Experience shows that large automation projects are risky and often unprofitable. Yet, pursuing these ‚white elephants‘ remains a core activity for large sales teams in nearly every logistics automation company. With the rise of numerous technology firms and unprecedented demand for logistics automation, the number of large-scale projects continues to grow. This article aims to explore why these projects often fail to deliver the expected benefits.
Why Discuss the Risk of Large Projects?
During my time in the warehouse automation industry, I rarely saw a large project that turned a profit for the system supplier. Some projects were beneficial for the customer—though certainly not all. In fact, most very large projects fail to meet customer expectations as well. Once a project surpasses an order volume of around €50 million, it typically takes several smaller projects just to compensate for one failed large one.
I once raised this issue during a project management seminar attended by professionals from the oil industry. To my surprise—since I hadn’t yet formed a strong opinion—they confirmed my suspicion: their large, prestigious projects usually lost money. Despite senior management’s attention, these initiatives routinely failed to deliver profits.
A few days later, on a flight back from a client meeting, I sat next to a commissioning engineer for a company that builds large mixing machines. He had worked on projects of all sizes in over a dozen countries. I asked him about typical project sizes and the profitability of large-scale efforts. Without hesitation, he responded, “No, the big projects all ended in a loss. But management keeps chasing them.”
While anecdotal, these examples—from unrelated industries like warehouse automation, oil and gas, and industrial machinery—point to a systemic issue. It appears that project risk increases non-linearly with size. I’ve since reflected on the reasons behind these failures and attempted to categorize them.
Reasons why Large Projects are Risky
Size Increases Complexity
One of the most important risk drivers is complexity. Simplicity is a virtue in system design, but what exactly is complexity? A complex system’s behavior cannot be fully understood by analyzing its components in isolation. If taken apart, the system ceases to exhibit the very phenomena we aim to explain. This is due to emergent properties—features that arise only when all components interact together. This contrasts with complicated systems, which involve many elements but still follow a reductionist logic.
Importantly, a large system isn’t necessarily complex. A very simple but large system—like a high-bay pallet warehouse—may pose few challenges because the key drivers of complexity are absent.
Two of these drivers are numerousness and interdependency. A higher number of elements leads to more possible combinations. This alone results in complication, not complexity. But when combined with interdependency—when a change in one element affects others—the system becomes complex.
For example, changing the load carrier in an automated warehouse might require adjustments to racking, conveyors, palletizing robots, dolly stackers, SKU allocation, net volume calculations, replenishment frequencies, and more. The robotic picking system may not support the new carrier. It might be unsuitable for cold storage or incompatible with delivery vehicle dimensions. You could easily miss something, and starting over might be the best option.
Or consider a customer increasing SKUs from 8,000 to 20,000: the entire logistics concept might need redesigning. A previously sufficient front zone loop might no longer work. A system optimized for worker productivity may suddenly need to support one-hour delivery, requiring an entirely different approach.
What makes this worse is that cause and effect are often not obvious. Effects can be indirect or delayed, making learning and adaptation difficult. There is time delay.
Two further complexity drivers are variability and variety. Variability is typically undesired and refers to inconsistency in inputs or outputs—like fluctuating demand or quality deviations. Variety, by contrast, is often intentional, providing customers with choice. But variety can induce variability—hence car manufacturers bundle options to reduce it.
The impact of complexity on risk is clear: it increases the likelihood of failure and makes root causes harder to identify. And because larger projects involve more money, the consequences are more severe—e.g., penalties for delays. So, complexity increases both the probability and severity of failure.
Size Becomes Addictive
Large projects can become addictive, drawing companies into a cycle of ever-larger commitments.
This addiction stems from three forces:
- Psychologically, large projects become the new normal. After securing a €100 million deal, €10 million projects may seem unimpressive—both to yourself and stakeholders.
- Economically, unprofitable large projects create a need for more big deals to recover losses.
- Organizationally, teams built for large projects may be underutilized afterward unless another major project follows.
As with any addiction, large projects must be approached with care, awareness, and restraint.
