Making the On Demand Business - apps4work/co.a4w GitHub Wiki

On-demand Manufacturing is the enabling technology but it's not the only technology that needs to be present to capture the on-demand garment business.

People are not ordering On Demand garments. They're not ordering them because no one is selling them and no one is selling them because no one is designing them and no one is designing them because no one is producing the software that will enable designers to design them.

And noone is producing the software because it's too hard. It's not hard to produce software or at least there are Legions of programmers quite capable of producing very sophisticated software but it's hard for those programmers to get into the Garment industry. It's much easier for them to go make money providing Slicker algorithms for banks to steal each other people's money more easily, than it is to make software that will make a better Garment. Or indeed most other physical products.

This is not because the product design and Manufacturing industry does not have computers or software. It's because the computers and software in the industry are first generation. They are computers and programs that do what we used to do, but do it in some way better.

We have replaced a man with a pair of scissors with a computer-controlled cutting blade. But they're still an operator they're controlling the cutter. The cutter requires digital input and so we replaced a pattern maker who draws lines on paper with a pattern maker who draws lines on a screen and so the lines can be passed to the cutter machine. On the whole We have replaced and human doing a job with a human using a computer to do the job. We have built computer aided design and computer-aided Manufacturing. Humans added by computers.

We need the second generation implementation of computers and programs for making products.

To get to that we need to look at the whole picture differently. Let us imagine that there will be a computer with a program that will make a product. It will take the idea in the mind of the designer or a consumer of something that they want or need and produce the physical thing.

And then let's ask what points in that process does the computer need help from a human.

And then, as a practical implementation, think about how we want to break up that single computer and that single program into many programs and possibly many computers.

So instead of conceiving of a process which is essentially humans working together with some help from computers, where humans tie together the parts of the process, we see it as an integrated computer process where humans intervene or assist where it seems useful for them to do so.

In this new vision it is obvious that there is a single digital representation of the product being worked on it. It is obvious that programs dip into that representation to find out facts that it needs and put into that representation the facts that it creates... Its work products.

It's obvious that humans do not need to be involved in telling a program where some data is or in moving data from one computer to another. These are things that computer and programs should be able to take care of themselves. Human might be involved in a decision to use a particular program to achieve some desired result and a human might be involved in a decision as to when to perform a particular step, but no human needs to be involved and how a program makes a program work or when it works when there is no human decision required.

In the current process we tolerate a lot of unnecessary human involvement. It is tolerable because it happens relatively infrequently and it is arguably necessary because humans are essentially controlling the process. For example our supplier of our nesting software emphasized how we might run the software over the weekend to save 1% of fabric utilization. The vendor was unconcerned that a cut file needed to be transferred to the cutter operator or that the operator needed to choose which cut file to cut next.

This is more than reasonable if you're about to cut a hundred thousand instances of the product. A 1% saving in fabric is a significant amount of money. And humans typically needs to be involved readying the factory for a run length of 100,000. And the decision that we are all ready to do it with the latest and best and correct cut file is a sensible human decision point.

On the other hand the on-demand business we can't spend the weekend optimizing the cut file because we can only spend five minutes making the entire product and there is very little point in optimizing the last 1% saving a fabric if we delay the production by 2 days. And the decision to cut one instance of that product isn't important enough to involve a human, and isn't really a decision at all, and happens every minute or so and would be an incredible waste of a human.

So a single vision of a computer that's controlling the manufacturing is obvious when the business is the on-demand business.

When you expand the vision to the complete design and Manufacturing the same notion applies. There is necessarily a process of going from an idea of what someone wants through the creative process of making the specification of a physical object that represents the idea and to all of the manufacturing rules that will cause the manufacturing machinery to make the product, and there is also human decision as to whether the product is something they want to be made, but the necessary involvement of humans and the order in which the involvement might happen becomes more variable.

The most obvious variation of order is the reversal of Manufacturing and the retail purchase decision. In the on-demand business we're not going to make it until it's sold. We are not going to speculatively stock shelves or warehouses with product that we hope someone will buy. We will not guess what size fit that someone or what combinations of options they might want or what fabric they might want the clothes made with. And we will not constrain the possible combinations of options, sizes, fabrics and so forth to what we have space to fit in the warehouse or on the Shelf in a retail store or in our working capital allowance. Nor will we markdown the products. I asked an assistant in Target how much a pair of jeans were the answer was "they are still $29.99". I was struck by the "still" because of the implication that a markdown was inevitable.

Less obvious reversals of sequence is the creative decision. A designer makes certain decisions. A designer obviously has in mind something that is the essence of their design. But somewhere prior to manufacturing every decision has to be made, since the retail purchaser it's only choosing among product that has already being made. In the on-demand business any decision is potentially deferrable to the retail customer. The designer may have ideas as to what fabric the garment should be made out of but if the product is going to be on-demand-manufactured it's pretty straightforward to allow a retail customer to choose a different fabric. That creates a whole new Route To Market for fabric manufacturers. Instead of selling into the product development cycle they have the opportunity to sell directly to the end-user. This obviously creates opportunity for simple variation such as the same fabric in a vastly larger collection of alternative colors. This can be handled straightforwardly by On Demand retail software that can present a garment in a color that it's never been made in, as easy as it can present the garment in another color that has never being made in.

