Whizz Systems on Competing in Silicon Valley

Estimated reading time: 11 minutes

Whizz Systems is an EMS provider located in Silicon Valley that has managed to survive and thrive through many of the industry's ups and downs over the past two decades. President Muhammad Irfan discusses the company's assembly and design services, as well as trends he’s seeing from the industry in the Valley. He also details how choosing the right type of customer has led Whizz to see sustained success in one of the most difficult and expensive areas for an assembly shop to survive. 

Barry Matties: Muhammad, please start with an overview of your company.

Muhammad Irfan: We’ve been in business for 20 years, helping customers design and launch new products. We bridge the gap between ideas and shipping their own product. Customers pick and choose what they want to work with us for, starting early on conceptual design to a working prototype, and everything in between from finalizing the design specifications, component selection, doing the hardware design, and developing any software, firmware, or production tests for the product. We also help with building the initial prototypes; debugging the boards or the system; doing all kinds of certifications, environmental, shock and vibration, FCC—whatever certifications that are required—and building the initial ramp-up volumes.

We’re efficient in helping startups walk through as fast as they like and keeping them disciplined and things documented, which is what they often overlook. Sometimes, running too fast is counterproductive. We make sure they run fast but in a productive way; that’s our huge value-add. And especially for startups to smaller companies, we are a great place where they can talk to us in engineering language and get a shippable production-worthy product from here.

Matties: What level of design are you doing—bare board design or the entire assembly?

Irfan: It’s complete hardware design, FPGA design, schematics capture, and signal and power integrity simulation analysis to make sure the circuits are going to work together as one circuit. We also do thermal analysis, identifying all of the problem areas, and designing heat sinks to mitigate any thermal issues in the system. What typically assembly shops refer to as design is CAD work in our industry, so what we do is way beyond the CAD work. We take a concept and do all of the design work, including hardware design, FPGA design, firmware, software, and manufacturing, giving customers the flexibility to engage us at any level.

Matties: What trends and technologies are you seeing out there?

Irfan: There is a lot of miniaturization and connectivity with IoT now taking shape. It was a buzzword in 2011; now, everybody understands that it is deeply impacting everything along with AI. We’ve kept our firmware team and software teams strong, so we offer a total solution. The framework for sensors talking to each other through a hub and some sort of an AI engine, making decisions at the edge, and relevant data being sent to the cloud and back is applicable to pretty much all industries now, and we’re in the middle of it.

Matties: You’re bringing computerization into everything?

Irfan: Yes, we are.

Matties: Many people are making the case that computerization doesn’t fit in North America, that you see it more in a large-scale production environment, and that the investment doesn’t even necessarily outweigh the benefit. What do you say to those people?

Irfan: Some aspect of it is true, but a lot of it is not because AI is going to affect every part of the industry. Even if it’s engineering-oriented, small-volume, early R&D-type work, there is a lot of intelligence gathered through those, and you can adopt better practices and learn how to do better designs. What they’re referring to is there are not large-scale production factories in the U.S. for an investment of that size, but we could scale down that investment to make it more relevant to each industry and the segment that you’re applying it to.

There are a lot of industrial IoT applications that are coming up; they don’t need large volumes, but they still all need certain things, and they cannot invest in these heavy R&D projects. Solutions that are needed for all of those and AI is at the heart of all of it. In our industry, for example, we were the first to make investments in those automated robotic storage systems from Juki and build software intelligence on top of that to learn how the material is utilized, where we run into issues, what type of materials we run into issues with, and what type of customers we run into issues with. That intelligence and know-how upfront will benefit any customer.

Matties: It goes beyond tribal knowledge; you need empirical data to manage your business that way.

Irfan: Right. And let’s say within design iterations, there are certain blocks you use. Most of the time, there is common known knowledge about each type of approach that has pros and cons. If you start to formalize that and you have a way to collect that data, then you can aggregate and have more intelligent decision-making. Instead of somebody looking at it and maybe missing a few use cases, a company that has an organized mechanism for collecting all of that data and drawing a use case will benefit.

Matties: You can avoid a lot of trouble before you find them. Can you share your annual revenues?

Irfan: We do not disclose our revenue. We have a solid foundation, good infrastructure, a strong team, and the best employee retention, and we keep reinvesting back in the business. We’re now looking at our strategy on how to grow to the next level and figuring what the right strategy is, but we don’t want to grow it for the sake of growth; we want to pick the right type of customers. One thing that helped us do well in our business was knowing who we don’t want to be, and not saying yes to all types of business.

Matties: What is the reason for your growth?

Irfan: Finding the right type of customer and sustaining that type of business model.

Matties: When you say the right customer, you mean you’re seeking to grow your revenue with more profit because it’s the right customer. To what degree?

Irfan: We believe that with 20–25% growth, we can maintain the quality of what we already offer the customer in this footprint.

Matties: By optimizing the systems you currently have with AI?

Irfan: We currently have extra capacity that we have already developed because we don’t want to be looking for extra resources when we already have booked the business where it’s a little too late. We already planned the capacity and have already incorporated that into our business model. We have the resources now, and we’re only running one shift for us to grow into the second shift as a natural extension to fill that growth. So, we’re in a good position to handle that growth. We can grow our business and double our revenue in this footprint.

