The Hundred-Year Chair · a chapter

Chaironomics

When I was fifty-two I decided I deserved grown-up furniture and bought a Herman Miller Eames lounge chair and ottoman from Design Within Reach, for what I remember as about three thousand dollars. Design Within Reach listed the certified article at eight thousand seven hundred when I drafted this account. I had also seen reproductions offered for a few hundred. I love mine, including the pleasure of owning the certified article. But that price comparison is not a teardown. It does not establish equal materials, workmanship, or service, or tell me how the price difference is divided among production, licensing, distribution, and margin.

I had investigated the gap between making cost and selling price long before buying the Eames. My second academic paper went after a claim everyone in design repeated and nobody had checked: that eighty percent of a product's manufacturing cost is set by its design. I called the method I used to check it product archaeology, and as far as I know I am still its only practitioner. Chase the eighty-percent number to its source and it dissolves into an informal survey at Ford. So Scott Pearson and I dug.

Consumer Reports had tested the automatic drip coffee makers and found they all brewed an indistinguishable cup, which let us hold function fixed. We tore every machine down to its parts and priced each part under one standard set of input costs. Price and cost turned out to be barely related. The Rowenta cost an estimated $5.92 to make and sold for $40.60. The Toastmaster cost more, $7.61, and sold for less than half as much.

The spread in estimated manufacturing cost had to come from somewhere. With comparable brewing results and the same input prices, we could examine differences in design: the architecture of the assembly, the geometry of the parts, the number of times a hand had to touch the thing on the line. That spread ran to about half the average manufacturing cost in the sample. It showed how much the design could affect cost without explaining every difference in retail price.

Once you know design drives cost, you can watch the industry play it as a game, and the clearest player is IKEA. IKEA starts with the function it wants and the price the chair should sell for, and works backward through margin and freight to a manufacturing target. If IKEA were making the Keel Chair they would probably call it the Köl, fix the price at eighty-nine dollars, and hand the designers a brief: give us the Köl at a manufacturing cost of twenty-three dollars, FOB Linyi. That discipline gives the world painted MDF and flat packs and M6 bolts. I am not sneering. Hitting twenty-three dollars without the chair looking like it cost twenty-three dollars is real skill. I will confess that during this project's naming spree I registered the domain. If IKEA ever wants the Köl it is waiting, and my price is not ten dollars and eighty-eight cents.

My own project had begun at the other end, with unwanted offcuts rather than a retail-price target. But I had already imagined kits and possible products. To price either, I would have to decide how many I meant to make.

Making one chair had let me do things I would have to account for if I offered a hundred. I could sort through the bin until I found a pleasing combination. I could offer up a part, plane away a little wood, and try again. If the stock was not quite what I expected, I could change the design while standing at the bench. That flexibility was valuable to me as a maker. It was not yet a repeatable production plan.

For a hundred chairs, I would need to make more of those decisions before starting. The cherry chair's assembly without fitting and the truss chair's drilling jig were steps toward repeatability. The jig let me transfer positions instead of measuring every stick afresh. But neither answered how long a batch would take or what fraction of the offcuts would yield the parts. I could sort the scraps into repeatable combinations, preserve the freedom to arrange every chair differently, or buy new stock. Buying new stock would simplify the supply problem while leaving my bin of leftovers untouched.

A kit would move some work to the buyer. I would have to choose which cuts and holes to make, label the parts, and put the right hardware in every carton. The buyer would need instructions for something I could explain with a gesture at my bench. A hundred chairs would not simply be my best afternoon repeated a hundred times.

At a thousand, I would also be asking which parts I should make at all. I had been answering versions of that question since long before the chair. I have had a factory of some kind almost continuously since 1999. "Of some kind" does considerable work here. The first was an extra room above the garage at my brother's house in New Hampshire. Most of the later ones had high bays, loading docks, forklifts, and air tools. Since 2003, though, we have made very few of the parts ourselves. Most of our work has been assembly and order fulfillment.

Our first overseas suppliers were in Taiwan. I remember a scooter T-handle that cost six dollars to fabricate in New Hampshire and a dollar thirty-five in Taiwan. At the time I understood most of the difference as labor cost. We later sourced more from mainland China, and more recently from Vietnam. Over those years I came to pay at least as much attention to what a factory made every day as to where it was.

A part might be a zinc casting, an aluminum extrusion, or an injection-molded plastic component. I would not send the same drawing to three factories and expect a useful contest. The material and the process affect the shape I should draw. Sometimes I develop more than one version before deciding which to pursue. I look for a factory whose ordinary work resembles the part I want, rather than asking a willing factory to learn an unfamiliar process on my order.

