Smoke Scout - The pitch

If there was a service that would cost $100 but on average save you $1000, you would jump at the offer. That’s what Smoke Scout is offering wildfire-threatened states across the country. For roughly 10% of the cost of typical damage done by wildfires, Smoke Scout will monitor at-risk areas around the clock and typically report outbreaks within half an hour of the fire’s start. Combining Smoke Scout with fast fire suppression could save states hundreds of millions in damaged property.

Smoke Scout - The fun part

I see two really fun parts about Smoke Scout. First, this is an interesting design challenge. The Smoke Scout UAV needs to have large enough surface area on the wings to collect solar energy to keep aloft day and night.  The UAV will also need to have a payload capacity of about a hundred pounds, almost all of which will be used to carry the fire sensing optics. The UAV also needs to be durable and not require frequent maintenance.

The other thing I find fun about this idea is the image processing and sensor development. The better job you do at image processing, the less optics you need, which saves weight and cost. So, building prototypes and spending a while getting the image processing close to perfect would pay for itself many times over later.

Smoke Scout - The growth potential

If Smoke Scout grew nationwide, it would have have over $2 billion in revenue and $1 billion profit. It it went global that number could go up by another factor of 5 or 10.

Smoke Scout - The monetization

Smoke Scout would provide the UAVs and operation as a service to states. Smoke scout would charge per square mile surveyed per year. I think $10,000 per square mile is about right. If Colorado wanted to always cover 5% of their state, it would cost the state $50 million a year, about $25 million of that would be profit for Smoke Scout. I’m assuming that the operation and maintenance cost would be $5 million. Even at a cost of $50 million a year, Smoke Scout pays for itself 10 times over.

Smoke Scout - The idea

Even the largest wildfires start from a single small source. If the fire can be identified early, within the first hour, then it would be much easier to put it out. The best way to keep an eye on large swaths of land is with an army of unmanned aerial vehicles (UAVs).

The UAVs would be something like the Silent Falcon, except about two or three times larger. With that additional size would come more payload weight for sensors and rechargeable batteries as well as a large wingspan covered completely with solar cells that would allow the UAV to stay airborne for months on end. The solar cells would be large enough to keep the UAV afloat and to charge the batteries, so it could stay up at night. The UAVs would communicate what they were sensing with ground-based control stations that would also give the UAVs their future flight paths.

There would have to be enough UAVs to cover the areas most in danger of wildfires. Each UAV would be able to see about one mile in each direction. If each UAV followed a track as shown below, the loop would be 72 miles, which could be accomplished in an hour and allow the UAV to sense the state of 144 square miles of ground.

So, it would take 900 UAVs to blanket Colorado. Realistically, only small percentage of the state is at high risk for wildfires at any given time. If only 5% is at risk, then only 50 UAVs would be required to keep the state safe. Each UAV would cost about $400,000 or $20 million to cover the entire state.

After a fire is detected, airplanes loaded with water or fire suppressant need to be scrambled to contain the blaze. These plans can fly at over 500 miles per hour, which means that a fire in practically any part of the state could be reached within 20 minutes of takeoff. If firefighting planes are takeoff ready within 10 minutes, then a fire could be suppressed within at most an hour and a half.

Now, there are already satellites that try to detect wildfires. The problem with existing solutions is that the satellites are in lower earth orbit (LEO) and only sweep over a particular area once a day or every other day. Satellites in geosynchronous orbit (GEO) would be able to stay over a single state; however, satellites in GEO would be dozens of miles further away from Earth making sensing the fires nearly impossible.

Smoke Scout - The motivation

In the summer of 2012, wildfires in Colorado burned nearly 4,500 acres and caused roughly half a billion dollars in property damage. That summer was extremely dry in Colorado, and due to the effects of climate change, summers may become even drier in some regions. The key to preventing extreme destruction from wildfires is to stop them early, before they can get out of control.

