How to successfully introduce new capabilities into society and your supply chain

There’s a saying in supply chain management– you’re only as efficient as the weakest link in the chain, and you can only move as fast as the slowest link. For instance, it’s great that the goods-to-person order-fulfillment system you installed can fill an order within 15 minutes of receipt; however, what have you really accomplished if replenishment processes prevent pick modules from being adequately stocked with inventory, it takes hours for manual pack stations to complete individual orders or shipping docks become quickly overwhelmed?

The same thing happens outside of the four walls of industrial facilities. We’ve likely all seen video footage of traffic in developing countries – cars, trucks, motorcycles, bicycles, pedestrians and livestock, all sharing the same road; even that high-performance sports car can only move as fast as the flock of sheep crossing the road or cow which chose that moment to lie down and block traffic will allow.

Something similar is happening now in supply chain given the constant introduction of new technologies and capabilities, both in our everyday lives as consumers and in the industrial settings in which we as supply chain professionals work. The rate of change and introduction of new tools promising to make our lives better, or improve our operations, is dizzying. Some succeed in finding a foothold, and yet others fade— in part due to the difficulty integrating said tools into existing systems, an important factor considering that few have the luxury to rip out and replace existing systems and processes without tremendous disruption.

In this article, we’ll reflect on how technologies and capabilities are introduced into society and the supply chain; why some are able to readily fit into existing systems while others are not; and suggest a framework to design improvements more likely to be successful— improvements that will strengthen the weakest link and drive efficiency and productivity.

Homogenous and heterogenous systems

New technologies are introduced into the systems in which we all operate through two different means: homogenous and heterogenous.

Homogeneous systems have similar performance capabilities, and any differences in capability that exist have little impact on each other. In a homogeneous system, the technologies, or actors, have the same performance capabilities and are governed by the same rules of the road. In other words, they’re all operating at the same speed and with the same capabilities to react to events in real time.

Now imagine a highway system built specifically for autonomous vehicles— all vehicles on the road would notice a red light at the same time and come to a stop; they’d all start up at the same time when the light turned green and could communicate with one another in real time. If we had the luxury to create a homogenous highway system, we could roll out autonomous vehicles fairly quickly.

The reality is that transportation systems – and most systems – are heterogenous. Not unlike transportation systems, most complex and large-scale systems consist of subsystems that vary in performance. Like the developing world example, the roads accommodate all sorts of transportation modes simultaneously, everything from sports cars to trucks to rickshaws to cows, with various levels of performance— flexibility, anticipation of intent and cooperative nature of conflict-avoidance are the key performance differentiators. When humans drive cars, they make adjustments based on real-time conditions and cooperate to ensure the overall safety of the operation, an example being one driver waving to another as a signal to proceed. In a fully- autonomous car, such cooperation from heterogeneous actors is more difficult. Similarly, if a cow or buffalo enters the road in India, cooperation is significantly reduced or stops altogether, since intent cannot be shared effectively.

Even as the technology for autonomous vehicles improves, it is likely to be years, if not decades, before a significant percentage of cars on the road are autonomous. There will be early adopters and late adopters, with the majority of the population somewhere in the middle. Such an adoption cycle means that heterogenous systems are a reality of life.

Another example of a heterogeneous system is air transportation, currently consisting of jets, turbo prop aircraft, gliders, helicopters, drones, balloons and space-launch rockets. In the future, there will be urban air mobility vehicles, drones of all sizes, internet-serving balloons and vehicles capable of supersonic and hypersonic speeds. The performance characteristics of these vehicles differ in terms of speed profiles, turn radius, and climb and descent rate; yet all will need to be accommodated.

We can see a similar example in the supply chain as warehouses implement robots to work alongside associates in operational functions such as sorting, picking and packing. In many instances, robots increase the productivity of these tasks… when all goes as planned. However, when things don’t go according to plan, or if a robot stops for one reason or another, humans are forced to intervene. While there is much discussion about flexible automation, people remain the most flexible – and creative – part of the system.

The world is full of heterogeneous systems that include various actors with different performance capabilities, all of which may impact, or interfere with, others’ performance. As systems trend toward digitization, artificial intelligence and increased autonomy, it will be important to accommodate and integrate them along with existing systems.

The challenge – one that may determine a new technology’s success or failure – is how to safely accommodate heterogenous actors with a wide range of performance capabilities into an existing system to maximize efficiency.

Integrate or segregate?

There are two approaches to meet this challenge:  One is to integrate a new technology or capability into an existing system; the other, if possible, is to segregate it.

Most transportation systems are examples of segregation. Interstate highways are designed for vehicles with higher performance levels; that’s why we don’t allow bicycles on Interstate 80. Meanwhile, many local roads have designated bike lanes to improve cyclist safety.

