Problem: Finding the right clothes can be hard.

Answer: An app that lets the user take photos of an item of clothing (whether from a magazine, online or in the real world) and used computer vision to find visually similar products.

Business model: Initially, affiliate commission from product sales. Later, we had talks about a B2B offering, embedding the technology on retailers' websites, but it didn’t pan out.

Status: Dissolved

We shut it down after realising there wasn’t sufficient customer demand to buy products through the app.

Learnings:

Figure out earlier what implicit customer assumptions needed to be tested:

Explore alternative business models sooner rather than waiting until you’ve released your product already and the market isn’t interested.

Don’t automatically do things just because a business advisor says you should — even if they’re wearing a suit.

Be straight up and open about the limitations of your product. Don’t hype it up to be something it isn’t — that leads to frustrated customers.

Problem: Finding the right balance of macronutrients to suit one’s personal dietary goals can be hard.

Answer: An app that lets you specify what foods you have and a desired range of how much you want to eat, then tries to determine an optimal plan that satisfies your desired dietary parameters (calorie intake and ratios of carbs/fats/protein).

Business model: SaaS monthly subscription.

Status: Just a landing page for now, but I actually implemented this with some R scripts and spreadsheets back in 2017. I’m working on a prototype using Phoenix LiveView now, but it’s really early stage.