Size Creates Ripple Effects
Because of their risky nature, and because they are most helpful for managers’ egos, larger projects receive more management attention than smaller projects. Besides the fact that some people will get jealous as they feel their smaller projects do not receive the attention and recognition as they used to, which makes more people want to pursue bigger projects (which brings us back to the previous point about addiction), this also has a more serious implication: If something goes wrong, additional resources often will be allocated to the big project to fix the problem. And obviously these resources are not normally sitting idle somewhere, but they are working on something, normally on other projects.
So instead of one big project blowing up, you could have five or so projects blowing up which cannot be finished on time since key resources are retracted. Since resources don’t magically multiply and good resources in particular tend not to be idle, chances are you will create quite a ripple effect in the moment you are putting out fires in a big project, which, in the end can lead to the situation that you are creating more problems than you solve.
I found this effect particularly powerful with IT problems in big projects. IT may be the one part of business which senior management understands least. And so senior management often believes that putting more resources on an IT problem will help solve it quicker when, in fact, the opposite might be the case. It will hurt all other projects, however, which will run the risk of being delayed.
Opportunity Cost
Finally, there is a category of cost that you may not even see: opportunity cost. The cost of the best alternative that you did not choose. Instead of keeping 10 people busy with one project worth €100 M, these people could have worked on three projects, each worth €30 M, and chances are each of these projects would have made more money than the big one. And this is the true cost of such a project.
Escalating Project Size can Mark the End of a Boom Period
Here is one macro observation.
If it feels like it’s raining mega projects onto you, and projects by tendency become much larger than they used to be, this indicates people are losing their risk aversion. They feel too safe.
You normally begin to feel safe when things go well for a long period of time. In the economy, this is during the boom period. And if the boom period has been long, chances increase that it will be over soon.
This in and by itself does not have to be a problem – because maybe you have signed the contract for this large project already. But if not, it should give you reason to double-check your potential future customers’ financial strength and vulnerability before you commit to something. There are some industries which remain largely stable even throughout economic recessions, such as the food retail industry (everybody needs to eat); and there are others which are quite sensitive to economic fluctuations. But even if someone’s industry remains stable, there can still be changes, for instance with respect to terms and conditions to obtain a bank loan for a big project. Out of risk considerations, banks may straight out reject loan requests – as they did in 2008 in many cases – thereby sending the entire economy down a spiral, with many otherwise promising projects being left unfunded and stopped.
Consider that large projects have long project lead time, both in planning and in realization. So even if things still look good today when a potential customer approaches you with a project, by the time you want to sign the contract things may look differently, and even more so by the time the project is under construction.
(This article was published pre-CoVid-19. Let’s say the economic downturn in 2020 is a case in point for many industries, though certainly not for logistics automation).
Which leads me to my next point.
If Your Customer Has a Problem, You Will Have One, too
Now, if we indeed face the situation that the economy – or at least one specific industry – is going down the drain and cash becomes a rare resource, your customer’s problem can very easily reflect on you and become your problem.
In projects, the cash flow you receive from your customer is often – and should be – somewhat balanced with the expenses you are having. This is important as a hedge against the risk that you have significant expenses – and the cash flow coming from your customer suddenly dries out.
But even if this is the case: We said earlier that big projects make addictive. It can very well be that your organization really needs the cash flow from this project for other reasons, maybe to finance growth, or a new building, or to compensate for hick-ups in other projects. So if that cash flow from your customer dries out, it can easily happen that your organization is in big trouble. And obviously more so, the bigger the project is.
Don’t Risks and Benefits of Larger Projects Balance out?
Occasionally I have heard the argument that with larger project, it is of course not only the risks, but also the benefits, that are significantly higher. So shouldn’t larger projects be “neutral” from a risk perspective?
I believe this is not true for three reasons:
- While you can clearly have bigger benefits from a bigger project, the direct benefits are quite limited. It’s more money that you can make. And, maybe more importantly, you can prove to the world that you are capable of doing large projects and that you should get another one. But still – this is quite limited. On the other hand, when things go wrong, there is no limit to the downside. In the worst case, one huge project going wrong can wipe out your company. So upside is limited, downside is unlimited. And this imbalance between risk and rewards is important to consider as it tends to become larger as the projects become larger.
