Sometimes, we see companies with fantastic products – and they fail commercially. Sometimes, we see companies with fantastic products, and they can even solve real problems for real customers – and yet, they fail commercially.
Now, there is certainly no shortage of reasons why companies can fail commercially in spite of having fantastic products, and I will not attempt to enumerate and discuss them here. Instead, I would like to highlight only one very specific reason which in my opinion often remains neglected, or even unnoticed altogether: competitors or substitutes that are good enough.
Good enough is a very powerful concept. I claim that having a product that is good enough is often preferable to having a product that is superior. Let’s have a look at three examples of products or systems in the warehouse automation industry that are failing to meet commercial expectations through the lenses of “good enough”.
Automated Case Picking Systems: The Rest of the World vs. Witron
I had the pleasure of working with the Food Retail division of SSI Schäfer for some years. At the time, SSI had some fantastic products and it offered an automated case picking system (called Schäfer Case Picking System, SCP) which, from a technical point of view, was most likely market-leading. Our system could handle the broadest range of SKUs. We could build the best (highest, most stable) pallets, taylored to each individual store’s layout. With the 3D Matrix system based on the Navette multi-level shuttle, we were able to offer the best combination of performance and sequencing. Our visualization system looked fantastic and our warehouse management system was well-suited for all food retail-related requirements. We spent good money on marketing the system, we presented it on all major trade-shows, and our sales engineers were fantastic and knowledgeable about all things food retail.
And then there was Witron. Witron was, still is, the market leader for automated case picking systems. I frequently dreamt of Witron, and these were not pleasant dreams. We competed against Witron in every bidding round, often on the short list, and to cut a long story short it shall suffice to say that Witron’s order intake and financial performance has never taken any noticeable hit due to our existence. We knew for sure we were better than any of our other competitors (Vanderlande, Dematic, Swisslog,… ). In fact, we did not even consider them serious competition. Also, we knew our solution was better than Witron’s – from a technical point of view. Even many of our leads could see that. There was but one problem: Witron’s solution might not have been perfect, but it was good enough. It was cheap and it was proven by dozens of references. Witron has built the same solution for twenty years. I have visited several Witron-engineered warehouses built between 2003 and 2018, and except for small details they looked effectively the same, and they worked just fine. Witron’s palletizing robot was most primitive, and still is. While Witron’s palletizing robot was one simple mechanic arm that would push cases onto a pallet from one side only, our palletizing robots were able to build perfect, beautiful pallets by accessing them from all four sides. Witron used Miniload cranes, decades-old technology, while everybody else used single-level or, even better (as we did), multi-level shuttles. Witron’s solution required two separate Miniload warehouses, one for storage and one for sequencing, while we would store and retrieve in a sequenced manner from the same warehouse, which is much more elegant and allows later order cut-off times. Without going into the details, we could shine with a good number of technical benefits over our competition, and over Witron in particular. Such a pity that customers did not care. If somebody offers the same product or system for twenty years (and keeps improving it on details), it gives them credibility. And it gives them a cost advantage that is impossible to match if you want to use the best and most fancy technology. On all relevant metrics, Witron were not the best, but they were good enough, which in combination with cheap enough and proven provides a value proposition that is almost impossible to beat if customers are risk-conscious and think economically.
AutoStore: Almost Always Good, Rarely “The Best”
Concept and technology behind AutoStore have been around for more than 25 years. It took quite a while for Hatteland to gain traction, and for good reasons: from a logistics point of view, having multi-deep storage for small-parts goods-to-person picking systems does not seem like a reasonable proposition. Once it was sufficiently demonstrated that the system actually did what it was supposed to do, and very reliably so, it began to spread like a wildfire. Distributor after distributor signed on, creating the remarkable situation that several companies would compete for the same customer to sell them the same product. (For this reason alone, Hatteland deserves an award for the best business model).
The commercial success of AutoStore in combination with an excess of venture capital and the silly hype of Microfulfillment Centers has led to a founding spree of new AS/RS systems for goods-to-person picking, as well as to comfortable funding of existing start-ups that did not really go anywhere in previous years. A warehouse operator interested in building a goods-to-person picking system for small parts today can choose between dozens of technologies and vendors, from the more traditional technologies like Miniload cranes, single-level (roaming or captive) and multi-level shuttles, and carousels to Alert Innovation’s Alphabot, Exotec’s Skypod, Attabotics, Fabric, stow’s e.scala, Storojet, Blue Robot, AMR-based solutions (Geek+, Quicktron, and the like), and Jungheinrich’s PowerCube, most of which have not existed ten years ago. The task of choosing an appropriate system has become exceedingly difficult for many customers. Interestingly, the increasing availability of different AS/RS does not seem to have had any impact on the commercial success of AutoStore and its distributors in any significant way (AutoStore insiders, please do correct me if I’m wrong). On the one hand, one reason certainly is that market demand for automated solutions in general has been high for some years. On the other hand, I simply cannot see that AutoStore’s younger competitors, any of them and all of them together, are making a dent on the market. Sure, everybody is able to sell a couple of systems as innovation projects. Only once the phase of innovation projects is over can we see if a system really gains traction on the market. And I simply don’t see this happening as of yet. In fact, AutoStore has eaten a good piece of the cake of incumbent vendors and their traditional systems, too.
