Automation gone sideways

Avatar photo
by Treena Hein

Falling in love with someone or something cool and new is natural, but while it may be exciting, it also carries risk.
For DC managers or company leaders who become infatuated with wondrous-sounding automation solutions, it can be a truly deadly mistake.

“History is littered with the casualties of companies that fell in love with technology or designs that were disconnected from real market demand,” Rob O’Byrne, group managing director at supply chain consulting firm Logistics Bureau, noted recently on his firm’s website.

Besides disengagement from the needs of customers, there are other serious reasons why companies fall prey
to mistakes relating to automation – mistakes still being made today across the globe.

To find out more, we contacted three experts: Charles Fallon, principal at consulting firm LIDD Supply Chain Intelligence in Montreal, Quebec; Jeff Christensen, vice-president of product at Seegrid, a provider of connected self-driving vehicles; and Nick Klein Schiphorst, director of business development at Dematic Canada.

Read on to learn their take on why companies go wrong and some of most common examples of automation misuse from the past and present. They will also discuss particular types of automation and other factors that typically lead to misuse, and how firms should go about automating correctly.

Why does misuse of automation happen?

Jeff Christensen

Everyone wants to jump into solutions right away, but that is fraught with peril if you do it too soon. We have to figure out what problems we’re trying to solve, what is the business problem specifically. You have to define those before you enter the solutions space. Without properly defining problems, you also can’t measure your solutions, and whether you’ve actually solved the problem, and it’s easy to make subjective judgments rather than objective ones.

Misuse happens when folks begin with an unclear or under-defined reason for putting in automation. A lot of industry news and marketing can make people feel like they are missing the boat to the 21st century if they aren’t actively planning or implementing an automation solution. But very few companies are good at identifying the real reasons that could make such a solution viable for them. Without clearly defined business objectives, it will be near impossible to navigate the hundreds of critical decisions that need to be made in such a project, such as fundamental decisions about what technologies to adopt.

Even when the business objectives are clear, the planning and implementation of automation is extremely complex and unforgiving. The most important part of the planning and implementation is the technology. A company’s WMS will change and a new WCS (warehouse control system) will come to play as a critical part in the operation. These systems must talk. And each system must perform its tasks. However, information technology is often not given a high enough priority on automation projects. And too often, this is the reason for delays and underperformance.

Klein Schiphorst:
One cause of misuse is design that has not taken all aspects of the business into account. Consider an automated system where the design did not take into account all characteristics of the products to be handled, such as types, sizes or weights. When certain products cannot be put through the automation – where the business case assumed they could – the negative impact on the ROI runs up quickly. It’s not an optimal use of the investment, and manual processes are still needed for ‘non-conveyable’ products.

Also, the impact of specials/promotions on the supply chain, if not taken into account in the design phase, will result in the automation not being able to cope properly with sporadic high demand.

Another cause of automation misuse is incorrect input into the design. For example, forecasting a 10-year growth rate of one percent versus an actual two percent rate means the system is outgrown way too soon. If the reverse is true, the system is too large to begin with and the ROI is pushed out many years.

What types of automation are most commonly misused, and why?

All types have the potential to be misused if they don’t line up with the defined problem.

Nick Klein Schiphorst

Klein Schiphorst:
I do not have statistics on what types of automation are most often misused and also do not believe that it boils down the type of automation – more to the proper/best design being engineered. That word ‘engineered’ is key.

It’s not so much that any particular kind of automation gets misused. It’s that there’s a pattern of faulty thinking that creates sub-optimal automation solutions. As I said, this starts with unclear business objectives. Then, in the design process, the systems are generally undersized due to improper modelling of the item and volume profiles to be handled.

People tend to forget that nothing works in a predictable, steady state – the whole point of warehousing and distribution is to buffer against supply chain variability.

What are some of the misuses of automation that have popped up over the past 10 years or so? What can we learn from the past?

Robotic arms, both small and large, have been tried, and they never really did the job so they’ve fallen out of favour.
Also, the idea of having a dedicated space, separating automated processes from manual, creates operational inflexibility. The industry is moving much more towards fluid hybrid environments where robots and humans work together using their individual core competencies. Forced segmentation will wane. Humans are really bad at boring repetitive tasks, and when they are bored they make mistakes and get injured. They are also bad at random access to know what to do next. Robots excel at both repetitive tasks and random access.

