In Chelmsford, a small city east of London, CloudNC operates a high-spec factory, which specialises in machining precision components for manufacturers in the aerospace, automotive, and oil and gas sectors. 

Before we moved in, it was a decommissioned soft drink production plant (the old Britvic building). We’ve transformed it into a modern facility where we employ some 50 people, all working to meet the requirements of some of the UK’s most prestigious and well-known companies and corporations. 

I’m incredibly proud of what we’ve built. The equipment is top of the range, and in particular we have a quality and inspection department that ensures that our components are accepted by the customer practically every single time. 

But… aren’t we a software company? So how did we end up running a factory…  and why?

‍

‍
Drilling down‍

The short answer to that question is quite simple: we’re trying to build software solutions that reinvent global manufacturing. And, it turns out, that’s an incredibly difficult thing to do: precision machining is a complex, physical environment with millions of possible ways of carrying out every task, in which things get hot, vibrate, wear out and break. 

As a result, creating the software requires more than just coding expertise: it also needs a deep understanding of the physical world and the intricate processes involved in precision machining. That's where the factory in Chelmsford comes in.

By operating our own facility, we have a dedicated space to test, refine and improve our software solutions - live. It serves as a real-world testing ground, allowing us to observe and analyse how our software interacts with the complex machinery, materials, and processes involved in precision machining. For example: if we want to really be sure we’ve found the best way to program a CNC machine to drill a hole on a Haas VF2 until the process is absolutely perfect every time, we can!  

That’s not all though. Having the factory means we also have a team of skilled machinists feeding back. If we’re struggling with solving a physical problem with code, our experts are often able to provide a way forward - precisely because they understand the physical nuances of the problem and what other customers would be looking for.

That feedback loop is critical, as the information they  share with the software team is at a level of detail way above what you’d receive from a typical manufacturer. To make our solutions like CAM Assist perfect, we need really specific scientific information - the kind that all customers don’t have the time to source, and likely wouldn’t share with us anyway as it’s so valuable. 

Sim Factory? 

In a computer game, it’s usually quite easy to build a factory. 

You select a flat-looking site, press the button, and an industrial plant crops up and starts working. Feed it with raw materials, and processed goods start coming out the other end. 

Now it turns out that in real life, it’s… not like that. At all. It’s actually super complex, and very expensive. 

We learnt that the hard way, trying to do it on our own as supposedly smart young techies, without relying on the experience and expertise of manufacturing people who knew what they were doing. 

One of the main mistakes we made though wasn’t to do with layout, or machinery, or staffing. It was to do with everything I’ve just said about why we needed a factory in the first place: to prove the software. 

At first, we thought we needed a factory that effectively doubled as a lab, so we could experiment. And that’s what we built - a gleaming facility stocked with awesome kit. It wasn’t optimised for sales, but given that we needed a testing ground, so what? 

Well it turned out, that became a big problem, very quickly, for two main reasons: 

1: Factories are very expensive and if your facility is optimised for testing software, and not to compete with other factories at the sharp end of the market who are all already providing what customers want, then you won’t meet your sales targets and you will burn a lot of money. Like, lots. 

2: Building a lab helped us iterate our software that worked in a lab environment. But in practice, that didn’t end up being that useful, because software that only works in an ideal fantasy world where every machine works perfectly, exactly as in the manual, doesn’t necessarily work in different circumstances. What we needed to be able to do was create software that can deliver parts to the requirements of the most demanding customers in the most demanding industries - not to our own idea of ‘good enough.’

After some time, we came to our senses and reinvented the factory (under the helm of Mark Duke) to create components for customers on a competitive, commercial basis. In the process, we learned a lot more about creating tools that clients would ultimately want to buy than we ever had when we were more interested in form over function, while also ultimately transforming the facility into a profitable enterprise that today meets the needs of customers varying from bespoke energy companies to some of the world’s largest aerospace manufacturers. 

Ultimately, I fundamentally believe that you cannot build automation software (in any industry) without being able to do the job you are trying to do yourself. We’ve seen start-up after start-up come and go across tech and the process is the same; the founders identify a problem, decide they can fix it with tech, discover the problem is a lot harder than they thought, burn all their money failing to create a solution that works for customers and fits into their workflows, and disappear. 

‍

Having a real factory to iterate our technology probably saved CloudNC from being just another one of those war stories. Now, as we come to launch our first software solution, CAM Assist, into the real world, I know that it’s been battle-tested to oblivion and that it does what it says on the tin. 

Will real users who aren’t based in our Chelmsford factory find things in it to improve? Of course - and I can’t wait to get that feedback and make our solutions even better and more useful. 

But for now, I’m very happy that when we sweep back the curtain and reveal CAM Assist to the world, it’s not just been coded by the best software developers and mathematicians - it’s also been worked over at length by real machinists who know how to programme CNC machines to operate at their best.

Don't just take our word for it. If you want to reduce your programming time by up to 80% - sign up here!