Easy Print - The pitch

Printers are a huge waste of money, they spend most of their time idle and can cost up to $1 per page if you don’t use them frequently. Stop throwing away money and use Easy Print. Print your document to the cloud and pick it up in a retail store near you.

Easy Print - The fun part

For me, the fun part here is twofold. First, detailing all of the costs in printing and finding clever ways to cut them. For example, Easy Print won’t need very many customer service representatives, which can be a huge cost. Second, making Easy Print truly easy will be a fun challenge. Really understanding how people will use the product and make everything seamless will be the key to its success.

Easy Print - The growth potential

So, above I show that I think each printer will make $2,200 in profit per year. I think in America this could grow large enough to have one printer for every 1,000 people, or around 310,000 machines. That amounts to a profit of around $700 million a year, and a revenue of roughly twice that.

Easy Print - The monetization

The Easy Print printer will look mostly like a big office printer, except a little simpler. The printer will only do the most basic printer functions: print color or black and white, one or two sided, and collate. The print won’t do everything under the sun, and by limiting the functionality, it will be possible to cut costs and increase printer speed (around 30 to 50 pages per minute). The printer will also have a large store of paper and ink, enough to print 2,000 pages. Lastly, the printer will be designed to need minimal maintenance and, when maintenance is required, it will be simple and require hardly any training.

On average, the printer will print around 200 pages a day and need more paper and ink every week. Because the printers will be other stores, the supplies for the printer can be stored in a back room. The maintenance time could be as short as 5 minutes, or $5 paid to the store (0.25 cents per page).

I believe that it will be possible to drive the total cost per page (paper, ink, maintenance, energy) down to 2 cents per page for black and white (4 cents for color). Between the profit for black and white and color, I expect an average profit of 3 cents per page; That makes $6 per printer per day, or about $2,200 a year.

Easy Print - The idea

You need a twenty page paper printed including two color plots, but you don’t have a printer. So, you go to easyprint.com and upload your pdf. The website quickly processes your pdf and gives you pricing options: 20*2 cent per page for light B&W, 20*4 cents per page for B&W or 20*4.6 cents for B&W and color (10 cents a page for color); The website is smart enough to figure out which pages have color and which don’t. You can also customize your order if you only want particular pages to get printed in color (not just all pages with color). After finishing your order, you see a map of easy print printers in your area (in grocery and corner stores). You choose a convenient location and pay for the printing. The website then gives you an order code to use when you reach the printer.

You drive to the store with printer, walk up and enter the order code, wait for the printer to print the pages, and take your pages and leave.

Easy Print - The motivation

Owning a printer just isn’t worth it for most people. If you only need to print something every few weeks (or months), it doesn’t make sense to buy a single purpose machine for the job. If you calculate the average cost of a printer over the 2-4 years of ownership, it comes out to around 10 to 50 cents per page (assuming 1,000 and 200 pages printed, respectively), and that doesn’t even include ink or the energy the printer consumes while idle.

But wait, there is already Google Cloud Print right? Printing from a Fedex Kinkos costs 10 cents per page just for black and white (50 cents for color). Those prices are on par with the using your own printer, so they don’t make much sense either.

The Perfect Tailor - The competitors

http://www.civali.com/
There are two differences between Civali and The Perfect Tailor. First, Civali wants to make custom jeans given the usual two measurements: waist and inseam. I think that doesn’t go far enough and that only a full scan can lead to the perfect pair of jeans. Second, they will operate online and will require about a week to get a pair of jeans to the customer, whereas The Perfect Tailor would aim to get it done in only a day. Lastly, they wouldn’t be able to show what the jeans would look like on the customer, which TPT could potentially do.

The Perfect Tailor - The pitch

No two peoples’ bodies are the same, so why wear the same jeans? The perfect tailor makes jeans that are designed to perfectly fit you. In less than a day, your new jeans will be ready to wear.

The Perfect Tailor - The fun part

Making the booth to get the Kinect or Leap Motion sensor to make the accurate 3d point cloud sounds like fun. Also, making the software to figure out how to make the cuts of jean material given a style and a 3d model will be a challenge.  There are a number of interesting technical hurdles to overcome with this project as well as some more practical ones.