Matties: And the market is in a great position to take advantage of that.

Irfan: Exactly.

Matties: It doesn’t sound like your customer base is market-specific; it’s more based on characteristics.

Irfan: Yes, and the type of engagement. Whenever there is an FPGA involved, that’s a great project because if they want us to develop the FPGA code for them, they usually start us at $500,000 and up. Usually, it ends up being $1–1.5 million for FPGA code. Then, the hardware is around $200,000–300,000, and another a $100,000–200,000 of NPI spend. It adds up to a $1–2-million project.

We use words like AI and IoT to identify the market, but generally, we’re a services company. Between big tech and all of these accounts, we touch all of these technologies. It tends to be large companies with well-defined deliverables from us, and startups who need a card or a system, but they don’t want to touch it; they want to work on their secret sauce (the software), so those are great customers for us. All of the chip companies needing bring-up boards, validation boards, and silicon reference design boards are a great fit.

We have years of experience designing, manufacturing, and managing EvaluationKIT business. We also become a natural extension for them in supporting semiconductor customers to productize the reference designs to a specific application and form factor, which is of tremendous value.

Matties: How many people do you have here?

Irfan: Just over 100 in this facility.

Matties: Is this your only North American facility?

Irfan: Yes. We also have a production facility in Malaysia.

Matties: Is it high-volume?

Irfan: Yes, but again, it’s not super high volume; it’s low- to early-mid-volumes. It’s also good for startups to know they can start here, get to some cost reduction, and go to a Tier 1 CM when they are ready.

Matties: And design a product that will fit into that manufacturing because oftentimes, people design a product that has to be re-engineered.

Irfan: Exactly. We’re very cognizant of that fact.

Matties: How many jobs a week or a month are you running through here?

Irfan: Usually, we have around 50–60, but sometimes up to 100.

Matties: That’s a lot to manage.

Irfan: We are organized in the process, and people do quick-turn, small-volume, and some high-value jobs. For some customers, the boards are usually small, quick-turn types with a lot of exotic stuff. The volume of jobs is high, but the volume is not; however, there’s always stuff going on. There are also customers who need large boards that take a long time. It’s a balance between those two types of customers.

Matties: How do you optimize or manage your production schedule? Is that a manual process still or a hybrid?

Irfan: It’s a hybrid but more tuned towards manual because of on-the-spot demand that shows up before the weekends and one-day turns. And that is why we don’t fill the capacity with cheaper business because if you tie it all down, we cannot be there for those customers when they need it; it’s hard to predict that. By design, you predict that uncertainty will always be required and you keep the capacity available.

Matties: Are those types of customers ongoing, though? Would you have some visibility into their work?

Irfan: Yes, we do, but not planned. There could be a delay in engineering or software development, and all of a sudden, it cumulates at the end towards manufacturing. Something could be happening, but two months later, what is happening this week? That’s common in this business and why a lot of shops cannot cater to that model. It’s hard to cater to that model. We grew with that DNA, so we are good at it.

Matties: Well, we’re hearing more and more of this type of situation, so it’s good to have the capacity and the skillset.

Irfan: And we constantly invest in new equipment to keep that capacity fine-tuned.

Matties: What sort of equipment are you currently investing in?

Irfan: We upgraded some of the inspection systems. We constantly upgrade our lines in terms of additional machines or replacing the older models with the newer versions, such as the SMT line, oven nitrogen capability, auto-press for press-fit connectors, and conformal coating. We don’t do a lot of production volume, but we realize the need for quality, so we invest in equipment even if you’re going to build 10 of something.

Matties: What’s your greatest challenge in this business?

Irfan: Growing the right type of engineering skillset.

Matties: That has to be incredibly difficult to find people in Silicon Valley and pay them a competitive wage and have that skillset.

Irfan: There’s a creative way that we overcame that challenge. We dedicated some space here for a startup ecosystem. Startups are always going to push the envelope and do cutting-edge stuff. Having them here and bringing additional engineers under our umbrella, they come up to speed fast because they’re closely working with those teams. And that indirectly then helps us develop our in-house capability.

Matties: And you have a relationship with these people?

Irfan: Yes.

Matties: Because not all startups work.

Irfan: Correct, or they end up going to other large companies and finding jobs there, so you have referrals. That’s how we’ve indirectly grown our business in the past also.

Matties: But it’s a heated economy.

Irfan: It is expensive to maintain these types of resources. That’s why general assembly shops cannot grow into this market. It is a difficult and expensive undertaking.

Matties: It sounds like you made a lot of smart moves in your business through the downturns and the upturns.

Irfan: In 2001, we were losing money because a lot of our customers disappeared overnight. We were debating about doing a big layoff and question how we could sustain the business. We were a small company, but we still had a sizable payroll, so we sat with all our employees and said, “Either we make a 30% workforce reduction, or we all take 30% pay cut, and the two founders go on zero pay.” Everybody unanimously said let’s take a 30% pay cut and keep everybody. That bought us a long-term foundation that they knew we’re doing it together. The decisions are not made far away where somebody is looking at dollars and cents and making a decision based on that. It fostered that environment that helped us retain a lot of our people in the up-market. They were confident that we were sincere and upfront with them.

Matties: Great. Thank you, Muhammad.

Irfan: Thank you.