Domain: Consumer
Status reached: Advanced validation
When: Feb 2023-June 2023 (intense work. Several months of ideation before).
Team skillset: Founder with senior marketing background
Thesis/Idea
Solution for healthy and efficient home-cooked meals for busy families
Started with a straightforward problem that many families face: Ensuring that you have healthy, home-cooked meals is a challenge for couples where both have full-time jobs, particularly those with children. The market was couples where both people work, have kids 3-12 years old, and have a good income but not so high that they would have a cook or other high-end options.
I started with the concept of a meal plan. These exist but usually cater to people dieting or to foodies. The idea was to offer meal plan solution designed for the needs of families.
We built something very lean as an MVP and quickly realized the meal plan necessitated too much work from consumers, for whom even answering 7-8 questions about their preferences was often too much.
When ChapGPT came onto the scene, it seemed like that might be an interesting option, though in retrospect it might have confused us. Started wondering if we can do something bigger with AI by learning what a family needs/wants. For two months, in another MVP, the goal was to see if we can have AI learn the person/family quickly and try to do smart meal planning. There was a strong component of chat telling you what to do when plus getting the shopping done. The suggestions would be some combination of things you can cook in a short amount of time, ordering some meals, giving you tips on how to make meals happen (put a pot of quinoa on in the evening and then you can have it for lunch tomorrow). The MVP was 2-3 families on WhatsApp with AI logic in the background.
Another idea/MVP was to try the B2B2X route – go through supermarkets or delivery apps to get info on families.
Why it didn’t progress
- The product was, and still is, very interesting to me, but it’s not big enough to build a company around it. Might be a feature.
- While VC money was needed to develop the product, it wasn’t going to yield VC returns.
- People want to eat healthy, but it will take time for them to want someone/something to tell them what to do. Maybe one day people will be used to a digital assistant.
- B2B2C was a dead end since supermarkets and apps don’t want to suggest food to their customers from a legal perspective.
Validation
-B2B2C idea: Got to people who worked with retailers. Spoke to 3-4 and it was very helpful. Mostly in Israel. I got to them via tech companies that work with the retailers.
-Product: The best validation was to try it on families. I tried it on myself and on very close friends but learned more when it was more distant circles.
-Best validation was doing it in a way that didn’t waste too much time and money. Tried it more manually on a few families.
Research
-Researched what exists and why it doesn’t work.
-Did a few surveys. Spent around 15,000NIS on surveys.
-Interviewed tens of families in IL and US to understand what they do, what they need, what isn’t solved.
Insights
-GenAI is good for scale but will take time until it does better work than people.
-Define the key thing you want in the solution; crystalize what you’re trying to do.
-Motivation: Ask yourself why are you trying to solve this problem. I realized that I was more interested in solving this problem, even if it resulted in a small company, than starting a $1B company (hypothetically). If you take VC money, it has to be bigger. I didn’t think this through all the way.
-Conviction vs. Risk – Conviction needs to be higher than risk.
Want to contact the founders? Sababa!