Even an “Easy” Kickstarter is Hard

So recently I launched a kickstarter. I picked a product that I liked, and thought, well, that it would be easy. I was very wrong. Making even a simple idea or product good involved a large amount of careful thought, work, and help. I hope that the first kickstarter is the hardest.

I’ll report back if I try this again.

The idea:

Binder Clip (1 of 1)

I had been searching for a good wallet for a while. I had finally settled on a wallet from Bellroy, the $89 dollar “Note Sleeve“. Insane price, I know.

Honestly I loved my wallet. It WAS thin. It was classy. It did serve me well as I traveled overseas for work. However, it did have a tendency to launch my ID and credit cards onto the ground as it aged, but rarely. So when my friend started using just a binder clip as a wallet; well, I fell eighty-nine dollars worth of superior. However, dang his wallet was thin, and damn if he never dropped a card.

So, since this is a post about making a product, I wanted to analyse some flaws in the clip as a wallet.

  • The clip damages the cards.
    • The binder clip actually puts down an large amount of force. It’s easily enough to crack/deform a credit card over time or destroy something as fragile as, say, a coin or plastc.
  • The clip does not protect the cards.
    • This seems like it should go on the previous category, but it deserves it’s own section. One of the properties of a wallet is to keep currency straight, protect business cards, and keep the credit cards from breaking over time like they tend to. A clip leaves them open not only to damage from itself, but anything around it as well.
  • The clip is silly looking.
    • Having a binder clip as a wallet by itself is not going to impress anyone. It’s clever, but it’s cheap. Obviously cheap. There’s something about a nice wallet that communicates intentionality. A sense of having it together. A clip does not do this.
  • The clip has a tendency to fly off the stack when it hangs on something.
    • This is not the first thing you’d consider, but while a clip has really good clamping force normal to its jaws, it’s really week in shear. So it has a tendency to twist off its bundle and then send itself flying off with its own spring force.

Then came a 3d printed prototype for an improved binder clip wallet. My first impression was that it was a very silly object, but after some use I decided that there was actually something there. It worked really well. It solved a lot of the problems to the clip wallet.

The first 3d Printed Prototype
The first 3d Printed Prototype

The first prototype of binder set out to solve a lot of problems.

  • Solved: The clip damages the cards. 
    • The wallet sides add a millimeter or so of extra plastic between the cards and the clip. Through months of use this has proven to be enough to completely stop any of the damage caused by direct contact with the clip.
  • Solved: The clip does not protect the cards.
    • By extending the edges of the plastic of the wallet to the edged of the card it offers full protection for anything inside of the wallet. It also has the added benefit of making it easier to put stuff in the wallet since you no longer have to clamp in the middle under the clip.
  • Not Solved: The clip is silly looking.
    • The 3d printed wallet still looked, well, hackish. This needed work.
  • Solved: The clip has a tendency to fly off the stack when it hangs on something.
    • The ramps on either side of the wallet eliminate the chance of any rotation, which keeps the clip secure. It also lets you slide the bundle of cards into a tight pocket easily.

Let’s make it a product: Design and Prototyping.

About half of all the prototypes made.

I knew that binder was a simple product. There’s nothing complicated to it. So it would most likely be an injection molded part. Which led me to a problem. How the heck am I going to prototype this?

Best Hackerspace

I remembered reading the Guerrilla Guide to CNC Machining a while back. If you’ve never read it I highly recommend that you do so. In my opinion, it’s one of the finest pieces of technical writing ever written. The hackerspace in Louisville, LVL1, had a vacuum degassing chamber and a CNC machine in disrepair. Knowing that I could make it gave me enough confidence to go ahead with designing for injection mold instead of 3d printing. Which meant that I could do vastly more complicated surfaces and smaller features.

So I set out to design and refine the halves of the wallet.

Montages are hard to do with pictures so here's a shitty collage of some of the many many iterations of binder.
Montages are hard to do with pictures so here’s a shitty collage of some of the many many iterations of binder.

