The potential market for automated micro-farming (backyard farming) is huge, but it will take a long time to to reach its potential. My question is, at what point would AutoMicroFarm (http://automicrofarm.com/) become attractive to investors (both YC and others)? Would 10% weekly growth for a year be key, or something else?
Two and a half years ago, we AutoMicroFarm founders had an interview with you, and you decided not to invest, saying it was difficult to see how AutoMicroFarm would generate the kind of growth startup investors are looking for. However, YC invests with infinite time horizon and is not afraid of risky-looking companies (http://blog.samaltman.com/new-rfs-breakthrough-technologies).
So what would YC or other investors like to see before investing?
10% weekly growth for a year will certainly get investors' attention.
However, for a company like this, I think the most important consideration is how you plan to build a monopoly. There are lots of companies proposing to build automated food growing units for families with very similar plans; what stops this from being a race to zero-margin pricing? Why will a customer buy your product in 5 years and not one of the many clones?
People refer to this in lots of different ways--Warren Buffet as the relatively benign sounding "moat" and Peter Thiel directly calls it a "monopoly". Whatever you want to call it, how do you plan to do it?
When will you have the first unit in customers' back yards?
How are you going to price it? If it pays for itself in 5 to 6 years, have you thought about going out with a leasing program? I think people almost never do enough on the financial innovation side.
To build a monopoly, the plan is to make the product open-source. On the data side, having data about all the different environmental conditions and how they affect plants and fish will also help build a monopoly.
The first unit will go in a customer's back yard in a few months. In my new-construction neighborhood of ~400 houses, I plan to find 10-30 customers; I've started talking and have several interested.
The 5 to 6 year payback period is conservative: it assuming the need to build a greenhouse to house the AutoMicroFarm for year-round production. At scale and without needing a greenhouse, the payback period drops down to 3-4 months (when compared to similar organic food).
Edit: as far as a lease program, the USDA has a family farm lending program with a low rate that we could tap into on behalf of customers.
Could you say more about how making the product open-source will give you a monopoly? Will the data be open-source too?
If the product is open-source, how will monetization work?
If you can get 30 houses in your neighborhood as happy customers, that would be a great start. You should give them super customer service--ie, go around to their units and make sure everything is working perfectly. The word-of-mouth recommendations from your first customers is so important.
What do the unit economics look like? How much are you charging for the current units, how much do people need to spend a year to operate them, and how much do they save on groceries?
My thinking is that the hardware and software will be open source, but not the data. Also, the OS license would allow modifications for the customer, but not for resale. Since the biggest part of the cost will be the "dumb" hardware (the containers and structure make up ~70% of the cost), it should be straightforward to sell the product at a profit.
Another point of monopoly would be to keep up a pace of innovation that is faster than the competitors.
I am offering the first customers a sale price of just the material costs. Also, they would get 24/7 support, and the CEO's (i.e. my) cell phone number.
Currently, similar products cost $1500+. My costs are $1000 (in quantities of one), and that price will come down 10-fold once I can manufacture 10k units per year.
At the current costs ($1500), the payback time is 1-2 years. Once purchased, the cost to operate is negligible (tens of dollars yearly). So they save $1-2k in groceries annually, depending on the length of the growing season.
I think you will have a hard time claiming that the project is open source except the data which you keep and monetize.
I strongly suggest considering other models. Why do you want to "open-source" this at all? Why not just make it hardware and software you sell, and recommendations about how to grow based off of your large data set, but very hackable?
That is a distinct possibility with plant "inserts" for the vegetable beds, one of the planned innovations I've had in the back of my mind. AutoMicroFarm could sell just the empty media inserts for a reasonable price, and also offer pre-planted, ready-to-fruit plants with inserts for a much higher price.
That way, those who are willing to wait weeks/months for their plants to grow from seeds/seedlings can save money, and those who want it now can pay the extra cost.
As for the "moat", I think the product with the best automation will win at least a defensible niche within the market. I would encourage you to start upmarket and not worry about price so much right now. Get really good at making the thing run itself; you can cut the price later. (But, I'm someone with a lot more money than time, so that colors my view here.)
I haven't polled my friends on this, but Silicon Valley (where I live) is full of such people. And a lot of us have yards, the area being resolutely suburban, and of course the weather is great (assuming this thing doesn't need too much water).
