your dog? Can a dog become an extension of your image and sense of style?
The real question. But do your style preferences extend to your dog? Can a dog become an extension of your image and sense of style?
Chapter 02: Validation. How do you test demand for a product that doesn't exist yet?
Chapter 02: Validation. How do you test demand for a product that doesn't exist yet?
Where AI comes in. The collars in these photos don't exist. Yet. A mock sale needs a product people can believe in. For a physical product, that used to mean moulds, samples and a photo shoot. With AI, I could show the product, build the store and bring in buyers – before casting a single collar. Image generated with ChatGPT.
Where AI comes in. The collars in these photos don't exist. Yet. A mock sale needs a product people can believe in. For a physical product, that used to mean moulds, samples and a photo shoot. With AI, I could show the product, build the store and bring in buyers – before casting a single collar. Image generated with ChatGPT.
About validation. So I started working on validating the concept. I began by talking to dog owners and people with specific styles. But how can you validate a product as quickly as possible – without investing thousands of dollars in metal casting and production machinery? That's where business product validation comes in!
About validation. So I started working on validating the concept. I began by talking to dog owners and people with specific styles. But how can you validate a product as quickly as possible – without investing thousands of dollars in metal casting and production machinery? That's where business product validation comes in!
The method: Mock Sale. “Presenting a sale for your product without processing any payment information.” It lets you test fairly faithfully whether someone would buy a product – without having the product yet. Smart and agile. D. J. Bland & A. Osterwalder – Testing Business Ideas (Wiley, 2019), pp. 288–289. Cost, setup time, run time.
The method: Mock Sale. “Presenting a sale for your product without processing any payment information.” It lets you test fairly faithfully whether someone would buy a product – without having the product yet. Smart and agile. D. J. Bland & A. Osterwalder – Testing Business Ideas (Wiley, 2019), pp. 288–289. Cost, setup time, run time.
Chapter 03: Build. Product images, a store and the ads – built in two weeks.
Chapter 03: Build. Product images, a store and the ads – built in two weeks.
Step 1 · Product visuals. So I created the product visuals. Three products – two collars and a coat – generated with ChatGPT: studio shots, close-ups and lifestyle photos. Draco – collar with forged scales. Axis – collar with two rows of studs. Misery – a warm dog coat with a metal soul.
Step 1 · Product visuals. So I created the product visuals. Three products – two collars and a coat – generated with ChatGPT: studio shots, close-ups and lifestyle photos. Draco – collar with forged scales. Axis – collar with two rows of studs. Misery – a warm dog coat with a metal soul.
Step 2 · The store. Then, a fake online store. Designed in Figma, coded with Claude Code. It looks and works like a real shop – except nothing can be paid for. 11 page types · desktop + mobile.
Step 2 · The store. Then, a fake online store. Designed in Figma, coded with Claude Code. It looks and works like a real shop – except nothing can be paid for. 11 page types · desktop + mobile.
The mock sale funnel. From product page to waitlist: 01 Shop → view product → 02 Product page → add to cart (key event) → 03 Reservation → sign me up → 04 Waitlist.
The mock sale funnel. From product page to waitlist: 01 Shop → view product → 02 Product page → add to cart (key event) → 03 Reservation → sign me up → 04 Waitlist.
Step 3 · Meta Ads. And set up all the ads in Meta Ads. One metric that matters: cost per add-to-cart.
Step 3 · Meta Ads. And set up all the ads in Meta Ads. One metric that matters: cost per add-to-cart.
Chapter 04: Measure. What the data has to show before a single collar is made.
Chapter 04: Measure. What the data has to show before a single collar is made.
What I'm measuring. What does the experiment measure? 01 (most important) How often do users add a product to the cart? 02 Which sizes do they choose – and do they combine products? 03 How much is a customer willing to pay for the product? 04 Do they sign up for the waitlist? A real declaration – although many people may not be keen on reserving. Question 01 decides whether Gore Dog goes on. The other three shape what gets made.
What I'm measuring. What does the experiment measure? 01 (most important) How often do users add a product to the cart? 02 Which sizes do they choose – and do they combine products? 03 How much is a customer willing to pay for the product? 04 Do they sign up for the waitlist? A real declaration – although many people may not be keen on reserving. Question 01 decides whether Gore Dog goes on. The other three shape what gets made.
The decision rule. I'll keep building Gore Dog if an add-to-cart costs me PLN 35 or less. PLN 15 – target. PLN 35 – go / no-go line. Why PLN 35? It comes from a detailed breakdown of production costs and my preferred return on investment. Bonus: I can also swap products and learn what people really want to buy – before any production starts.
The decision rule. I'll keep building Gore Dog if an add-to-cart costs me PLN 35 or less. PLN 15 – target. PLN 35 – go / no-go line. Why PLN 35? It comes from a detailed breakdown of production costs and my preferred return on investment. Bonus: I can also swap products and learn what people really want to buy – before any production starts.
What I've learned so far. Two things this project reminded me of. 01 I love product validation and building products from scratch – across disciplines, too. 02 The Meta Ads panel is the least readable system on the planet.
What I've learned so far. Two things this project reminded me of. 01 I love product validation and building products from scratch – across disciplines, too. 02 The Meta Ads panel is the least readable system on the planet.
