Tempo Automation: Setting the Pace for Low-Volume, Quick-Turn Assembly
While on the SMTA International show floor, my good friends Michaela Brody and Paul Benke tracked me down and suggested I talk with Jesse Koenig. They told me he had an intriguing story, and sure enough, he did.
Patty Goldman: Hi Jesse, please tell me about your company, Tempo Automation.
Jesse Koenig: We started the company mid-2013, so it’s relatively new. We had the same overarching goal then that we have now, which is to make electronics development much easier and more seamless for our customers, electrical engineers, to let them spend their time on what they're good at, which is designing electronics, instead of on manufacturing logistics. That's always been the goal of our company.
We started by riding the wave of the 3D printer revolution to make a desktop PCB assembly machine that would do solder paste deposition, pick and place, and optical inspection. We spent about a year and a half developing those robots. Many engineers were excited about it, but after a lot of research and talking to customers and testing out that robot with customer jobs, we began to realize there's a much bigger market for taking care of the entire process for the customer and making it more integrated into their workflow.
When we were going to build and sell these robots, a lot of what we were doing was developing the software to run the robots and making it much easier for that process to fit into the customer’s CAD design process. We're still focusing heavily on software but now we're selling our service instead of selling a machine. We're buying what we find to be the best available hardware, the best available machines, and then focusing our engineering resources on software development. So, we’re building our own software for the customer to interface with, and to run our factory.
Goldman: What is your position in the company?
Koenig: I'm a co-founder and VP of Technology. I oversee our facilities, equipment, PCB vendors, and some aspects of software development. I work with our software engineers, doing product management for the software that will interface with the various machines, robots, manufacturing personnel, and PCB vendors, making sure the proper data and algorithms are being manifested in the software.
Goldman: And what does your factory do?
Koenig: Our service is turnkey PCB assembly, so a customer uploads their CAD design to our web application and they get information like DFM feedback and they can process their bill of materials themselves. Our software has access to the inventories of our distributors so we can go through a customer’s BOM, make sure the part numbers are real part numbers, that there are no inconsistencies in their BOM. For example, a quantity that doesn't match the number of reference designators in their BOM—we point out things like that, which would normally be a phone call or email a couple days later.
We display these things in real time, tell them if a part is not in stock at distributors and let them select a different, equivalent one from a list we display or choose to send us the part. Then we give them a quote that shows them why it costs what it does and why it takes as long as it does. So we give them opportunities to make changes, make it faster or less expensive. And then they press the order button and three days later they have the boards in hand, assembled and inspected by X-ray and AOI.
Goldman: Who makes the bare boards?
Koenig: We have a few partner companies with arrangements to get the fastest turn times. We're getting most of our bare boards with a one-day turn time, and we have arrangements with these partners where we have our systems synchronized so we can send them orders without getting quotes first, to make it even faster.
Goldman: Is this largely for prototypes?
Koenig: Yes, largely prototypes. Our quantities are one to 250. Someday in the future I think there may be applications for the kinds of technology and software and automation that we develop to get into higher volume projects, but right now we're just focused on being the absolute best at low volume.
Goldman: You see a need there.
Koenig: We see a huge need there and the difference is greater there. In other words, when you're only making 10 boards, if you have to do a ton of manual set up, if there's a lot of human labor that goes into it at the front of the process to get 10 boards made, that hurts a lot more than if you're making 10,000 or 100,000 boards. That upfront cost is not as bad then. The difference with our software and automation is you get a bigger delta on smaller jobs. That's where we focus now because we can give customers a dramatically better experience than they have at other places.
Goldman: How long have you been doing this part of it? I know you were working on something else before.
Koenig: We set up in our current facility in the first quarter of 2016—that’s when we set up the SMT lines that we now have. We're growing now; we just raised a financing round and so we will be hiring more software engineers. There’s a lot of technology that we want to add to make the customer experience even better, to make turn times even faster, and to make quality even higher. Then we will be adding equipment, too—adding SMT lines. And we’ll be moving into a much larger facility in 2018.
