Product Lifecycle Management is dead; long live product lifecycle management
Product Lifecycle Management (PLM) is one of the cornerstone systems for managing information in complex manufacturing environments. It makes sense: Companies and engineers need a way to manage the information that make up a product’s identity. Similar to how version control systems (VCS) like git are critical to modern software development. Unlike modern version control systems, though, PLM systems were developed by companies that assumed a no-connectivity workflow.
The problem is that PLM has been wrapped up by methodologies and software vendors from decades ago. Processes designed for an era before the internet and software built by companies from the Industrial Revolution surely can’t be the way we do things in the Information Era — can they? Over the decades, companies have come to dominate Product Lifecycle Management, and the management of a product’s lifecycle has been captured by their whims.
The world has changed in the past few decades. The internet changed the way and the speed at which we communicate, build products, and experience the world. However the computers, phones, and wearables that we use to transmit those new experiences continue to be built the same way we built their predecessors.
New problems in transportation, agriculture, and climate have put a focus on physical hardware products. And in 2019, we are at an inflection point with the tools and methods we use to design and deliver hardware. Just as software engineers moved on from waterfall, local data centers, and disconnected teams for designing, developing, testing, releasing, and maintaining software; hardware engineering is on the brink of changing its methods and processes to deliver products that are faster, safer, and more environmentally sustainable than their predecessors.
To realize the potential of new compute and information systems, it’s important to understand the limitations of the way hardware companies use PLM today. PLM systems are used in the engineering or development “phase” of a product’s lifecycle. This phase includes design, prototyping, and initial validation. Once a product has been designed and deemed validated by engineering, they hand responsibility to manufacturing/production. The systems used to manage this phase in the product’s lifecycle are ERP (Enterprise Resource Planning) and MES (Manufacturing Execution Systems). Despite its name, PLM systems are tossed out in a critical part of a product’s lifecycle — the part where the product is built. At best, some of these systems have integrations with others. In many cases, companies spend massive resources moving information across systems manually. In the worst cases, information is siloed in the system where important information is concerned (e.g. PLM systems know about design decisions while only ERP systems know about manufacturing failures).
In the past, information disconnect was the cost of doing business in hardware. And with most of the resources going into “non-recurring engineering”, companies ate this cost by hiring highly-skilled engineers to babysit manufacturing once engineers were “done” developing the product. Today, product delivery expectations are compared to software delivery times. Customers want things fast. Even when customers are not applying that pressure, the number of competitors taking on hardware challenges means that “non-recurring engineering” is now very much recurring. Engineers have to constantly innovate. And with that, the cost of manually ensuring quality makes hardware product development unsustainable.
Designing and engineering a product in only the “development” phase using PLM is no longer valid. Design has broken out into the production and operations phases. Companies are expected to constantly innovate. If they don’t, their competitors are close in their rear view mirrors looking for ways to out-innovate. Whether the competitive threat is technical, efficiency-related, or a new business model, it is too risky for companies to spend years on development and then stop when the product starts getting built. There is no “we’re in development” or “we’re in production”. Companies need to reference key decisions from the “development phase” for regulatory/certification purposes. An engineer may need to make changes based on a recurring failure mode he/she sees while “in production”. The information generated during different phases is now critical across phases. Companies are always in development and in production.
There is no “we’re in development” or “we’re in production”
This has always been the case. Companies just used to eat the cost. When Apple had issues with its iPhone 4 signal reception (Antennagate), it was not able to adjust its manufacturing processes to address the issue. Rather, they spent a lot of money building a a case to customers that they were holding the phone incorrectly. When Apple admitted the issue, they shipped out bumpers to compensate for the design error. More recently, Boeing originally designed its MCAS system with redundant sensors in case of sensor failure. Through years of development, engineers made design decisions that short-circuited their way through manufacturing and test to ultimately deliver faulty airplanes. Boeing will be realizing this mistake over the coming months, if not years.
There are positive examples of companies that are already embracing how new information systems can allow them to change the world. Tesla and SpaceX use data across phases of a product’s lifecycle to continually improves other “phases” of development. For example, Tesla collects massive amounts of data from vehicles on the road to improve its self-driving software and the hardware that supports it. SpaceX has been using data after every flight since the beginning of its Falcon and Dragon product lines. This wasn’t just to improve its ability to land a rocket on land after delivering a payload to orbit, but the data is used to continually improve cost of manufacturing and operating their rocket — the real differentiator from its incumbent competitors that cost orders of magnitude more.
Most of the 21st century’s manufacturing companies haven’t been born yet.
Most of the 21st century’s manufacturing companies haven’t been born yet. The mass manufacture and availability of consumer electronics is just the start to what we are capable of doing when we understand the information coming from the world around us. The movement of information across design, build, and operation will radically improve the speed, quality, and meaningfulness of the products we build in the physical world. For this to happen, hardware companies must rethink the way they create and use information to deliver hardware quickly and safely. By breaking down the silos across a product’s lifecycle, we will get back to the real meaning of product lifecycle management.