I eventually arrived at a shape I really liked. The wallet needed the twin ramps at the side to provide the support for the clip and to make it easy to slip into a pocket, but this made the wallet feel thick. It also meant that the user gripped a wedge to get a handle on the wallet, which is not good because wedges do what wedges do. So my first priority was to find a way to make the wallet visually thinner and improving the grip, while making sure the wallet looked good at the end of the day. I sketched a lot. A whole lot. I finally arrived at the idea that the wallet takes its final shape from. At the edge of the wallet is a triangle that is completely flat and as thick as a credit card. When using binder this gives the wallet a feel of being no thicker than a card, even though it is about four times as thick. From this triangle the ramps come up. It actually works to visually thin the wallet, since the ramps are now triangular and your brain is telling you the triangular section is thin by feel it quite thoroughly tricks you.

So, I set out to get my first prototype and see if it worked.

First I had to repair the CNC machine at the hackerspace. The space had inherited a Zenbot years ago with a nice Wolfgang Engineering spindle. I also went ahead a built an enclosure for the machines. They were to live in the woodshop and it got very dusty in there.

Newly Housed CNC Machines
Newly Housed CNC Machines

Equipped with a working CNC machine I set out to buy materials. I followed the guide fairly well. I managed to score some really good deals on milling bits from ebay sellers. I got great Made In USA endmills from a milspec pcb fab shop out of chicago for nothing, however I had to order from china to get ball end mills. I really searched and ended up with a very useful collection endmills for around $20. 4x 2mm ball, 4×0.8mm ball, 3×0.5mm flat, 3×2.4mm flat, and one bonus bit. They were all two flute since I would be working in plastics and renshape mostly.

What $20 well placed dollars in end mills will get you.
What $20 well placed dollars in end mills will get you.
Difference in quality.
Difference in quality.

Next problem was Renshape. Renshape is great for making molds but it’s super expensive. It’s a fiber in a plastic filler that doesn’t stretch, absorb water, shrink, or expand and machines super easy. You don’t want anything cheaper than Renshape 460. I thought I had found Renshape 460 cheap from carbon3d, but they were actually repackaging the much worse Renshape 440. The surface finish from it is relatively terrible compared to the Renshape 460. Honestly you can get the same finish by milling and then sealing the heck out of MDF. If I were to do this again I would drop the extra cash to get the board of precision board plus from Inventables. It’s just as good as Renshape 460 and a little cheaper.

With CNC, Cutting Bits, and Material in hand I set out to mill my first mold. I shopped around for software to do the CAM paths, but eventually settled on Fusion 360. IT IS PHENOMENAL. IT IS ALSO FREE! Seriously, there is nothing better for the hobbyist than fusion 360 for CAM paths.

Fusion 360 rocking some of my later cam paths.
Fusion 360 rocking some of my later cam paths.

I did have some hiccups at first. Setting offsets, dealing with files, and all sorts of little details. I did manage to avoid plunging the bit into the table or breaking any… this time. Here are the top five things I learned running this cnc machine:

  1. To hold down a block of renshape use ducktape, then double sided tape on the tape, makes cleanup super easy.
  2. Rather than using one super large GCODE file with all your paths and bothering with toolchange codes in LinuxCNC save multiple files from fusion 360, it remembers your position between file loads and moves.
  3. Renshape is magic butter that milling bits love. You can tear through the stuff. Leave the default speeds. Also it wont dull your bits.
  4. Set up three times. Be super mega ultra exact about getting your z-height just right for a good finish. Take your time. Heck take a break and come back and do it one more time. Check it again. Nothing worse than ruining a two hour long machining job because your z-height was off by 0.1mm
  5. If you can’t tram your table (the zenbot has no real tramming ability other than shims and swear words) tram the work piece. Make the first operation a face operation.