The Keel Panel's four little brackets gave me that choice. I had printed the prototypes in plastic, then designed versions for injection molding, die casting, and extrusion. Three processes, three part designs. I wanted a bright finish that would sit comfortably beside the fasteners. I also wanted straightforward assembly. I chose an aluminum extrusion: metal formed through a shaped die into a long, constant section, then sliced into short pieces. Each bracket needed two drilled and threaded holes, one through each of its two fastening faces.

I found Mia Chow at Homey Precision in Dongguan through Alibaba. In my experience, the speed of these exchanges is striking. With a drawing and specifications ready, I can often have a substantive conversation with a knowledgeable salesperson within an hour during their working day, and a quotation in less than a day. I have also spent months getting through quotations, meetings, and contracting with domestic suppliers. Those are my experiences, not a stopwatch comparison of two countries. For this bracket, the exchange soon narrowed to the feel of the edges.

Mia's first quotation was fifty-three US cents apiece at four thousand pieces, with a separate tooling charge of $260. Four brackets went into each chair, so that quantity would equip a thousand chairs. I asked whether sandblasting would remove the sharp edges adequately or whether the parts needed vibratory finishing, tumbling with media to soften the cut edges. Nineteen minutes later she replied that she had discussed it with the engineers. They needed to vibrate the parts before sandblasting and anodizing. The revised quotation was fifty-eight cents.

The tooling would cost more than four hundred of the finished parts. At four thousand pieces it added six and a half cents to each bracket, before delivery. For a single experimental chair I could print another version without buying a die. A thousand chairs made the die a modest part of the bill. I had not sold a thousand chairs, but the quote gave me a different way to make their hardware.

I asked for forty samples and for the parts to be packed four to a bag. Mia offered the samples without a separate piece charge, with tooling and sample delivery to be paid. The four-per-bag instruction was as much part of the order as the angle in the drawing. Whoever assembled a chair would need a complete set.

There was still a finishing question. Claude had insisted on masking the threads before anodizing, keeping the finish out of the threaded holes. I had been skeptical, but masking remained an option in the drawing I sent. Mia came back after consulting the engineering team. They did not recommend it for this use. She also pointed out that the quotation had not included masking. I replied, "No need to mask," and sent a revised drawing. The finished threads still had to pass the specified gauge. The change concerned how the factory got there.

Mia later sent photographs of the samples. Three little brackets stood on a surface, one with a thread gauge inserted. Another photograph showed bags of finished parts. I thought they were beautiful. "Great. The parts look nice," I wrote. Mia said she would arrange delivery the next day. The photographs showed a result I wanted to hold in my hand; checking the delivered parts and trying them in a chair would come next.

For the wooden keel I was working with Ivy Yu at EFS Furniture in Vietnam. Their quotation listed an unfinished beech part without inserts and a finished part with inserts, priced at a quantity of a thousand. The finished version was about six dollars. It was a component quote, not the cost of a chair or a promise about what I would pay to receive one at my door.

I ordered eighty-four samples: twelve unfinished without inserts, then twenty-four each in natural, walnut, and ebony finishes with inserts. The sample prices were higher than the production quote. Ivy asked for color references. I told her I cared more about uniformity than the exact shades, and sent examples from the Eames lounge chair. She replied with photographs of their walnut and ebony colors. I thought they would work.

We were also working out the cartons. I wanted two keels per carton, in packaging suitable for the eventual shipment to a customer in the United States. Ivy estimated the weight and obtained a delivery quote of $1,150 for the samples. I thought the carton weight was too high and asked her to check using my estimate. The revised delivery quote was $880. She had made clear that the final amount would depend on the actual packed samples. It was possible to be delighted by the price of a piece of wood and spend much of the conversation on the box around it.

As the samples were being developed, Anne, another member of the EFS team, asked what the part did. She wanted to understand the insert installation and choose the appropriate manufacturing method. Could I explain its function or send a picture showing it in the finished product? I had sent specifications for a stick with holes and metal inserts. She wanted to see the chair.

The keel's quality was still to be established. So was the time required to turn batches of purchased components into finished chairs or useful kits. But I had people working with me on the parts, asking about the edges, the threads, the colors, and the way a fastener would be installed. Much of the workshop I depended on was on the other side of the Pacific. I was grateful for Ivy and Mia's hustle, and for Anne's question.

One, a hundred, and a thousand were different making problems, not milestones I had already passed. The bracket quote showed what tooling could do for one small component. At much larger quantities, a factory could put the entire chair through a single tool.

