Musings on Industrial Software

At Schematic Ventures, I’ve invested in a few industrial software companies tackling automation and met with countless more. Over time I’ve noticed a common pattern: companies are following the strategic playbook of the modern computer software industry. However the industrial software market (or at least automation software) is still in its infancy and faces different market dynamics. A better guidebook would be the success stories from the early days of computer software when hardware dependencies were still a major consideration.

Early Days of Computer Software

1950–1970

In the 1960’s, mainframes were special purpose machines built for specific industries. Customers would purchase both hardware & software together, often customized with the help of consultants, to solve a specific problem. If a new problem arose, a new mainframe was needed. Software was designed either by hardware companies or in direct collaboration with hardware companies.

There was thus no absolute line between systems software and applications programs, and neither was it clear that the former would invariably come from the hardware manufacturer and the latter from in-house or independent efforts. (1)

1970–1990

The industry changed over the next two decades as the IBM antitrust suit, the microcomputer, the multitask operating system and new software distribution methods (cartridges, tapes, disks) emerged. A core value proposition of the personal computer became its open architecture and ability to run third party software.

VISICALC is a shining example of successful third party software of this period. VISICALC was able to port to a wide variety of microcomputers given common architectures and operating systems. However, despite the microcomputer’s new flexibility, one issue kept VISICALC tied to hardware partners: the customer. The install base of microcomputers was still relatively low so first-time computer buyers purchased both hardware and software together as a bundle. This meant VISICALC indirectly depended on the value proposition of hardware partners during the purchase decision and, with this in mind, VISICALC wisely decided to launch exclusively with the promising Apple II.

Guys with MBAs would wander in, curious to see what this “personal computer” hype was all about. They would see VisiCalc running, immediately recognise its usefulness, and demand a copy. And when the sales assistant said that it only ran on the Apple II (then selling for $2,000) they would tell him to add that to the order. (2)

1980–1990

After the install base had grown, the software market shifted into a third, independent phase. For example, a customer, having purchased an Apple II for VISICALC, could then install additional programs without further hardware consideration.

In the OS/2 world, device independence is an inherent feature of a multitasking operating system with a graphics interface… Since software will have to be written to the operating system, it will be up to the systems programmers to maintain [hardware] compatability, not the applications developers. (3)

Which brings us to today’s software market. Computer software customers most likely own compatible computers with sufficient capabilities. Deployment channels are now digital and web architecture is beginning to eat away at the differences between operating systems. Today computer software companies succeed without much thought given to hardware dependencies.

Market Dynamics in Industrial Software

Bringing exciting advancements in machine learning and sensors, more and more start-ups have entered the industrial software market. These companies are breaking apart industrial products into higher margin software and lower margin hardware (industrial computers, robot arms, conveyor, etc). However this approach does not eliminate the hardware dependencies, it simply externalizes them.

Direct parallels can be drawn to the market dynamics of the early days of computer software:

  • Hardware may not exist in the configuration appropriate for the software
  • Customers may not already own hardware and will have to purchase both hardware and software together

Following the modern computer software playbook, industrial software start-ups ignore hardware dependencies until it is too late. The fault lies squarely with the start-up if the correct third party tools and hardware ecosystem necessary for the software do not exist at launch. To avoid this, companies should give hardware and service dependencies real consideration at design stage:

  • What are the hardware dependencies of my software?
  • Does the right hardware exist? Can it be built by the time of my software launch and is someone else willing to build it?
  • Does my customer already own hardware or will they have to purchase hardware with the software?
  • If they have to purchase the hardware, is the hardware a commodity?
  • Who will install the hardware to run my software?

The onus is on the industrial software company, not the customer, to make sure the third party tools, hardware ecosystem and behaviors necessary for using the software exist.

If you would have any questions, please shoot me a note at julian@schematicventures. Thanks for reading!

References

(1) Haigh, Thomas. “Software in the 1960s as Concept, Service, and Product” IEEE Annals of the History of Computing. January–March 2002.

(2) Naughton, John. “Why a Simple Spreadsheet Spread like Wildfire” January 2016.

(3) InfoWorld Jun 15, 1987.