Similarly, in order to maintain the overall efficiency and safety of air transportation operations, gliders and drones are prohibited from operating near airports or at higher altitudes where commercial aircraft fly.

Of course, everyone wants flexibility such as allowing cyclists on any road, but doing so would create significant efficiency and safety challenges. In order to strike a balance between efficiency, throughput, and safety, transportation systems follow a common mantra:  Flexibility is afforded where possible, and structure is designed in where necessary.

Another way to describe the design principle for heterogeneous systems is: Integration where possible and segregation where necessary.

A good example of this principle is the introduction of robots in warehouses. In piece-picking operations, autonomous mobile robots with significant onboard safety features work in coordination with order-selectors on the floor; meanwhile, larger, robotic palletizers capable of inflicting serious harm or death to a worker are segregated away from manual-palletizing operations.

An age-old problem

Many of the tools and technologies available today are new to supply chain managers, such as Artificial Intelligence, Machine Learning, 3D Printing and Additive Manufacturing and drones. But the challenges associated with how to introduce these new technologies into existing systems have always been with us. In fact, incredible technological advancements have been consistently introduced into society over the last century, if not longer. It’s worth looking at some of those, as well as the lessons learned on how various technologies were successfully introduced.

Looking back, we find there are three categories of introductions.

In the first, an introduction offers improved productivity or performance, typically with minimal, if any, negative impact on the existing technology or uses. The first generation of personal computers, for instance, improved the speed of calculations and other data manipulation and gave us the ability to digitally store documents.

In the second category, new technology can co-exist with existing technologies and uses— the electric car is an example of this category. Based on most auto manufacturer’s current plans, the gasoline-powered car will likely become obsolete; but until then, electric and gas-powered cars will continue to co-exist for quite some time without interfering with each other.

In the third category, a new technology or improvement replaces what’s currently available over a period of time. In this category, technologies are introduced to improve convenience, through better performance or at a reduced cost, or for safety, such as seat belts, airbags and retrofits that quickly become mandatory. Technologies that improve convenience such as Netflix are driven by customer needs, either stated or implied, and market demand. Similarly, the media for storing music – or movies – evolved from vinyl to cassette to CD to digital storage, with each taking significant market share from its predecessors.

Here are some interesting examples of technology-based products introduced to society and how their introduction was received.

1. Automobile: The early stages of the automobile required three people to operate:  One person to drive, one person to walk in front of the car with a flag and a third person with a horn in hand. This introduction allowed pedestrians and horse-drawn carriages to stay out of the way of faster moving vehicles. Over time, all three functions were integrated into the car itself; horse buggies needed to be retrofitted with warning and caution signs in back to provide car drivers with adequate, advance warning to avoid collisions.
2. Regional jets: When regional jets were introduced, their performance characteristics differed from conventional commercial aircrafts in terms of climb/descent rates and nominal and maximum speeds. Such performance-characteristic differences required that air traffic controllers learn and understand these aircraft to better manage them along with other existing airspace operations.
3. Small drones operating at lower altitudes (less than 55 lbs. and 500 ft., respectively): Given that small drones have different performance characteristics than conventional aircraft (speed, climb/descent rates, piloting), enabling their integration into the national airspace system requires a different set up. A new construct called Unmanned Aircraft System (UAS) Traffic Management (UTM) was developed to safely accommodate small drone operations beyond visual line-of-sight at high densities. The construct focuses on cooperative, digital intent-sharing by leveraging third-party services for functions such as planning and tracking as well as management by exception (informing which airspace is not available to use rather than giving specific step-by-step directions) and letting UTM users develop and execute their plans by staying away from each other.
4. Larger drones (operating along with other manned aircraft at higher altitudes): Significant research and development efforts were necessary to determine the requirements to safely enable larger drones in the national airspace system. Without pilots onboard, air-traffic control communication and tactical detection of other traffic must be achieved through automation rather than humans for these drones.
5. Computers: When personal computers were first introduced, they needed to be inexpensive enough but with enough performance improvement to spur consumers to buy them. While we are likely to continue to see passionate early adopters of new computing technology, widespread adoption requires significant performance improvements at a reasonable cost. This is called “minimal viable product strategy;” instead of creating the most complex (and hence most expensive) product based on many needs, the computer has to strike the right balance between just enough performance improvement and affordability to build the market momentum. Early computers were an example of segregation rather than integration into our daily lives— we had to go to a separate room where these computers were located. Over time, computers, including laptops, tablets and smart phones, became more robust, portable and integrated into our daily lives.
6. Segway and scooters: The Segway is an interesting example of a technological introduction that promised to revolutionize transportation but never took off as expected; that’s because local authorities feared heterogeneous operations could pose significant safety risk to pedestrians. Segway did find a market in niche applications such as police patrols, tours in urban centers and getting around in malls. With shared bikes and scooters, we’re seeing a new attempt to improve short-range transportation in urban locations. We would argue that scooters are a good example of a minimal viable product that increases convenience for users in downtowns and city centers.
7. Commercial space transportation: At the time of this writing, Richard Branson and a crew from the Virgin Galactic space-tourism company had just safely launched into space before returning to earth and landing in New Mexico. In the next few months, we will see similar flights from companies run by Elon Musk and Jeff Bezos— all examples of increased activity in the commercial space-transportation sector. When space transportation was left to the government, airlines accepted flight delays and longer routes to accommodate launch windows and avoid debris fields to ensure safety. However, as commercial space operations continue to increase, an interesting dilemma is emerging:  Who has priority to access the airspace? In this example, the introduction of new technology has some impact to users of the current system. In order to maintain the highest safety, airspace is segregated for short periods of time until the space vehicle is launched or re-enters the earth’s atmosphere. While research is underway to ensure that such segregation is minimized in terms of duration and airspace volume, other operations could still be affected. The question of equity and access become increasingly interesting if there is a potential overlap of interests.
8. Voice-based products: We are beginning to see the emergence of voice-based technologies designed to accept natural language input and provide desired outputs such as access to on-demand information and changing environmental conditions. While these products are not always accurate in their interpretations given variation in voice inputs and accents, they provide key capabilities that offer a minimal viable product useful to consumers. Future voiced-based inputs will be integrated into devices such as computers, televisions and even home appliances, enabling faster actions relative to typing, much like what we see in smartphones.