Goldman: Where is your customer base right now? Is it mostly local or across the country?
Koenig: When we started out it was mostly local. There’s still a substantial Bay Area segment, but that’s less than half our business now—it’s spreading across the country.
Goldman: How's that happening? Is it word of mouth?
Koenig: Yes, plus we have salespeople, and marketing. We do trade shows. We don't have a booth at this show, but we do maybe eight shows a year that our sales and marketing people go to.
Goldman: What kind of trade shows?
Koenig: Medical electronics, aerospace electronics, manufacturing and design. We just had a booth at PCB West down in Santa Clara.
Goldman: But it sounds like your real strength, though, is in your software so you can provide this kind of turnaround for your customers and remove frustration.
Koenig: Absolutely. Turnaround is a big part of it, certainly. I think the software and automation goes into two broad categories. One is customer experience, and there are a lot of facets to that. The other is operational efficiency. The one that's more fun to talk about is customer experience, unless you're an investor or an accountant, and then they like to talk about operational efficiency so that we can have better operating margins, because manufacturers oftentimes don't have great margins; the more automation, the better the margins.
The customer experience is everything that we've been talking about, so the fastest turn time, customers love that of course. And spending less of their time on phone calls and emails because they can self-service their jobs. They can get the design right the first time because of the transparency that our software gives them during their design process. They can be designing and using our web application as a tool in that process.
Then automation leads to higher quality, and that leads to a better customer experience of course.
It also gives more education for them, so they understand not to do certain things. They don't have to put a blind or buried via in there or they don't have to make a trace as thin, and it's going to save them a lot of money or save them turn time. They may not have known that. There are a lot of electrical engineers who are smart, who really understand circuits, but they don't know manufacturing implications. They don't realize it's going to make their board take three days longer or cost twice as much when they put in these features. We give them education and transparency on these topics. So all of that goes into customer experience, and it’s really exciting to see how much our customers appreciate it.
Goldman: Excellent. Tell me a little bit about the other side, the operational efficiency side.
Koenig: Sure, absolutely. With that, a lot of it goes pretty much hand in hand. If the customer is getting a better experience by not having to do phone calls and emails over two days, that's also our people saving time not having to do those phone calls and emails. Saving time not having to figure out what they meant about an MPN that wasn't typed in properly or other mismatch in the BOM or parts that are not available. We don't have to spend our time there.
Now that we get the bill of materials that's already been processed by the customer through our software, we know it's all available and we can automatically order those parts. If we get board specs that we know our bare board manufacturer can make, we can automatically send them that order without a lot of back and forth with them. That saves us a lot of time.
We can automatically program our machines because we already know the specs fit with our machines’ capabilities and we can make our machines talk the same language. A lot of times in electronics manufacturing operations each of the machines is sort of an island unto itself. The operator goes in and programs this machine and spends a couple of hours and then goes to the next machine and they're inputting a lot of the same information.
One of the things our software engineers are working on is getting into the digital workflow of these machines, finding how these program files are formatted that go into the machines, and figuring out how much of that we can automatically generate based on the data that's coming into our web application. That leaves a lot less for the operator to do, which makes for a lot less labor going in the first time, and far fewer mistakes. Mistakes cause either quality problems and/or rework for us, whether it's reprogramming machines or physically reworking boards, which costs a lot and kills margins. These things together make for a process that has less labor and less rework, which does great things for margins.
Goldman: Tell me, what is your background?
Koenig: My background is in aerospace engineering. I have an undergraduate degree in physics and a master's degree in aerospace engineering. I worked for about a decade in the aerospace industry. I was doing guidance navigation and control engineering for satellites and airplanes. I would develop simulations that modeled the environment and the vehicles, and develop algorithms that would allow these vehicles to automatically control their pointing throughout complex maneuvers, as well as their position in space. I would work with the software team to turn those algorithms into flight software.