I don't think a lot of people here want to spend time gardening, but lots of us love fresh vegetables. That's an opportunity, it seems to me.
ph0rque, in addition to all the good ideas here, think about a solution to the problem that all gardens and farms have - that what you want is not always at the right harvesting time. Ie: if I want to eat a carrot, I want it now, not in 2 months. How can you solve that problem with your system? Bigger volume? Exchange with neighbours with their own AMF? etc...
Also I'm definatly with @sama on the Nespresso model. Don't even bother selling seeds because some will not even germinate. Sell the pods (biodegradable of course) that "plug" into a prepared receptacle.
Thanks for the insight, jonnycowboy. I'll think about the problem you pose. It can be somewhat mitigated by staggered planting (in the case of carrots) and continuously-fruit-bearing plants (e.g. some tomatoes), but there must be a better, more elegant solution.
Thinking about it more, I could introduce the subscription model by knowing how far along each insert is in the harvest cycle, and offer to ship new inserts right as the existing insert harvest is winding down.
I wonder... you could advertise something like this as a health product. Let's say you have Celiac and you need to eat gluten free. Using this AutoMicroFarm, you can grow exactly what your customer needs (some lettuce strains may be hard on the stomach, some much better etc), so using your combined data from all units, customers, you can make sure that a certain AutoMicroFarm unit is completely fine tuned to that customers health needs.
Celiac is a silly example, but hopefully you know what I mean! I'd definitely be looking into this because it's realistic, and could help people a lot. Another wee example is that a lot of mass-produced foods are probably genetically modified and have odd chemicals which of course a lot of the population have adverse reactions to -- using an AutoMicroFarm unit, you can make sure that the food is completely organic, no chemicals, you have a record of everything that's gone on so you know you're safe to eat it.
Though, I'm probably talking crap!
p.s. I'm into hydroponics, etc. Done a few things like this unit with Tessel and stuff so I'm excited to see where you end up.
Thanks, I'll definitely keep that in mind! I have a friend who eats only organic for health reasons and to avoid ingesting the odd chemicals you mention, and he's definitely interested in an AutoMicroFarm.
There may be overlap with the market for food spectrometers, https://www.kickstarter.com/projects/903107259/scio-your-six... . You are providing food supply chain transparency at the point of production rather than consumption. This is related to software security efforts to build reproducible binaries from open source, to guarantee production integrity.
Edit: here is Nussbaum's list, pulled from the Wikipedia article:
Senses, Imagination, and Thought
Affiliation [social interaction]
Control over one's environment (political and material)
Even that is very relative. What quality of housing? Of food? There are lifesaving medical procedures that even the richest struggle to afford, if the condition is rare enough (IIRC someone on HN posted about paying for genetic testing for him and his wife, $8k each, to figure out what their son's condition was - is that something you're "poor" if you can't afford?). It's even clearer with education - is basic schooling enough? Are you "poor" if you can't afford to do a PhD?
We should definitely start with the necessities of life. Food and clean water, hygiene (it sounds dumb, but being able to bathe is a luxury in some places), dental care is probably more important than medical just from the standpoint that more need it...
Shelter would definitely be on the list. Energy probably.
Transportation probably not.
How to define the minimums... not entirely sure. Simplistic formula are liable to be downright unpleasant (daily calories) if distorted in situations like the "lowest bidder provides". Anything more than simplistic will soon turn into a designed-by-committee mess.
Clean water is the only easy one to define. Measured in gallons or liters, no contamination from toxins, no contamination that makes it unpalatable to a reasonable person.
Even energy can take many forms... electricity, cooking fuel, heating. If you live on the equator, no need for heating, but in northern climes it is a matter of survival. And the exchange rates between various forms of energy aren't constant, not even within a single locale.
> Eventually, we'd come to learn how wrong we were, how hard the road ahead of us would be, and that selling the company would mean the end of what we'd worked so hard to build. But that night was for celebrating.
So ominous... any hint as to what happened before the next blog post installment?
> So ominous... any hint as to what happened before the next blog post installment?
Looks like they started a new company called TaskTorch. Considering this acquisition happened in 2010 and TaskTorch hasn't really launched yet I'm going to assume they stuck around big company for a few years, hated what they were doing to the software they loved to build and eventually quit after taking their sticking-around bonus.
Good way to announce your new company if I do say so :)