The Flat-Pack Frontier, Pt. 2

Adventures in Digital Fabrication


In Part 1 of this series, I examined the research, design, and build process for the Zip Tie Lounger. The chair itself is less important than the process; it is a means of “research through making”, an avenue of investigation into the brave new world of digital fabrication.

So, with chair finally in hand, what next?

Traditionally, young designers looking to break into the furniture business have had three options: work in-house for a big company; attempt to license a design to a manufacturer and earn royalties on sales; or start a studio from the ground-up. In-house jobs present a classic conundrum — can’t get a job without experience, and can’t get experience without a job. Getting a design licensed is a similar problem. Big studios like to approach mid-career designers with a proven portfolio and substantial following before making investments in tooling and marketing.

Starting a studio of one’s own is the riskiest of all: failure could cripple the next several years of work, digging an already-deep student loan hole ever deeper. Capital is hard to come by without a track record of success or any meaningful collateral. Space, equipment, software licenses, marketing, taxes, employees. . . small-run manufacturing is a world of thin margins and constant threats to stability.

So a whole generation of designers have turned to the internet in an attempt to gain leverage: sharing work on portfolio sites, trying to get designs featured on blogs, or launching ideas on Kickstarter when traditional funding fails to materialize. But these avenues have their own dangers. Anything on the internet is easy to rip off, obscuring authorship with half-truths and distant manufacturers. Crowdfunding is a whole lot of work, fraught with supply-chain problems and the ability of the internet to amplify disgruntled customers.

Zip Tie Lounger profile.

A new start-up out of London, OpenDesk, was founded last year to lower these barriers to entry. Designers prototype projects, upload photos to collect crowd feedback, and then host free downloads of AutoCAD cut files and assembly instructions for home production. If (like most of us) you lack a CNC machine, you can tap into OpenDesk’s network of fabbers to have the design cut, sanded, and shipped to your door for a fee. Designers get to set their own royalties, local makers get business, and OpenDesk takes a small cut.

In January, I emailed to inquire about becoming one of their designers. I received a healthy dose of encouragement and a list of requirements: a viable built prototype, photographs of said prototype against a clean background, a .dxf cut file for download, and a PDF of assembly instructions. While basic, that list had about 120 hours of work embedded in it, not counting my wrong turns along the way.

First page of assembly instructions for OpenDesk.

I eventually muddled through it all, with some great feedback from the folks at OpenDesk. The Zip Tie Chair is now hosted under a “freemium” model on their site — free to download for non-commercial use or pay to have someone else fabricate it, in which case I make a tiny royalty. You can take a look and download it here.

Easy! Right? I bypassed all the traditional gatekeepers, leapt (somewhat) gracefully over all the barriers to entry, and got my fledgling design to market. But is that market ready for a flat-pack chair put together with zip ties? Which brings me back to the original questions I set out to answer in Part 1 of this series: is digital fabrication really faster, cheaper, greener, or better than making things with traditional means?

Was it faster?

The Zip Tie Lounge Chair took roughly 200 hours to design, build, and ready for distribution. In a commercial sense, that is a tiny amount of time to get a product to market — my chair’s timeline is not even a rounding error in man-hours for an IKEA development team. However, if I use a freelance rate to account for my time, 200 hours at a rate of $50/hour, that is $10,000 worth of effort expended on an entirely speculative venture. That represents an enormous dollar value for an independent designer or startup firm, especially if that is time taken from potentially paying clients (although not true in my particular case).

In terms of actual manufacturing, faster is debatable. A bigger machine, with a vacuum table, could make the cut in probably half the time. If the Fab Lab was larger, sold plywood in-house, and had a conventional wood shop alongside, I could’ve walked in with a thumb drive and walked out with a fully-finished chair, saving a lot of time and driving. Certainly a commercial fabber is better-equipped than I was to execute the file in a reasonable amount of time.

The process, both development and production, is definitely faster than typical timelines in the furniture industry. However, it is still a long and costly path for a sole proprietor just getting his or her practice off the ground.

Page from my sketchbook quantifying the process.

Was it cheaper?