So! Now I have a positive for a mold! Time to pour the silicone. For the urethane resin and the silicone mold material I just followed the Guerrilla Guide’s suggestion. Innovative Polymer IE-3075 and Quantum Silicone QM 262. There’s lots of reasons to pick these two and Guide covers it really well.

A successful disaster.

So I put some mold release on the positive and following the guide exactly mixed up the silicone, degassed it, and poured. It was a successful disaster. It worked, but I had a ton of bubbles in the mix. Miraculously, no bubbles on my part.

I had an issue where no matter how long I degassed the silicone (or resin) it would always bubble up. What I eventually figured out with the help from Michal Zalewski of the Guide is that you can actually pull such a vacuum that the resin starts to boil, or worse: one component of the two part mixture starts to boil. The solution? Run the vacuum until you see the bubbles expand and then collapse. You’re done at this point, there are diminishing returns. For silicone take the cup of mix and pour a thin stream from high up to get the rest of the bubbles out. For resin, brush the inside of your mold with resin and then pour low from one spot.

I waited 24 hours and had my first mold! Next came quite an embarrassing round of experimenting to get the parts to come out. Eventually though. I started to have success.

Some success at last!
Some success at last!

 

Here are some things I learned:

  1. There is a real and serious time limit on the resin. You have three minutes. Practice your movements like your a bomb diffusal specialist or a surgeon for best results.
  2. Always use mold release.
  3. Get a really good mix. I bought a badger hobby enamel mixer and that helped a lot.
  4. These parts were poured four months ago and are still bendy. Buy a scale.
    These parts were poured four months ago and are still bendy. Buy a scale.

    BUY A SCALE. BUY. A. GOOD. SCALE!  Especially if you’re mixing small quantities of resin. You need a solid .01g scale. Don’t mess around, the little drug scale you can get at a gas station or a kitchen scale won’t do. I got a great Made in USA used educational scale off ebay for around 25 bucks. I had serious trouble getting parts that were either too brittle or too bendy. The second I got a real scale these problems went away.

  5. You need absolutely zero colorant for a lot of color. I ended up just dipping then end of my paint mixer in my colorant and then mixing up a batch of resin for a serious color. A lot of my first parts were a really dark purple because I was just using way too much colorant.
My friend Matt, with the wallet prototype he's testing.
My friend Matt, with the wallet prototype he’s testing.

Anyway, after about two weeks of waking up in the morning pouring a mold, going in at lunch, removing a part, rinse, and repeat. I started to get good at it. I also had enough parts to start testing the product out. I started to hand them out to friends. They liked them quite a lot, which was great! They even said, un-promted, they’d pay $10 bucks for one, easy. Which was around the price I was hoping to sell them for. It was super heartening.

There were a lot of changes and drafts that came out of this stage. For example I had 1.0mm of clearance between the clip and the wallet edge. This was too much and the wallet could twist almost three degrees off center with this much clearance. I tightened it up to 0.25mm (the magic of this resin process is the accuracy you can get) and that solved that. I also experimented with adding ramps and curves to the edges to make insertion of the clip easier, but none of these really helped at all and the clip was pretty easy to insert in the first place.

Eventually I felt I had a product that was ready for market. So next came the kickstarter.

The Campaign:

The campaign was way harder than I thought it would be. The plan in my mind was, wake up two days in a row, slave over some cool graphics, upload, launch. First off. Making nice graphics is hard. My girlfriend is an industrial designer and a very excellent graphic designer. I spent about six hours on a set of icons after four hours of sketching, and asked her to take a look at them. I could tell that she had some thoughts, but didn’t want to hurt my feelings. She asked if she could help out. I thought about it and said, duh, yes please.

My six+ hours of unprofessional work.
My six+ hours of unprofessional work.

So if you’re ever wondering what the difference is between someone who’s sort of good and someone who is a dang professional. Here is what four hours of my girlfriend’s time came up with:

Four hours of professional work.
Four hours of professional work.

