Once upon a time, if you had a fever, a doctor might prescribe leeches to cleanse your humors. If you wanted to travel from Chicago to California, you went on foot or horseback. And if you wanted to manage and track a production process, you attempted to implement a legacy MES.
Thankfully, we discovered penicillin, and we invented the internal combustion engine (to say nothing of electric vehicles!). The MES market, however, is stuck in the days of the Pony Express.
I wanted to open with this comparison not to be cheeky or clever, but because it encapsulates so much of what is wrong with the conversation around MES at the moment. What, really, do leeches, horses, and MES have in common? They’re historical solutions to challenges we still face today. Whether you cite market research, buying behavior, or vibes, it’s clear that problems MES solves are more pressing than ever. The way we think about MES, however, isn’t.
My goal in writing this piece is to argue that we’re thinking about MES in a misguided way, and that this wrongheaded conception of MES is preventing organizations from realizing the promise of digital transformation. Framing the question as “is MES still relevant?” or “is MES dead?” misses the point. Drawing a line in the sand between “integrated MES” and “Siloed Apps” misses the point. The conversation we need to be having is why we let one period in the history of MES dominate our thinking in the present, and whether or not a modern digital strategy can be built on outdated technologies and architectures.
So let’s sit with the opening examples a little longer.
The discovery of penicillin didn’t just lead to the advent of antibiotics. It changed how we understand the transmission and treatment of diseases. When cars were invented, they didn’t just improve upon the horse. They opened up a new era of mobility and freedom across America, and in so doing remade the physical and social landscape of the country. We are at a similar inflection point with MES. The next generation of composable, app-based solutions represents a revolutionary approach to the challenges once tackled by traditional MES (to say nothing of all of the challenges that traditional MES couldn’t address). These new solutions redefine who creates them, accelerate the delivery of value, improve integration across IT and OT environments, and revolutionize how systems are governed, managed, and maintained. Crucially, they enable manufacturers to achieve order-of-magnitude productivity gains that were never, and will never be, possible with traditional architectures. To make myself clear: if your cutting edge MES is a monolith “lifted and shifted” to cloud, you are talking about whether a new saddle makes a horse faster.
Pony express rider. Undated watercolor illustration.
Most folks in manufacturing are familiar with the apocryphal Henry Ford quote, “If I had asked customers what they wanted, they would have said faster horses.” So too, with MES: If you ask traditional MES vendors what they want, they’ll say composable monoliths. Modern consumers want electric cars. Manufacturing leaders want to run digital businesses with the flexibility and agility to transform. We need to stop limiting our thinking on digital transformation to the terms set for us by historical MES. So it is in that spirit that I say: MES isn’t dead, but the horse is.
What should we expect from an MES?
The crux of my argument is that, when we talk about MES, we’re not talking about a technology or any particular system that goes by “MES”. Instead, we’re talking about a solution space.
What we currently think of as MES evolved over the course of the last 30 years, and when we say “traditional MES,” we’re talking about a mix of solution architecture, business model, and solution delivery model. This can be tough to disentangle, but the mixed reputation of MES is only partially due to the rigid, complex nature of the systems themselves. All told, the solution space for MES involves infrastructure, service architecture, system lifecycle updates, and integrations. This directly bears on how manufacturers plan their MES projects. While 70% of manufacturers are considering further deployment of existing MES systems, nearly 45% feel the system is too difficult/expensive to maintain, upgrade, and scale. Indeed, many of the frustrations with MES come down to high total cost of ownership, vendor lock-in, long project timelines, and poor ROI — dead or alive, someone has to pay the bills…
So let’s talk about why traditional MESs are no longer the right solution for organizations interested in digital transformation.
It helps to start with some foundational observations about the nature of manufacturing operations.
- Every manufacturing operation is different. Even if there are only so many archetypes for how things get made, there’s infinite variability from plant to plant, even from line to line in a single plant.
- Operations are dynamic. There are innumerable reasons why a process might change over time, but a discipline rooted in continuous improvement inherently believes things can, and should, continuously change.
These two facts have massive implications for how technology should function in operations environments. Technology at every scale needs to enable an agile way of working. From a system perspective, this means that there’s no such thing as an off-the-shelf solution or module, and systems need to adapt to the changing needs of an operation.
With this in mind, we can look at where manufacturing as an industry is going. When I talk to executives, and when I look at what’s happening in the market, I see the same things over and over:
Flexibility is key — There are two related but different types of flexibility that manufacturers need in their production systems. First, manufacturers need the flexibility to adapt their processes to changes in demand, product mix, supply chain, and other factors in the face of a volatile macro climate. Second, the manufacturers I talk to want to be able to continuously improve their processes, and continuous improvement demands flexibility, both from a system and a cultural perspective.
Traditional MES are not flexible. There’s a reason manufacturers joke (and some aren’t joking…) that they have to change their process to fit their MES. Manufacturers should not have to accept systems that prevent them from adapting their business to the vagaries of the economy, consumption patterns, or politics, and they should not have to accept leaving productivity on the table because a system is too complex or expensive to update. Rigidity is risk.
Long value timetables aren’t acceptable — Depending on the size of your organization and the complexity of your use cases, price tags for MES easily climb into the millions. If the development timescale for an MES is 18-24 months, then your organization is on the hook for a massive amount of money before any value is created. If you need any customization post go-live, open your wallet. Total cost of ownership for MES can often far exceed the initial quote.
When we talk about legacy MES no longer being fit for purpose, we’re talking just as much about business models, delivery mechanisms, and time-to-value as we are about technology. If your MES is central to your digital strategy, are you comfortable making no progress toward that strategy while you wait years (and pay) for that MES to deploy?