- The only way to have a larger project risk neutral is to grow your company at the same rate as the potential project risk. But that’s almost impossible and pure theory.
- The margin you could earn with a project grows with project size in a linear way. The risk of failure grows in a highly non-linear way. Thus, risks and benefits are only balanced in small projects; the bigger the project gets, the bigger the gap between risks and benefits grows.
Therefore, a larger project will always be more risky than it will be beneficial, all other things being equal.
Conclusion
So – large automation projects are risky and are likely to be unprofitable. With all that said, what should we do? Should we walk away from large projects and let competition try?
One of the most important things we can do to avoid a problem is to be aware that (a) there may be problem and (b) to understand why. So I hope above discussion provides a useful contribution to awareness and better understanding of the risk in large projects.
You can’t always choose which direction a project goes. But when you can, you should have serious discussions with your management, with your engineering team, and with your customer. My frank suggestion is to be clear about the risk involved for both, the customer and yourself, when project size escalates. My experience is that customers appreciate honesty.
Another thing to consider is that large is not the same for everyone. Witron has been hugely successful with a copy-paste approach. When their customers let them do what they think is right, even a large project is fairly low risk for them. Yet, when the customer has other ideas and attempts to dictate Witron on how to do the project, even they will run into chaos, as it occasionally happens to them. So it is not „large“ alone which works as predictor of unprofitable chaos, but rather the combination of „large“ and „new concept“ which turns out to be hazardous. Outside of Witron’s copy-paste world, most large warehouse automation projects are highly customized, however.
Also, better understanding of the risk of large projects should reflect on your project portfolio in the sales funnel. It may be totally acceptable to pursue one super-large project if your company is not dependent on its making profits. After all, these projects do challenge your organization and help you grow competencies and a reputation that you may otherwise never be able to obtain. But having too many large projects is likely to be unhealthy for your organization. As a rule of thumb (for the logistics automation industry), not more than one hot project in your sales funnel should exceed 50% of your target order income of a given year.
I’d be very curious to hear about your experience with large projects. If you want to share your experience, send me an email or reach out to me on LinkedIn.
Some Additional Remarks
This Article is mostly about Suppliers – How About Customers?
Based on several discussions with experienced professionals from the logistics automation industry, as well as on own experience, it is fair to say that super large projects in the vast majority of cases not only make suppliers unhappy. They also fail to deliver on customer expectations. Way too often, they result in a lose-lose situation. Fascinating, is it not?
Large Centralized DCs versus Microfulfillment Centers
There is an interesting discussion currently going on in the field of online grocery services. While some eGrocery players bet on very large highly automated DCs, Microfullfilment has become a competing fashionable concept. Australien retailer Coles felt they had to defend their decision for a large centralized DC concept [1] after investment bank Jefferies called US retailer Kroger’s deal with Ocado a potential „multi-year mistake“ [2]. Nobody really knows which concept is going to yield most benefits since neither of them has been tested at large scale over a longer period of time. There is good reason to believe that unit economics are much better in large centralized DCs, but that statement is not limited to Ocado’s concept and technology.
[1] https://www.afr.com/companies/retail/coles-defends-150m-ocado-centralised-fulfilment-centre-deal-20191017-p531ir, last access on 2019-12-24
[2] https://www.cnbc.com/2019/10/10/this-fulfillment-method-could-a-multiyear-mistake-for-grocery-chains.html, last access on 2019-12-24 . It should be noted that the Jefferies report is based on a some wrong assumptions and there is good reason to believe that correcting these assumptions will lead to opposite conclusions. There does not seem to have been any debate regarding the validity of the report, which is concerning.
Podcast Version of the Article
I have recorded a (slightly rewritten) audio version of the article for the Warehouse Engineering Podcast. You can find it on Youtube or listen to it below.
Update History
- 2023-01-13: Minor revision
- 2023-02-18: Minor linguistic revision
- 2025-05-11 Major linguistic revision