The purchase price is calculated off of an estimated wholesale plywood price, a low ($65) hourly rate for cut time at a commercial fabber, and an optimistic 2 hours ($50) of hand-finishing time, totaling $139 before shipping. Categorically comparable furniture pieces from Target or IKEA are less than $50. I will qualify the pricing comparison by saying it will vary tremendously based on where the customer lives and what material options they pick. However, having access to a CNC router changes the whole equation: the chair suddenly clocks in at about $30 in materials, plus the time of the DIYer.

In the OpenDesk model, royalties are paid to the designer if someone elects to have their chair fabricated by someone in the OpenDesk network. Designers can set a flat fee, or a sliding percentage royalty.

On a $139 purchase price, at a 5% royalty rate, I will clear $6.95 per chair. That works out to 48 chairs just to re-coup my material costs. After that, I’ll have to sell 1,439 chairs to make back my time. If I had my own studio, and was building chairs one at a time, the design, development, and material costs would be built into each semi-custom piece, allowing me to clear a profit as each job went out the door. With mass-manufacturing, the development costs are all front-loaded and the chances of an eventual payoff vanishingly slim.

OpenDesk is working on an app-store model wherein the downloads themselves could be monetized. The potential for profit in the download model is huge — priced low enough, say $5 for the .dxf cut file and assembly instructions (less than a full album on iTunes), with half of that going to the designer, I’d need 4,000 downloads to recoup my time. While 4,000 is a lot, it’s far less of an ask of the consumer since all they have to do is download, not actually follow through and build the chair.

There is a lot of potential for distributed manufacturing to be cheaper, but as of right now, the technology is not widespread enough to radically bring down costs over conventional processes.

Screen shot from the OpenDesk design onboarding process.

Was it greener?

Most of the carbon emissions for production were tied up in transportation: 60 pounds to prototype, or 197 pounds to ship 75 miles in a box truck (an arbitrary radius I drew for testing purposes). Plywood itself, depending on what study you prefer to cite, can be considered carbon-neutral or even carbon-negative, so I assumed no emissions solely from the material. For the machining costs, I came up with 1.21 kWh of electricity for 55 minutes of run time on a ShopBot PRS Alpha hooked up to 220V power supply. 1.21 kWh, even from a coal-fired power plant, only added up to 1.7 pounds of carbon.

Material efficiency is harder to quantify. The cut itself utilized 52% of the half-sheet, which makes for a lot of wasted wood. Furthermore, the remainder of the sheet is very hard to use for something else, as it is weird-shaped fragments. The best way to reclaim the waste would be to wedge some small parts for another project into the cut file. Plywood can’t be burned or composted due to the adhesives, but new, soy-based alternative glues could allow for more sustainable disposal.

Was it better?

The greater question hanging over the concept of distributed manufacturing has less to do with the the manufacturing and far more to do with the distributing. The costs, in both money and carbon, are mostly on the distribution side. Digitally, when a third of Americans still lack access to broadband, it can be tough to even get ahold of the idea of the chair.

On another level, while touted as open-source and bar-lowering, this whole exercise still had significant barriers to entry. I needed access to specialized training to navigate the drafting software and engineering challenges, as well as reliable broadband, photo-editing software, a digital SLR, hand tools, a car, and a couple hundred dollars of disposable income to gamble on an experimental project. To utilize the product of my work, a consumer needs, at the least, an internet connection and a good bit of cash; at the most, they need a car, technical knowledge, access to a CNC machine, and some hand tools.

By that measure, Instructables is much more radically open-source. It can be accessed by phone (no broadband needed), and many of the projects use simple tools and scavenged materials that average folks have access to.

As with any new technology, it will take time for these friction points to resolve themselves. I have no doubt that CNC wood milling will come into the consumer mainstream eventually. Imagine a powerful retail company (Wal-Mart, Target, Amazon, IKEA) putting a set of CNC routers in the corner of each of their (many, many) distribution centers, churning out furniture on-demand and extremely close to point of delivery. Amazon is already building a robust same-day delivery service; conceivably, with access to a network like OpenDesk’s, they could process orders, cut parts, sand, finish, and ship ready-to-assemble chairs nearly overnight.

But, for now anyway, only three people in the world have a Zip Tie Lounge Chair (I gave two earlier prototypes away). I’ll be watching closely over the next few months to see how it gets adopted, tracking downloads, builds, and page views.