“I can’t rely on the vendor or consultants to manage this for me” — Operations professionals create more value for their company when they have adequate tools for optimizing their processes. They create the most value when they own their systems, and can quickly adapt OT and IT to meet the needs of production. Even without moving to a robust citizen development model of solution design and governance, there’s value in reducing reliance on external parties. If you don’t control your solutions, you don’t control your future.
Vendor lock-in is incompatible with an open ecosystem — There is broad consensus that the manufacturers who do best in the digital era will be those that compose solutions for their unique needs using an ecosystem approach. The crux of this approach is that no one vendor can cover every technology, use-case, or contingency a plant might face. And no one vendor should. The ecosystem approach posits that manufacturers will accelerate the value they create by engaging an ever growing number of solutions and new technologies. The extreme end of this argument is open sourcing everything. No one wants to be locked into and limited by a single vendor’s offerings. And yet a rigid, closed MES is constrained by its monolith architecture, severely limiting your ability to engage an ecosystem.
Manufacturing leaders: don’t settle for this status-quo. You can, and should, expect more from your MES. As long as we conflate MES with bloated, expensive solutions implemented in an old school command and control fashion, we’re going to keep missing the promise of digital transformation.
What we Mean by “Composable, App-Based MES”
If manufacturing is moving away from monoliths, what’s the alternative?
If there’s one concept I see consistently, it’s composability. Gartner helped in defining the term, and they have put forth an excellent definition of the composable enterprise that stands on three key core principles (context, architecture, and modularity) that are then expressed through core architectural properties of autonomy (building blocks can be easily changed), discovery (building blocks are identified, monitored, and managed), modularity (building blocks have discrete, bound functions), orchestration (building blocks agree on their method of interaction).
But “composability” around MES is often invoked loosely, usually to mean a modularization of parts that allows discrete feature sets to be recombined into novel compositions. We can look to analysts for a more rigorous definition. Still I would amend the common definition in two ways, with one challenge and one extension. First, I would challenge that, while vague composability sounds great in principle, engineering a platform that allows for composable solutions is difficult in practice, and doing so effectively requires a number of architectural prerequisites. Be wary of vendors who claim that, overnight, their legacy MES are now composable. Take it from Gartner: “In general, it is fair to assume that the longer an MES system has been in the market, the more effort that is needed to fundamentally rearchitect it.” Second, I would extend the definition to say that building a composable MES in practice is as much a culture as it is a property of the technology.
Let’s take this further by thinking through the technological pre-requisites for a composable MES. The first is architectural. Composable MES must be cloud-based, built on containerized microservices (indeed, we could even turn this into a metaphor and say that true composable MES are the microservices equivalent of production systems). The second is a rich and robust set of APIs. Composable solutions can communicate across and beyond the constituent applications to other business and OT systems. The third prerequisite is a common data model that allows all applications to connect and communicate across disparate production and business processes. Ideally this data model should be flexible enough such that manufacturers can adapt it to their unique context. And the final prerequisite is a no-code development environment. Virtualizing a legacy architecture won’t cut it. A solution isn’t composable if the vendor needs to do the composing.
Onto the extension of the common definition: what does a system with these architectural prerequisites allow from a cultural perspective? For one, it allows solutions to be developed in an incremental fashion by citizen developers—frontline engineers who are trained to problem solve. These folks understand their operations better than any vendor ever could.
Composability is about creating efficiency by allowing for the rapid development, extension, and iteration atop common rudiments. With a composable system, you can reuse components of your solutions – entire apps, logic written for a button, a step – because the syntax applies across any environment you would use it in. Keeping the system harmonized works because all applications are built on top of a shared and easily interpreted data structure. Composability is a property of applications, but that’s only possible because the platform they’re built on enables applications to be designed, extended, and recombined at multiple levels of the solution hierarchy. Finally, the platform you use for composing solutions should have governance controls built in, so that citizen development projects can progress rapidly without introducing chaos or risk.
What this means in practice is time-to-value, cost savings, and productivity gains that were never possible with legacy MES. We (at Tulip) see our customers implement MES—not MES Lite—but MES on aggressive time scales. An industrial manufacturer replaced their homegrown MES with Tulip at multiple sites in less than a year and reduced defects by 30% across 60 lines.
A pharmaceutical CDMO overhauled its legacy ERP and MES in 9 months, deploying an MES with Tulip, connected to SAP and composed of 90 apps from weigh and dispense to cleaning, mixing, and drying. A luxury goods manufacturer deployed a human-centric MES and improved productivity 18%. A medical device manufacturer implemented Tulip for fully validated MES, including eDHR, with a greenfield production site in less than 6 months and cut defects by 70%. I could go on.
Ultimately, app-based, composable MES allow you to have the core functionality of an MES while retaining the freedom to extend your solutions into new and emergent areas. One Tulip customer summed it up best: “Given the investment required for a full MES, it wasn’t really something we thought we could ever justify implementing. We realized that we could do it with an app-based approach and solve so many additional problems.”
Manufacturing’s “Faster Horse” Moment
Given how much is at stake, it’s no wonder that MES vendors are scrambling to carve a space for themselves in this new composable world.
And that leaves manufacturing leaders with a decision to make. In an era undergoing a shift no less seismic than the advent of the moving assembly line, manufacturers are being told to line up to see this season’s stable of faster horses. Which begs a few questions: Do you really want to build your digital transformation atop an outdated architecture? Are you prepared to hitch your future to a system that will constrain it? In an era when true transformation is possible, are you willing to settle for a faster horse?