Is that because AutoStore is “the best”? I don’t think so. AutoStore is a good system with impressively high reliability and availability and short project lead time. And for companies coming from (often messy) manual systems, it can offer quite a leap in terms of performance (and orderliness) of the warehouse. When talking to companies who have deployed AutoStore in their warehouses, however, I am often surprised about the relatively low picking performance achieved at the picking stations. The reasons for low picking performance can be manifold, and not always is AutoStore to blame (though anecdotal evidence indicates that there seems to be a tendency that more bots should have been included in the initial setup). The important point is, however: these AutoStore clients are still happy. Have they got the most productive system they could have got on the market? Often not. Have they got the most bang for the buck? Sometimes not. But they sure got a system that broadly does what it is supposed to do in the most reliable and predictable way, a system that everybody understands, and one that is installed in an almost plug and play fashion. Except when the pareto distribution of items to be picked is too flat, it is a system that almost always works with almost any order structure and sequencing requirements. Rarely is AutoStore “the best” system, however: Miniloads can make much better use of building height, shuttle systems can support picking orders with much lower lead time and are much less sensitive to the ABC distribution of orderlines, Exotec’s Skypod and stow’s e.scala have much lower requirements on the floor slab, AMR-based solutions can be deployed even faster and come with even less infrastructure overhead, Jungheinrich’s PowerCube can deal with higher bin weight and can be built twice as high. But even though AutoStore almost never scores best on any particular metric, it is either very good or, more often, good enough, and most certainly it is a safe bet. “Nobody ever got fired for buying IBM”, and the same might be true for AutoStore. Which is why life will be very hard for Jungheinrich with their recently launched PowerCube, which attempts to be a better, yet technologically more complex version of AutoStore. If you had €10.000 to buy one car for you and your family, would you buy a VW Golf that is known to work, or an Alfa Romeo? See…
Celluveyor: Unnecessarily Fantastic
The last example is Cellumation’s system Celluveyor. Roughly speaking, Celluveyor is a conveyor module with built-in sortation and diversion functionality. The element consists of a number of individually-driven wheels that can move cases in any direction. You can think of it as a bunch of connected drive units that would normally move mobile robots, flipped upside down, and in fact, that’s where the technology came from originally.
Due to the versatility of these drives, you can use a Celluveyor element for a range of applications. For instance, you can use it for sorting, for descrambling after layer-wise depalletizing, or for pre-building pallet layers prior to palletization. None of these applications is new, all of them were possible before with much simpler, less sophisticated technology (conveyors). Using conveyors for descrambling requires more space, but is exceedingly simple. Pallet layers don’t require pre-building, you could simply use an articulated robot and build the layers right on the pallet. And for sortation, there are plenty of existing solutions. The Celluveyor can do it all. But it’s not needed, because it is fixing a non-broken part. Except for those rare cases where space is of essence and you want to save every square meter you can, I suppose conveyors will be good enough. Celluveyor is here to stay, but I do not expect Cellumation to reach mind-boggling sales numbers for it.
Summary: What “Good Enough” Really Means
The three examples provided above have in common that the alternatives to the technologies or concepts that I described as “good enough” are often better on various metrics, are technologically more sophisticated – and typically are more expensive. For each additional increment of technological sophistication, we get decreasing marginal value and disproportionately increasing cost. In simple terms, the additional increment of technological sophistication still makes the products or systems better, but they are becoming even more complex and even more expensive than they are becoming better. “Good enough”, on the other hand, means we are riding the sweet spot on a Pareto curve of value (Y-axis) and cost (X-axis). And that’s the beauty of it. Sometimes, you need to go for technology that allows you to meet extreme requirements. Sometimes, customers want the very best solution they can get. But more often than not, companies are trying to make economic decisions, one important aspect of which is “proven to work”, whereas a highly sophisticated and expensive solution overshoots (Christensen) a large share of the market. It is totally possible for a product to be better than necessary, and this typically makes the product more expensive than necessary, too. A commercially successful solution does not need to be the very best, most sophisticated one; the people in the customer’s buying center who are making the decision often simply want something that works and they don’t want to lose their job over a fancy yet risky decision. So, they are choosing the VW Golf, not the Alfa Romeo. Which does not mean that there is no market for Alfa Romeo, but the market for “good enough” is much bigger than the market for “the best”.