But the concept can be extended given the ever-increasing capabilities of representational software. We should expect to be able to show the benefits of fabric that differ in more physical aspects. For example fabrics stretch differently or are semi-transparent or bend differently. We should expect the state-of-the-art representation of a garment to be three-dimensional photo-realistic movie of the potential customer wearing the Garment in a realistic usage scenario. Nobody will buy a wet-suit if they can't see themselves swim through the water in the wet-suit. And the manufacturer of a fancy zipper, I'm sorry Aqua closure, will want to sell his product on the basis of the appearance and function of the aqua closure in real live action.

Therefore it the design decision as to whether to include a better but more expensive Aqua closure or a bigger or smaller one or whatever else might be reasonable variations in the wet suit May in fact become part of the purchase decision rather than a design decision.

I am not saying that every person will Design every garment from scratch. Some garments, like my socks, are going to have no purchaser input into them they're going to continue to be made in their millions of identical socks. And they're still going to cost a few cents to be manufactured and sold for less than a dollar. But the consumer will come to expect to be able to make choices in Greater variety then they can now. To add or remove a pocket, or some other feature that makes it special for them. Software will become capable a varying the design in ways the customer wants but doesn't know how to do. Visual feedback by being able to show the customer in the dress that she wants being the primary determinant of whether she's going to buy it. Unlike the decision to manufacture a hundred thousand copies of the dress, the decision to purchase 1 copy of the dress is going to involve less professional decision makers. Just the purchaser.

It is easy to see the American Consumer come to regard the Garments shown on an On Demand Retail site in the same way that Americans view menus in Restaurants. Menus are merely suggestions from which to compose what you really want. "Hold the mayo" and "Hold the pockets" become equally viewed as entitlements.

Garments are well-positioned in price for the on-demand business. I designed my own house starting with a blank sheet of paper and $50 house design software but I used a professional to draw out the construction plans and he used an engineer to ensure that the structure would stay up. There's no way anyone is going to bet hundreds of thousands of dollars on their own design without professional approval. (But, incidentally, that $50 software, 20 years ago, had far more 'intelligence' built in to it, about designing a house, than the best $100,000 software today for Dress design has intelligence built in to it for designing a dress).

But spending a few tens of dollars to make dress that you designed doesn't seem like a prohibitive risk.

Even Manufacturing dresses On Demand where the manufacturer or the software designer was underwriting the accuracy of the translation of the idea into a physical product is not that big a deal. Consider Starbucks. If you order a fancy coffee and then tell them that the coffee they made wasn't consistent with the long list of instructions you gave, they'll throw the coffee away and make another one.

Shifting design decisions to the retail purchaser creates a whole new opportunity for making money. Today all the decisions that affect the price of the product are made prior to the involvement of the actual consumer of the product. She is left with a take-it-or-leave decision (or may be leave it until it gets marked-down). The decisions are made with a heavy bias towards lowest cost. Other things being equal, the right commercial decision is always "lowest cost". But those decisions are made without knowing whether other things are equal. Is it worth paying a few cents extra for a "better" thread? That's a hard decision because you don't know whether it is "better" enough to justify the extra few cents, but what you know, for certain, is that the extra cents will increase the cost, and the cost is clear and obvious for everyone else to see. It is easier for you to be seen to be successful keeping your costs down. You may never been seen contributing the "better" thread to the product. So the production system is biased to towards cost. (Even in an industry in which the actual production cost is a small fraction of the retail price - like Books and Garments and PCs).

If you move the decision to the purchaser, it's a whole new ball game. "$1 extra for Carbon Fiber reinforced thread", as an actual proposition to the purchaser that pays the $1 and is aware that the threads broke on her last wet-suit, may be a far better business. "$10 extra if you buy the dress 'optimized to make you look cute'" is a better proposition made to the person who would look cute, and who sees herself looking cute, and can decide if it is worth $10 to look cute, than a software proposition to a production designer to make the dresses she designs look "better".

And $10 times the number of people that you can make look cuter, is probably a lot more money than the software licence the designer will pay, because "cuter" is in the eye of the beholder and the $10 is in her wallet, but "better" is debatable. The margin on the "$1 extra for Carbon Fiber reinforced thread" is probably close to 100%, versus the margin you get if you can sell the idea to the designer.

The On Demand Business will not take over the entire garment business. There will always be a Long Run garment business. My socks and my tee shirts will continue be made in Long Runs of millions. Harry Potter novels will be printed in Long Runs, not On Demand. Broadcast TV isn't going away in favor or personalized On Demand content (although look at the effect of Facebook on TV consumption). You-Tube hasn't replaced movies or cable TV -- but look at how they interact. (Take John Oliver's HBO show and its YouTube channel as an example).

The On Demand sector of the garment industry will have similarly interesting interactions with the tradition garment industry. The ability have just exactly what you want will attract the crazies and the extremists, and the trendsetters and the early adopters. If you are invest a million unit run, to populate the shelves for the coming fall collations, you might wanted to pay attention to the On Demand Garment business as the place were trends that might jeopardize your investment or help you tune it, would show first.