At the recent Supernova 2007 conference, they did a session where startups presented, and to mix things up, at the end they told us that one of the companies was fake. Most people clued in, because the presentation had been funny, and had a few obvious business mistakes, but at the same time many commented that it was chosen well, because they would like it to exist. The fake company, ZapMeals claimed it would let you order delivered food from quality at-home chefs and caterers, with a reputation system that helped you choose them by quality. GPS-enabled delivery companies would show you where your meal was as it drove to your home.

The interesting thought here is that the restaurant business is notorious for business failures. This is mostly because it is often done by inexperienced entrepreneurs who don't know how to run a business, and because it's a hard, competitive business.

It did strike me that a system like this could be a good means not only to cheap food, but also to let people find out if they can really make it in the food business. That's because all the would-be chefs already have a kitchen, so no investment is required, though they would presumably need to get the same permits and health inspections any delivered caterer does.

A good reputation system would make this workable. Once a chef got a good and trustable reputation for both good food and on-time preparation, it's not hard to see diners willing to try it.

Even more interesting would be the efficiencies of chefs coordinating their cooking. For example, it's often not much more work to make 20 servings of a dish than it is to make one, especially if you know well in advance that you will do so. The problem is getting those servings to the customers fresh and hot. That's not so hard for dishes with a standing time, but that's only a subset of dishes.

When you cook at home, of course everybody has the same dishes, but we have a different mindset at a restaurant or when ordering out. Not always -- many are content to get one or two Pizzas for the family, or a bucket of chicken, because the price is so low, and asian foods are normally shared. But otherwise we expect to pick our dishes when paying for it. To allow that, the chefs would either have to make the same variety of dishes that a restaurant does -- with the associated costs, or we might consider a fancy delivery system.

Such a fancy system would probably depend on a group of chefs living close together. Each would make a large number of servings of their specialties, and the delivery unit would quickly gather those into the same van, which would then zoom out to the homes, and with computerized brightly printed labels, give each family their custom order. As long as the geographic scale does not get too large, it's pretty workable. Of course, the goal is to not eat up the savings made by cook-at-home chefs with the costs of delivery.

A chef could start out making little margin over the cost of ingredients, to get people to try their food. Then as reputation grew, so would prices. At some point the chef would get to a reputation that would give them the confidence to open their own sit-down restaurant, or they might raise prices so that the delivered catering business becomes a lucrative alternative.

As you know, I often write about self-driving cars. One of the side-results of that would be small, electric quick-delivery carts, ranging in size from a Roomba to a shopping cart. These carts would also work well and solve this problem. Owned by either the family or the chef or a delivery company, they could cause efficient and quick delivery.

It's possible that autonomous delivery carts might even enter the roads before self-driving cars, once people became convinced they would not hit pedestrians. Because they are low mass, even if they did hit cars they would rarely hurt people, causing instead only insurance covered damage. They would roam less popular streets, take up very little room and be very efficient compared to a person driving just to drop off or pick up an item. With just cargo inside, the risks are far lower. The only real risk would be how drivers react to them. I will blog more about these in an independent post.