A classic misuse is technology which commits a business to exactly one way of operating, which makes a stark line between manual and automated parts of the process and creates intrinsic barriers to change.

Charles Fallon

This is a hard question to answer because the last ten years we have seen some new technologies emerge and we are still on the first generation of those technologies. Specifically, for case-handling automated storage and retrieval systems, I think folks are finding ways in which this technology can be deployed effectively. Originally, they were often implemented as a means to stock and handle slow-moving product. This can be very effective; however, people have to keep in mind that these items tend not to generate significant labour if contained in a separate picking zone.

Otherwise, there are interesting applications at the manufacturing level of this technology both in terms of work-in-process and finished goods where the technology can create opportunities to add value that conventional systems could not.

Amazon’s purchase of KIVA convinced many across the industry that this was a game-changing technology, and several years later, you see many copy-cat technologies on display at trade shows like Modex and ProMat. However, interestingly, Amazon’s latest Canadian DC, which is 600,000 square feet in Calgary, Alberta, uses a Honeywell-Intelligrated materials handling system with classic piece-picking technology like unit sorters. I think that speaks volumes for how effective KIVA has been in Amazon fulfillment centres and should raise a note of caution in anyone looking to invest in a copy-cat technology.

Klein Schiphorst:
In the past, systems were designed to handle one product type and suddenly the market demands that various products come through the same DC. If the automated system is not designed to be flexible and cannot handle the new products, the automation becomes only useful for a piece of the business…In today’s market, end users are trying to find ways to be flexible in their system designs in order to be able to take on whatever the business decides they need to get into, to survive in the market.

Are there types of DCs that are more prone to misusing automation, in terms of their layout, geographical location, category of goods
and so on?

Many emerging companies – especially those staffed by Silicon Valley types – want to leap into automation. I have seen countless exciting new companies hobbled by bad decisions because they thought their innovative brand or product needed to be supported by innovative distribution systems.

European facilities are more likely to automate physically and have a higher emphasis on safety standards, but plants tend to be smaller which can make automating processes easier to tackle. Plants are larger in the US, which makes automating more complex and we look to productivity gains first.

E-commerce fulfillment has to move toward automating due the product mix, and automation there can be prone to misuse if DCs move too quickly into solutions that might not be best. I think the expectations of the market have surpassed the capability of E-commerce firms; they are working to catch up.

How do distribution firms choose the right road ahead?

Klein Schiphorst:
It appears that more and more supply chain experts are becoming increasingly savvy in regards to what to expect from automation. With today’s intense networking in the relatively small word of DC logistics and the availability of information, those with supply chain responsibility are at least asking the right questions and homing in on the key points. This will reduce the ‘misuse’ of automation.

The future is bright for automation in the supply chain world. Consumer demand for more variety, faster delivery and faster-changing SKUs will drive the need for more and more flexible automated systems and the key providers of these systems are spending many hours designing the systems of tomorrow.

I think understanding the business process planning, industrial engineering and financial planning (to understand ROI) are all important, and don’t forget to define your problem. You shouldn’t just automatically automate everything. Look for where you get business value and the most out of your human workforce and future automation workforce. Think before you walk or sprint. The growth rate of automation in DC will jump, and the technological capability to keep up is available.

With each generation of investment, the industry will learn and avoid some of the mistakes that we’ve seen in the last ten years. For some, this will mean turning their backs on automation completely. For many, this will mean a much more sober approach to the justification, design and implementation of these systems. And developments in self-guided vehicles should broaden the opportunity for hybrid materials handling systems where automation augments conventional, labour-driven systems.

Key take-aways

• Knowledge of how to properly automate is growing, and design of effective automated systems is evolving quickly.
• Automation misuse happens because firms lose sight of customer needs, jump into solutions before properly defining the business problem, or chose the wrong technology.
• All types of automation have the potential to be misused if they don’t line up with the defined problem.
• Even when business objectives are clear, the planning and implementation of automation is extremely complex and unforgiving.
• The industry is moving to fluid hybrid automation environments where robots and humans work together using their different core competencies.

Do it right
• Scope your business requirements more completely than you think necessary.
• Assess ways to solve the operational issue other than automation.
• Ensure the IT associated with the automation system is properly and completely implemented.
• Beware the pitfall of copying technology just because it’s used by others in the industry.
• Design your automated system to be flexible and handle new products.