Typical mistakes when scaling the network - Part 3/4: Strategy

"All men can see the processes whereby we perform, but what none can see is the strategy out of which excellence is evolved." (The Art of Logistics)

Welcome to Part 3 of our article series on scaling logistics network. We have gotten ourselves an overview about the timing („How to find out when new capacities need to be created?“) and the location aspects („What needs to be done to identify where to ideally place these new capacities?“) of such networks. When tackled in the context of growing businesses these two questions typically have an operational and periodic character. One should consider one’s capacity curves and business cases every six months and, if growth is steep enough, place the required additional capacities on a map. There are, however, countless strategic questions that set the framework for operational process we just described.

„What is the network strategy that we are operating in?“ Our strategy defines a majority of the ground rules that prevent timing and location matters from becoming complex, multi-variate problem statements. A possible path to such a strategy and potential issues along the way are described in this article.

How to define a network strategy for e-commerce logistics

Developing a network strategy is the theoretical starting point of scaling your network. Before thinking about timing and location, one should ponder an optimal amount and the best-suited types and sizes of nodes and connections for a network of a given scale.

In reality this is rarely the case, for perfectly understandable reasons. One of those is insecurity, with respect to planning accuracy on the one hand and to stability of business approach on the other. At the point in time at which a company may still change its network setup with next to no switching cost there is rarely perfect clarity regarding material flows and customer expectations during the next 12 months, let alone during the next 10 years. Another reason is the meagre short-term return on investment of such an endeavour. Developing a strategy requires resources, oftentimes even the help of external experts, but benefits scale with the network, are marginal at first and difficult to measure in general. Once these benefits did scale to the intended scale, they become a significant competitive advantage. The long-term return of such efforts is very appealing, but when a network is still young, strategy projects are rarely a top-priority.

Instead, e-commerce companies iterate step-by-step, towards a solution that is then perceived as optimal. One logistics location, capable of everything they need, for starters. Once it hits full utilization (or, preferably before it hits full utilization) a clone is added to fit the setup and scale as simple as possible. An alleged point of no return is approaching on tip-toes, since at some point the network is too stuck in its ways, to perform a strategy update at more or less justifiable switching cost. This point of no return may be considered purely theoretical, since any company interested healthy logistics in the long run should be motivated to optimize its strategy at any point in time, even if the breakeven for that optimization is a few decades away.

So, how to avoid such a pickle? First and foremost, it is advisable to give the topic of logistics strategy some thought, as early as possible. If that is not possible for whichever reason, it may be beneficial to maintain a setup that is “strategically flexible”, in other words: actively keeping potential switching cost low. Operating thus one should then periodically check whether a long-term definition of an optimal logistics strategy is achievable and kick-off a change project without hesitation once it is. In any way it is worth mentioning that, by tapping into industry knowledge and an outsider’s point of view, an external partner can help significantly when defining a such strategies.

Once this hurdle is overcome and you find yourself in a kick-off workshop on you new logistics strategy you may ask yourself which decisions are grouped under the terms “logistics strategy” or “network strategy”. We argue that the former actually includes the latter, in addition to a series of technology and process decisions that are taken on object-level (e.g. warehouse-level). One may call these decisions “process strategy”. Optimal technologies and process landscapes obviously also depend on the network setup. Thereby the logistics strategy keeps it all together and ensures harmony between the network and the object level, or between network and process strategy. The network strategy, being our focus for this article, treats the aforementioned objects like blackboxes with respective characteristics and functionalities.

Despite not considering the object level in detail, there are countless decisions to be taken. Network objects can have very different features and requirements, in pretty much any thinkable intensity, and may be combined freely. One could pair classic warehouses with satellite locations, with cross-docks, buffer or central storage locations, own return or even distribution centres, just to mention a few. On top of that one may establish several different steering strategies in nearly any setup. You can store your wares redundantly or assignment-based. One may assign single or grouped fulfilment centres to certain markets, or proactively move assortment towards likely buyers. Indeed, unlimited possibilities.

The path to an optimal network strategy typically features at least two steps. First, one needs to identify and define suitable strategic concepts. A good starting point for that may be a catalogue of requirements (customer, business, legal, environmental, …). This process already roughly evaluates these strategic concepts implicitly, since those that do not make sense intuitively will end up discarded.

Subsequently, the identified strategic scenarios may be evaluated quantitatively. That is pretty challenging, due to the rather high level of uncertainty, but it is possible and not necessarily a huge project. Based on this the best-possible setup can be identified and fine-tuned via an optional third step. That last step may indeed be considerably effort, requiring complex modelling and possibly specialized software, but it thereby confirms the results of step two and provides valuable insights with respect to sensitivity and sizing, in addition to a more reliable benchmark for retrospective controlling (not a small thing!).

In conclusion this topic may well be considered one of the most challenging but also one of the most exciting in the field of logistics. I’ll gladly discuss individual claims of this article or your own experience in this context, anytime via LinkedIn / Xing / Contact.