Doing The Dirty Work

Back to news

By: Isaac Brown, Landmark Ventures

In recent years, IIoT and I4.0 startups have been founded under the premise that many industrial operators are “data rich and information poor”. The idea was to put an analytics layer on top of these troves of data to boost productivity and make businesses more predictable. It’s true that most industrial companies are information poor – but sadly, precious few should be considered data rich. Sure, their environments *could* be generators of large amounts of data, but most are not capturing it in a meaningful way (and if they are, the data is often just sitting in a box – an actual box – waiting to be queried after an expensive operational issue, not being used predictively).

I’ve watched this play out over and over – industrial analytics vendors get their sales prospects all excited about the wonders of machine learning, but when they get into the factories, oilfields, power plants… they find that there is insufficient data acquisition to deliver on the promised analytics around productivity, maintenance, quality, OEE, scrap, etc. This is why it’s so important for industrial solution providers to “do the dirty work” and help with the data acquisition layer.

The state of data acquisition across the industrial ecosystem hasn’t changed that much in the past 5 years. You would be an optimist if you think it will change that much in the next 5 years. If you’re trying to sell analytics solutions into industrial operations, you had better be prepared to deliver the data acquisition infrastructure, and get your hands dirty. This often means deploying sensors and connectivity systems, whether proprietary or not – it’s plenty useful to be good at integrating the hardware from other vendors, you don’t need to build the full stack yourself, just be able to deliver it. And you need to get good at plugging into existing old-school operational data sources.

Many of these industrial tech vendors have gotten good at hacking industrial control & SCADA networks, which can be valuable data sources. Another good starting point is a data historian, and many legacy environments do indeed feature historians. If you’re not familiar with OSIsoft and the PI System, it’s worth reading about. You might say OSIsoft PI was the first IIoT platform – it’s been around for about 40 years, and an impressive amount of operational data across all the industrial sectors is captured and stored in these systems. I continue to be surprised by industrial tech vendors that A) can’t or don’t want to crack a SCADA network and B) have not built PI (or other historian) connectors.

There’s another thing here that comes up a lot – the need for industrial operators to be “cloud-ready”. Frankly, any industrial operator that is avoiding the transition into the cloud is in trouble, this is part of the reason why the Fortune 500 refreshes itself so frequently – because these companies lack vision and get left behind. The advantages of moving to the cloud are so hilariously obvious that it’s hard to believe companies won’t do it. Somehow, tons of industrial companies (in every sector/geography) refuse to put operational data in the cloud. This is changing, but too slowly. Many industrial tech vendors require operational data to be shared into their cloud platforms, which is a major inhibitor to their ability to win customers. Some smart vendors also offer an on-premise capability, but the majority don’t have the resources – or are too lazy and/or stubborn – to provide an on-premise solution. Some cloud-based vendors are indeed winning deals and achieving growth, but they are the exception to the rule – the industrial tech movement would be progressing more quickly if more vendors offered an on-premise version (or if industrial operators would catch up to the rest of modern enterprises and leverage the myriad benefits of the cloud).

It’s worth pointing out that many of these issues get blamed on one central issue: security. In fact, the industrial cyber security vendors are perhaps the fastest growing segment of the industrial tech market – we work closely with Nozomi Networks, who is achieving very impressive growth. There are admittedly serious risks to connecting industrial environments, but we see that oftentimes these concerns are oversold by executives looking to maintain the status quo (using it as an excuse to not do anything innovative). Moving data aggregation to the cloud puts relatively little at risk when it’s separated from the control networks that operate equipment and facilities. The benefits of the scale far outweigh any investment required in basic data security, and one could even argue that cloud platforms offer BETTER security due to standardization and platform effects. Even for solutions tied directly to SCADA/ICS, there are advanced security solutions to mitigate risks (Nozomi et al.). Industrial tech vendors need to come together to tell this story more completely to their collective customers.

There are a lot of reasons why industrial tech startups are struggling to scale, but certainly one reason is this misalignment with their expectations of digital maturity in their customer segments. It’s unrealistic to think that you’re going to go out and find a sufficient amount of industrial customers who have mature data acquisition programs and are thrilled to put their data in the cloud. In the meantime, solution providers need to do the dirty work and meet customers where they are.

Tags
heavy automation robot arm machine in smart factory industrial,Industry 4.0 concept image.
Top