The $100,000 Mistake
Manufacturing the Ecoqube Part 1
Kevin Liang of Aqua Design Innovations shares how he made two $100,000 mistakes in his Kickstarter journey for the Ecoqube desktop aquaponics kit: the first in manufacturing, and the second in fulfillment. This is the story of the first mistake.
For our interview, Kevin invited me to his home where he has a beautiful tea room. We sipped jasmine and tieguanyin tea while he relived his first taste of mass production.
Not knowing any better, when he was first introduced to the factory he works with, he sent them all his design drawings, expecting them to be able to make the product. He wired the factory a big chunk of his savings ($40,000) to cover half the cost. The other half of the payment would be sent to them when the product was delivered. When the first production sample was delivered, it had multiple problems: The filter was leaking, there were bubbles in the clear plastic right up front where the LEDs were, and the clear part of the filter lid where you could see the water flowing through was bent and didn’t fit. Kevin, anxious about his $40,000 and not disappointing his Kickstarter backers, flew out to China immediately.
“We addressed all the parts that had issues one at a time.” The leaking filter was the easiest to fix. The problem was due to thin plastic. The mold had already been cut, but fortunately taking away material from the mold to make the plastic thicker is a modification that can be done without having to make a whole new mold. “This is just a part of the process,” Kevin explains. “If I didn’t speak Chinese, I could have easily made up my mind and said [the issues] were because they’re Chinese and this is the type of work they do. As Americans the gut reaction can be ‘What do you mean it’s leaking? You should be fixing it!’ but the factory is just involving you in the development. The first sample is the rough draft.”
The second problem was the bubble. The Ecoqube light is held up by a clear piece of injection molded plastic, bent at a right angle. The plastic had to shoot up through this 90 degree angle in the mold, and so there were always bubbles right in the front of the product. “If the bubbles were always in the same place, we could pass it off as a design feature. But they were always in different places, sometimes on the right, sometimes on the left. It was horrible!” The factory offered to redo the mold, and have the cost split 50/50 (half paid now, half paid later), but Kevin didn’t have the resources to redo the mold. He went home and thought about it and did some research, and figured out that he could cut the mold so that there was an extra plastic piece sticking out the front of the light holder. This way the bubbles would get pushed into the extra piece, and the piece could be clipped off.
The third problem was the filter lid. The frame of the filter lid is black ABS plastic, with a clear (separately molded) polycarbonate piece in the middle to show the waterfall, inside, which reflected the LEDs. “I really wanted it. It was beautiful.”
But when the polycarbonate cooled after coming out of the mold, it bent 2 degrees and didn’t fit in the frame! Complicating things, the polycarbonate cannot be clamped while hot, or else marks are left on it. After a lot of experimentation, Kevin settled on a process: After the plastic cools for 30 seconds (enough to not leave marks when touched), the workers put it in a rig to clamp it, so that it only bent by 1 degree instead of 2.
One degree was still too malformed for a good fit, so he had the mold altered to add clear pegs on the plastic edge of the lid, and the workers had a jig to drill the corresponding peg holes in the frame. The pegs were glued into the peg holes. The filter still needed to be sealed, but using glue was difficult because it leaves marks on the plastic if you get it anywhere other than the seam. The last remaining piece of the puzzle was using custom 3M tape to secure everything into place.
This was a lot of labour (normally injection molded parts come out basically complete), but the labour was cheaper than making a new mold because of the low volume. The factory probably made 2000 units in total in order to produce 700 units that were satisfactory.
Finally, the Ecoqubes were shipped to San Diego, and Kevin rented a section of a warehouse. “It was 650 square feet that fit perfectly one container, right up to where the garage door closes, because that’s all we could afford.”
When Kevin inspected the Ecoqubes, he realized that a third of them had a problem: The pump was noisy. Kevin says “You might think, oh they shipped us a defective product. No! Jasmine [the owner of the factory] tested every single pump before they put it in. But once you put it in a confined space it starts making sound. So it’s a design problem.” Kevin had Jasmine send new pumps, and spent his Thanksgiving and Christmas of 2014 with his friends disassembling the filter, putting in a new pump, testing it to make sure it wasn’t noisy, and gluing it back together.
Kevin remembers the intense anxiety. “Man, I want to puke when I think about it. Feels like yesterday. But it’s all about the process, the journey”
From injection molding to sheet acrylic
Kevin had finished the design of the Ecoqube C, the next iteration of the Ecoqube, in November 2014. But they were swamped with the fulfillment of the original Ecoqube and didn’t launch the Ecoqube C Kickstarter until April 2015. LaunchBoom helped with the video and landing page and Kevin did everything else. The Kickstarter was an absolute success — they sold 6000 units, and sold an additional 1000 units between the end of the Kickstarter and the beginning of the fulfillment. An order-of-magnitude increase in volume meant flexibility to fix some of the problems from the original.
The acrylic tank had been outsourced in the original Ecoqube. It had been tested before being sold, but Kevin was finding that there were leaks that appeared over time, and the seal would sometimes break in 3 months. Jasmine, the factory owner, told Kevin: The new order is large, there was a problem with the previous supplier, and the machines for making the tanks out of sheet acrylic have become 50% cheaper than they were in 2014. Kevin made the decision to have the factory set up for putting together the tanks out of sheet acrylic. Jasmine hired Xiagong, a plastic engineer, on commission to help them figure out what machines they needed to buy and how to set up production and train the workers.
It turns out that sheet acrylic manufacturing, while cheaper than injection mold tooling at low volumes, comes with its own challenges. The workers have to be trained to glue the parts. Kevin made it clear to the factory that bubbles in the acrylic glue seams were unacceptable, and the factory wasted a lot of plastic while training the workers because of bubbles. “Jasmine is putting in more work and more time, and they had to waste a lot of product to figure it out, but her margins are significantly higher than another factory… Manufacturing moves from country to country, but what stays is high end, high margin goods. Not a me-too product. This is because of the technical skills required.”
“My first hundred thousand dollar mistake was making all the molds.” Kevin says they only use one of the molds out of all the molds he made for the first Ecoqube. When they get enough volume, they will use injection molds again. The minimum order quantity is 5000–10,000 units, so Kevin wants to be moving 10,000 units a month before they go back to the molds. But the acrylic affords them some flexibility with being able to iterate on their product: “Our superpower now is making a product quick that’s functional, and launch from the ground and ship it.”