What I’m getting at is that if you have access to someone who’s better than you at something, please use their talents. It will make your project so much better for it. A portfolio of some of her work is here.

Recording the interview section of the video. Nice lighting!
Recording the interview section of the video. Nice lighting!
Recording the audio for the advert part of the video. A nice microphone was essential. Though we didn't need much more than imovie to edit.
Recording the audio for the advert part of the video. A nice microphone was essential. Though we didn’t need much more than imovie to edit.

Next I wrote some copy and a script for a video. This took quite a few drafts. We rented a good camera and went to it. I highly recommend getting good equipment. I thought my girlfriend’s mirrorless camera would do just fine for the video; getting a Canon 5D made a huge difference in how professional the video looked. If I could do one thing better though I would get a good microphone set-up for the interview video. The audio is terrible, but there was no real time to re-shoot.

A small complaint on Kickstarter here. The editor is awful. Truly and terribly awful. We had to redo the graphics three times to get them to work with kickstarter’s terrible editor.

I also spent some very serious time calculating and recalculating the goal for this Kickstarter. I had originally set it a 10k, which would be enough to buy an injection mold tool and do the run. However, after using my trusty gram scale and recalculating the shipping. After researching all the possible hidden costs. I had to bump the number up to 12k. Even after all this, I am set to make at most $1,000 dollars off this Kickstarter if I meet my goal. It’s really a labor of love at this point.

Anyway, we submitted it to Kickstarter for review, got accepted and now it’s time to launch.

 

The Launch:

 

Here we go.
Here we go. Looking. Okay:)

I did a ton of research and decided that Tuesday at 9:00am CST would be the most successful time to launch. This hits the East Coast around lunch and hits everyone else’s lunch too. Weekends are not so good. The biggest thing that helped with the launch is friends and family. I was blown away by their support. There may have been a single manly tear in my eye, but I plead the fifth. In two days we made it to six percent and I’ve been really pleased to see that I’ve started to pick up more and more pledges from people I don’t know.

Campaigns
Campaigns

I’ve also been running small ad campaigns on Adwords, Reddit, and Facebook. Reddit and Facebook have been the most successful, and the most affordable for driving clicks to the Kickstarter. Adwords is doing pretty well too. I’ll have to cancel them soon (tomorrow) because they are eating into my budget fiercely. However, looking at the analytics it has helped drive a few random pledges to the Kickstarter, and from what I’ve read that’s important to Kickstarter’s algorithms for verifying a good Kickstarter.

 

Bonus rendering of the final injection mold geometry. I would love to see this in real life:)
Bonus rendering of the final injection mold geometry. I would love to see this in real life:)

Anyway, now I’m here waiting.  It’s a difficult product to market. In some ways it’s a silly product. In many definite ways it’s the best wallet I’ve ever used, but I had to use it to truly appreciate it. It is hard to communicate a tactile sense for how good something is with words and pictures and requires that the consumer has a good imagination for that sort of input. It is a risky bet. I’m hopeful though and the experience of creating a product, campaign, and launching with ads and everything has been very educational. Next on the list is getting the Kickstarter on some blogs. The ultimate goal being to get the Kickstarter to page one or two of the “what’s popular” section.

It was an enormous amount of work for something that’s set to make so little, but I would 100% do it over again any time. If you’re interested to see the final kickstarter you can check it out here. Either way, thanks for reading about my adventure!

Woo! I did it!
Woo! It’s real!

