Save Cost by Flattening the Tech Stack

What is Architecture and What’s a Tech Stack?

Architecture is an umbrella term for the ways that companies put together software and solutions to solve their business problems. You often hear people talk about a tech stack for the way that different solutions stack or a layer upon one another to get companies closer and closer to their goals, which for many manufacturers, boils down to cutting costs and saving money.

Mike Bowers is Chief Architect at FairCom where he helps companies reduce costs in their manufacturing processes by finding the right technological solutions to solve their biggest problems.

Leah Archibald: You’ve been architecting solutions for almost 35 years now. During that time, have companies high level objectives changed, or are we really still talking about the same things: Lowering cost and improving profit margins?

Mike Bowers: Of course lowering cost is always top of the mind for manufacturing because your cost per unit makes you more competitive or not, and so you have to reduce costs.

The Challenge with Most Manufacturer’s Tech Stack

Mike Bowers: The challenge in the manufacturing world is the technologies that they’re using – I know this will sound bad – are really 1980s technologies: 40 years ago. The technologies used in the IT world are 40 years more advanced than the technologies used in manufacturing for gathering data, collecting data, storing data, delivering data, and integrating data.

All the big buzz about machine learning and Artificial Intelligence requires data, and that data is trapped in the equipment and is trapped in these proprietary protocols. And then you pay people like me to come in and write a custom code to get data out of that machine into your software system. And it’s very expensive to hire me or anyone like me to come in and write that custom code. Once you pay your a $100,000 or $200,000 to integrate a few pieces of equipment, you’re like, “Okay, we’re done. We can’t afford anymore.” And then the CEO goes, but I need more data. I need machine learning. Oh, well, give us another couple $100,000 or give us another million dollars. Because we’ve had to pay someone so much money just to get a proprietary solution.

Leah Archibald: I’m hearing you talk about the real problem of a tech stack. Now, I’ve often heard tech stack as a positive term, because we’re stacking solutions. But you’re talking here about the stack of expenses. Each solution is an incremental expense, and as you get closer and closer to the end goal, there’s a danger that if you’re not integrating your data systems you’re just stacking your bills higher and higher.

Mike Bowers: That’s absolutely right. They call it technical debt, and it’s a very real thing. Factories have technical debt from when they were first built. And a lot of factories were built 40 or 50 years ago. Or even factories built in the last 20 years using these older technologies. They’ve created technical debt that’s stopping them from being modern. And so a new factory today who is looking forward, not with an old school mentality, the competitive edge that those guys get with a new tech stack will blow away everything everybody has out there.

Leah Archibald: Once you have the right data, what can you do with it to save costs over the entire product lifecycle? How can you use the data you’ve already collected and back it into your design process? So you’re not trying to recoup costs later, but you’re actually trying to save costs from the beginning on your new products?

Mike Bowers: When the data is in a real enterprise-quality database, you can discover what worked well in your product and what didn’t. For example, I ran a team in my last job of 80 business intelligence engineers. Those guys went out to the business and said, how can we get better value out of our equipment or our product? How can we make a better product? They would learn all the right questions to ask, take it to the data, ask the questions, and get reports and information back to the owners to say, “Here’s how you can improve your product.”  These data engineers can go and do machine learning on the data. They could do analytics, they could answer any question you have. And if you’re not getting the right data to answer the question, well get the data. You can now afford to collect all the data you want and get it to the right people to answer any question you need.

Leah Archibald: So what is making it more affordable? Is it advances in technology today that are eliminating some of the tech stack bill overrun that we were talking about?

Mike Bowers: Today we can bridge all the technologies in one technology. We built one engine that does all these things. It is a broker and it is a database, and it is an app server, and therefore now it can bridge across protocols because it’s one thing.

Leah Archibald: So what I’m hearing is there’s almost a flattening of the tech stack.

Mike Bowers: Exactly.

Leah Archibald: And then you can bring that data into the other places of your business where you need to have those insights, whether that’s product development or sales and service.

Mike Bowers: That’s exactly right. By flattening it and pulling it into one thing, you open the doors to whole new way of working.