A key lesson for the introduction of new technology is that it must be introduced carefully with the ability to interact or interoperate with existing ecosystems to prevent significant disruption to other users or preexisting operations— that’s one explanation for the slow introduction of autonomous vehicles.

Designing technology of the future

If new technologies are to be successful, we have to design future systems to be interoperable and accommodate the performance differences in heterogenous systems. Interoperability means that these tools can coexist with an understanding and accommodation of each other’s performance differences, similar to baby-proofing the house for a new child’s arrival so that everyone can co-exist – or interoperate – safely.

In a supply chain context, interoperability is a critical design characteristic whenever humans and robots work in the same space, otherwise we risk injuries like in industrial times when it was not uncommon for machinists to lose fingers to less-forgiving machines. Future systems will need to be designed to ensure no harm will be done to any actor in a heterogeneous system.

We see this play out in the current generation of autonomous vehicles: When they don’t know what to do, or confront an unfamiliar set of circumstances, they simply stop. In some respects, it’s not unlike what happens in our third-world example of a cow straying onto the road, bringing traffic to a halt— that reaction impacts efficiency but also maintains safety and removes uncertainty.

Unfortunately, however, that option isn’t always possible or viable. A jet airplane, for example, can’t come to a quick stop if a drone unexpectedly enters its air space. In that instance, we have to build in the interoperability among heterogeneous actors to maintain the overall safety of operations and the individual actors. It’s critical that strategic plans, tactical actions and intent are clearly understood by all to avoid collisions.

If a new technology or capability stands alone, it’s unlikely to have a negative, or even minimal, impact on existing capabilities (or even help improve effectiveness of the state of the art); in that instance, it doesn’t matter whether the system is heterogeneous or homogeneous. In such cases, a minimal viable product strategy can be deployed to build products that improve performance and are likely to be adopted quickly due to their affordability. An example could be dashboard operating systems like Android Auto, which improve the driving experience and have little impact on other cars on the road.

In the table below, we propose a framework based on ease of introduction of new products, improvements, capabilities or technologies into society, based on their performance characteristics and intended level of interaction with pre-existing systems.

While it would be easier to introduce new technologies into homogenous systems, the reality is that heterogeneous systems will continue to exist as a result of the differing nature of objectives, capability needs and cost of operations that compose a system. The key principles to make heterogeneous systems work efficiently and safely are segregation where absolutely necessary and integration where possible to afford maximum flexibility.

In order to integrate heterogenous actors, interoperability must be built-in to maintain awareness for all actors while ensuring overall safety and efficiency. The minimal viable product strategy allows for the introduction of improvements and technologies likely to be adopted by society and capable of being integrated with the rest of the system.

About the authors
Parimal Kopardekar, Ph.D. serves as the Director of NASA Aeronautics Research Institute (NARI).

Shekar Natarajan is the chief supply chain officer for American Eagle Outfitters.