Then I got into program management and business development, also in aerospace. Aerospace is great and I'm still fascinated with it, and we have customers who are aerospace companies, which is fun for me. All these machines and vehicles are fantastic technologically. Also, things like sending people to space and thinking about sending people to Mars is fun work to do.
After a while I started to want a change. The aerospace industry has very, very long timelines and I was starting to want something that's a little more immediate, and more accessible to many people. I was looking around at consumer electronics companies, trying to get to know the consumer electronics industry. I went to an Internet of Things meetup in San Francisco, which is companies presenting and people networking, in the Internet of Things segment of industry. I met my co-founder and our CEO, Jeff McAlvay, there. He already had this idea of making electronics development easier and faster with desktop robots. He had worked on a very early prototype with a couple of people that were helping him to convert a 3D printer into a pick-and-place machine.
The next iteration was going to be a new robot and he needed to bring it to Maker Faire, New York where he had reserved a booth to show it off. We were in San Francisco when we met, where he was working on this thing. When he met me, he said, "Yeah, I'm going to Maker Faire, New York in a couple of weeks. I have to bring this robot so we're working on building it." I said, "Where are you on it? How much of it is built?" He said, "We haven't started actually building it yet." I said, "But you've built other ones. This is just another copy of something you've already built, right?" He said, "No, we've never built it before."
In my mind, I'm thinking this guy is a little bit nuts, but I like it. I like that boldness. I started working with him then. We got a robot out to Maker Faire; it wasn't necessarily the most functional robot I've ever seen, but it attracted a lot of attention there and it got good conversations going, and it also got a conversation going with a person who ended up being our third co-founder. His name is Shashank Samala; he met Jeff at that show in New York. That's the story of how we started.
Goldman: And what you started out to do has since evolved. That's interesting.
Koenig: Absolutely. When we were building these robots we started to take assembly jobs from customers to test out what our customers needed and what capabilities they needed in this robot that we were building. We were taking the jobs as part of the development process for the robot, but then, in talking to so many customers, we decided that even though there is a market—if you could build a beautiful version of that robot that we were building, there is a market for it—there's a lot more R&D that we would have had to put into it to get it to where it needed to be. Then we were looking at the size of that market versus the size of the market for people who don't want to deal with the robot themselves. Some engineers want to have a robot to do their stuff, many others want someone else to take care of the whole thing. Especially because even someone who has that robot, they might need X-ray and they don't have X-ray capability. They need reflow, so what are you going to do about reflow?
We had some answers, but at the end of the day a lot of people want every single step to be done at a very highly professional level. We started to think how we could give them the most frictionless experience and let them spend the most time possible doing design work. To just take data from their designs and magically, from their perspective, produce a board in their hands three days later. That was our vision and so our work is towards that.
Goldman: I take it you've been growing, by leaps and bounds since then?
Koenig: Yes, we have. Right now, we have about 45 people in the company plus a few consultants. I think we'll be growing quickly. It's all relative; you hear some crazy stories in Silicon Valley of people adding hundreds of people in a year as a startup. I don't think it will be that fast, but we will be hiring fairly aggressively over this next couple of years. We are growing. We're a venture-backed company, so the goal is certainly to become very large and to become the gold standard for quick turn, low volume electronics for any application.
Goldman: Do you plan to become a worldwide presence?
Koenig: Yes, eventually worldwide. I think you asked before about prototypes—yes, it is mostly prototypes. It's not 100% because there are times when a customer needs 20 or 50 boards for some lab system they’re setting up or some trackers on some vehicles, or whatever it might be. That's never a high-volume application. In that case, we can build their actual final product as well. So sometimes it's the final product, but most of our business is prototyping.
Goldman: We have a term for that called low-volume/high-mix, with high-mix meaning lots of different jobs.
Koenig: We are 100% low-volume/high-mix. That is for sure.
Goldman: Jesse, thank you, this has been a very interesting and refreshing to talk with you.
Koenig: Sure, thank you. My pleasure.