The Perfect Tailor - The growth potential

As stated above, with 100 storefronts TPT would have $12 million in profit. For comparison, The Gap has about 2,500 stores in the US. So, if this grew to The Gap’s size, it would have 300 million in profit.

The Perfect Tailor - The monetization

The Perfect Tailor would have basically the same business model that other clothing stores do. The biggest differences are that TPT wouldn’t use semi-slave labor in developing countries, it would use local labor. Second, it wouldn’t need the complex supply chain. The only inputs are denim material and thread. The only waste is excess denim, which could be minimized using clever cutting patterns. Like some of my other startup ideas, profit here will come from minimizing the amount of human labor. The fabric isn’t nearly as expensive as several hours of wages. According to indeed.com, the average seemstress salary is $25,000 per year, or around $13/hour. In the Bay Area, I expect it would be at least $15 to maybe $20 an hour. Ideally, the labor time could be kept to about two hours per pair of jeans. I would aim to sell the jeans at $90 dollars per pair. That may look like a lot of profit (30 dollars in labor and probably about 10 in materials), but there is also the cost of running the store. Thankfully, the store wouldn’t need a large footprint, because it only needs to have a few booths for scanning and a back room for sewing (or sewing could be done off-site), or about 300 square feet. In silicon valley, I was able to find some locations with roughly that amount of square footage for $25 to $30 per sq.ft. per year, or a rent of around $750 per month for the space.

The last major cost would be that for customer service in the store. One annoyance with staffing a retail store is that most of the employees are idle, or not fully utilized, a large fraction of the time. It’s only during peak hours at night and on the weekends, when most people do their shopping, that you actually get your money’s worth out of the labor.

If some of the sewing machines are in the back of the store, TPT doesn’t have to eat all those costs. The store would be staffed mostly by tailors who also do customer service, when there are customers. When the store is idle, then they go back to sewing. If window shoppers don’t require much interaction, then a sales rate of 25% when the tailor interacts with the customer could be reasonable. I expect that it would take around ten or fifteen minutes of interaction to complete the sale (not walltime just interaction time). With those numbers we need to add on one hour worth of sales cost to each pair of jeans, or 15 dollars.

The retail front rent cost per pair of jeans depends on how many jeans TPT can sell per month. I think weekday sales could be about five or ten pairs. Over the weekends I expect 25 to 50 sales per day. The would sum to 300 to 600 pairs per month. The rent cost is then at most $2.50 per pair.

Putting it all together, I find that the total cost of the jeans is $40 (to make the jeans) plus $18 (to sell them). If each pair is sold for $90, then the profit is about $30 per pair, or $10,000 per storefront per month.

If this eventually grew to have 100 storefronts, TPT would have $12 million in profit per year.

The Perfect Tailor - The idea

Jeans almost never fit right, especially if you are a women. It’s not just that there are many different styles of jeans, it’s that jeans need to be fitted really well to look their best and people are so different that if you gathered a hundred women only a few of them would be able to swap jeans and look just as good as the original owner. People have many different shapes, so tailored jeans are the only way to guarantee a fit.

Imagine walking into a department store, browsing to find the cut and color you want in your jeans, stepping into a booth to get scanned, and coming back a few hours later or the next day to pick up your new pair of custom fitted jeans.

The hardware to do this sort of scanning is already available with Microsoft’s Kinect, and Leap Motions hardware, which should be out sometime this year, is even better. What would be required is to make a booth where one or several 3d scanners can sweep over the customer’s body to collect the 3d point cloud which will later be turned into the model of legs.

Next, the fabric patterns that need to be cut are calculated using the model and the desired style.
These patterns are then cut out of the jean material, by a CNC-like machine, and given to the tailor to sew together into the final product.