 

 

Bad Design Thought: The Gas Pump Fallacy

2015-07-05 20.40.47

It goes like this: You walk up to the gas pump, swipe your card, push the regular button, put the fill nozzle into your car, pull the lever, and nothing. Why? You look at the screen, and there scrolling on the screen is a message, “Remove Nozzle Then Select Grade”. You selected grade, then removed nozzle. How could you? You uncultured brute. Which brings me to the bad design trope:

The Gas Pump Fallacy: “Enforcing order to an input when the order needed is arbitrary, or blocking one input for an equivalent input when both are acceptable”

The pump could just as easily have asked you to select grade then remove the nozzle. As far as the software and valves are considered it’s meaningless. The pump needs two pieces of information, “the nozzle is out,” and “the grade is selected”. Since these are binary conditions that have to be met to begin, the order is arbitrary.

“Alright”, you say, “but I have to ask the user to give me the input, shouldn’t the order match the request?”

Well, no. Ask the user to input the information in whatever order you want, but accept it in whatever order it’s given. Another example of this design done right is the automatic registers at, specifically, Walmart. The register needs to know when to start scanning items. It figures this out by waiting for an interaction. The Walmart register lets you scan an item to begin the process, or it lets you push the start button. It offers you both, but accepts and needs one. The Kroger terminals, however, will not start scanning till you push the start button.

Finally, a third example of this being applied in software is the iPhone lock screen. If your fingerprint isn’t recognized the iPhone happily swipes the screen for you to the PIN input. However, even though the touch-id is no longer prompted for, the iPhone will still accept and is looking for a proper fingerprint input. The coder could have easily justified accepting only the pin at that point for security or code reasons, but it would not have improved the interaction.

The gas pump is the best example I could think of. This is because the user will develop a pattern based on the pump they are most used to. It is a common object and the designer has no control over the patterns they learn. A good designer will recognize when the pattern is meaningless and design a pattern agnostic system. Forcing a user’s hand is meaningless, the best interaction is one where the object being interacted with vanishes into the background, and only it’s function is left visible. A good toaster toasts bread. It has a lever and a knob. It does not force you to attach it to your WiFi before beginning the toast cycle.

 

Adventures in 3D Printer Design Part One of Many

I have a Prusa i2 that I built at a build-a-thon years ago. It works great! Well, it did, but time kept on moving forward like the big jerk it is, and everyone built much better printers.  Now it only works okay. I’m saving up for a move across country so I don’t want to spend too much of my hard earned dollars. Luckily though, past me hated money, so I have a lot of spare parts right now.

My plan is simple. Use what I have lying around all the parts from the prusa i2 to make a printer that’s really above average.

My inventory is thus:

How ain't it broke yet. It is.
My trusty old Prusa i2, run hard, put up cold.
Vitamins not for eatin'. 'cept for the brave of course.
Vitamins not for eatin’. ‘cept for the brave of course.

 

With these limitations I set out to work. I went through a couple of revisions

A couple of revisions
A couple of revisions

 

 

 

before I figured out what I’m calling my no. 1 rule to printer design (no. 1 rule pending to change).

1. Design the movey hot plastic bit first. It determines the size of everything else.

So, without any more delay. Here’s where I am with the x carriage.

Very Skinny, Very Fancy
Very Skinny, Very Fancy

I’ve done some neat stuff here that I would like to shamelessly brag about. Feel free to shamelessly flay my ideas to shreds in the comments. If we cant be dicks to each other on the internet then what did we invent it for?

Cool stuff:

  1. Did you know there is a RepRap Vertical X-Standard? I did not, but my roommate mentioned it to me so now this fits it.
  2. The center of mass of the whole thing sits right in the middle of the two horizontal 8mm rods. This should let me move it really fast without tearing all the things to pieces.
  3. There are lots of bearings.
  4. It’s got one of those bird shaped quick change filament things.
  5. It’s really really skinny. Right now with the belt tensioner/bearing retraint assembly, this thing is only 56mm wide.

That’s all I have for now.

My next step is to finish the X-Carriage. To do that I need to design the belt tensioners, then print it out and do test fits of everything. Rinse repeat.

After the X-Carriage comes the Vertical Assembly. This will cover my X and Y. I am trying to stick to the same movement as I had before so I don’t have to buy new belts.