A formidable answer to the problem of inexpensive seating is the white plastic stacking chair familiar from rented tents and taco stands, injection-molded from polypropylene in a single cycle that runs, by the accounts I can find, around seventy seconds. The industry calls it a monobloc. Out drops a whole chair, warm, ready to stack.

In my course I cost a molded part with a formula short enough for a napkin: mass times material price, plus the machine-and-operator rate divided by parts per hour, plus tooling divided by mold life. For an illustrative estimate, two kilograms of pellets at two dollars a kilogram is four dollars. A hundred-dollar-an-hour machine turning out sixty chairs an hour adds a dollar and sixty-seven cents. A quarter-million-dollar mold over a million pulls adds twenty-five cents. Direct cost, just under six dollars.

Under those assumptions, materials account for about two-thirds of the direct cost. That is not a retail price or an audited factory quote. It does show how a process that produces a whole chair in one cycle can leave little direct labor to allocate to each object.

For someone who needs many inexpensive seats that can be stacked out of the way, these are substantial advantages. My willingness to spend an afternoon making a wooden chair does not answer that person's brief. A low purchase price also tells me nothing by itself about whether an owner will value a chair or keep using it.

A further comparison is a chair I think of as the nice monobloc. Mario Bellini's chair for Heller is one molded piece of fiberglass-reinforced polypropylene, the same process as the parking-lot chair and not quite twice the material. The price I recorded at Design Within Reach was about three hundred dollars a chair. My estimate of the resin cost was about seven dollars. The rest of the selling price cannot simply be assigned to the designer's name or the store. Resin is only one cost, and a buyer is choosing a finished chair, including its proportions, not buying pellets by weight.

Move from those molded chairs to my own builds and labor floods back. A wooden chair is a dozen pieces that have to be cut, shaped, fitted, sanded, and finished. Every one of those verbs is a person's time. I have built enough chairs to feel it in my own hands. When I add up where the hours go, the wood is a minor line. A rough rule of the trade puts a made object at about a third material and two-thirds labor, and slides further toward a quarter material as the work gets more custom. For a chair built one at a time, material can be well under a fifth of the total. The rest is the shaping and the fitting and the finishing, which is time, which is wages.

In the monobloc estimate, material is most of a small direct cost. In my one-off builds, my time is a much larger expense. That does not make material cheap in every chair or make costly timber irrelevant. It means that knowing the species or the weight of the wood is not enough to know what the finished object costs.

Freight matters for the whole chair as well as the keel samples. A chair is an awkward thing to move because it is mostly empty space. You are shipping a volume of air with some sticks around its edge, and past a point freight is charged by the space a thing occupies rather than its weight. A container holds only so many cubic meters, and an assembled chair spends most of its cubic meters holding nothing.

This explains an astonishing amount of the furniture you own. It is why the monobloc stacks: a stack is one floor footprint carrying two dozen chairs. Stacking is a freight strategy wearing the clothes of a convenience feature. It is why so much furniture arrives flat for you to assemble. A knocked-down chair ships as a slab of dense material instead of a box of air, and the company has quietly handed you the assembly labor, the expensive ingredient, to do for free at your kitchen table.

When IKEA reworked one office chair to pack tighter, the count per container went from 2,750 to 6,900. They more than doubled the chairs by chasing out empty space. My decision to remove the truss chair's arms for shipping had a cost in assembly as well as appearance: someone at the other end would have to put them back.

Repair brings another person's time into the costing. I want a future repair to be small and understandable enough to be worth undertaking. That favors access to fasteners and parts that can be made again. It does not favor handwork for its own sake: the more labor a repair demands, the harder it may be to justify. A factory-made chair can be repairable, and a handmade one can be troublesome to repair.

Some choices seemed inexpensive ways to help: a common measuring reference, visible connections, a finish intended to be renewed. Others required more work now to make later replacement possible. Whether that investment pays off depends on someone actually making the repair. I can reduce some obstacles, but I cannot put an obligation to spend the time into the chair.

My notebook has two costing exercises. The factory-build page at the end of August carries component-price ideas and "$199 complete." The later scrap-build page begins "Material $0," adds fasteners, a carton, and labor, and puts "$199" beside "Price." These were proposals, not records of sales or realized profit. The zero assigned the offcuts no new material charge; it did not make the work of turning them into chairs free.

That accounting choice says nothing about the environmental impact of making the original panels, preparing the scraps, or building a chair. The financial ledger could not settle whether wood was a better environmental choice than plastic.

A chapter of The Hundred-Year Chair, drafted and refined in public. The same text lives in the book; the interactive tools and the fully-sourced technical notes live here. home · the book · the technical notes · hundredyearchair.com