Another feature made possible by this setup is letting customers see what they would look like wearing the jeans before they are even made. After the 3d scan is complete, a computer-generated pair of jeans could be placed on the image of the customer, or even shown in real time if there is a screen in the room. I’ve seen hacky examples of this already done with Kinect.

Also, the store would have about a hundred sample pairs of jeans. They wouldn’t fit perfectly, but the would give a good idea of how it would look overall and of the quality of the material.

The Perfect Tailor - The motivation

Lean manufacturing is a radically different way to approach making products. Instead of keeping around large warehouses full of intermediate parts, lean manufacturing suggests making as much of the final product in one place as possible and to only request more supplies as they are needed. It’s almost exactly the opposite of how normal manufacturing has worked for over a century. The end goal is to create the final product that the customer wants with the least waste.  

My goal with The Perfect Tailor is to apply those principles to jeans.

Modern Art Share - The pitch

Tired of your Monet? Like to swap it for a Renoir? With Modern Art Share, you can rent out your art and enjoy the art of others in the meantime. The best part is, if your art is in demand, you can make money while still getting to keep beautiful art in your home. Modern Art Share handles all of the transportation and insurance. You just say what piece you want to rent out and Modern Art Share takes care of the rest.

Modern Art Share - The hard part

The key to MAS is making everything seamless and simple for those renting out the art and those receiving it. The people who would rent out their art are used to high-end services and MAS would have to be very focused on customer service.

Modern Art Share - The growth potential

The amount of revenue MAS would make depends on the amount of art in circulation. Just one art middleman, Sotheby’s, sold over a billion dollars in art last year. There is probably more than $100 billion in art that would be viable for circulation. If MAS got 10% of that, that would amount to $100 million a month or over a billion dollars in revenue per year.

Modern Art Share - The monetization

MAS makes money by charging people rental and transport fees for the art they are renting. The monthly rental price will be roughly 2% of appraisal price, half would go to the owner, the other half to MAS. MAS could also be a middleman for art sales if someone really likes a piece they are renting.

Modern Art Share - The idea

You love your paintings, but you’d also like to change things up from time to time. You set up appointments for Modern Art Share to pick up your painting and to deliver your new piece. An MAS representative shows up, evaluates the condition of your painting, gets your signoff, and takes it away for someone else’s later enjoyment. A few days later another representative will stop by to drop off the art you will be borrowing. You repeat the process of swapping art every few months until you want your originals back.

The process is fairly simple. When everything goes well, MAS picks up a piece of art and ships it to a customer or takes it to a warehouse for temporary storage. Renting art has two part pricing: a mostly fixed cost for delivering and taking back the art, and the rental cost per month of the piece. The monthly cost will be dependent on the value and demand for the piece. The prices could range from $100 to $1,000 a month for various pieces. If you let someone borrow a piece, you get to take a cut of revenue (about 50%).  Art owners could also keep their art in their houses until someone wants to rent it, and only ship it out at that point.

One important aspect would be damages. It’s important that each piece is appraised before any transaction occurs so that the necessary level of care can be determined and any impacts to the piece can be documented at all stages. In particular, the condition of the piece should be precisely determined before it’s shipped because small accidents during transport or in the renter’s home could result in large value losses. It would make sense for an MAS representative to come by to see the piece before it is moved so that it can be appraised and so that high quality pictures can be taken so others can decide if they’d like to rent it.

Modern Art Share - The motiviation

You know you've made it when you can own original pieces of art. Well, at least, you’ll feel like you made it. The downside of owning a nice piece of art is that looking at it gets a little boring after a while. It would be nice to still own the art, but get to see other nice pieces from time to time.

SSLR - The pitch

For too long taking great photos was a chore: use a tripod, a shutter release, a gray card, and only shot landscapes at dawn or dusk. SSLR bring the fun back to photography. Just go out and enjoy taking photographs. Let SSLR’s computer vision wizardry take care of the rest.

SSLR - The fun part

I love computer vision, and this projects oozes it. Doing super resolution imagery, HDR, and clever white-balancing all sounds like a lot of fun.