Humans in modern supply chains
Supply chains are complex systems, possibly among the most complex ones ever engineered by mankind, encompassing people (many), machines (diverse) and software (tons). The perspective I recently posted on DDMRP generated quite a lively discussion. This lead me to further ponder the fundamental differences between Quantitative Supply Chain Management (QSCM) and DDMRP. At their core, these two visions are profoundly disagreeing on the role of humans within supply chains.
QSCM is firmly anchored in the classic IBM vision:
machines should work; people should think;
In contrast, DDMRP adopts a mass education stance, best summarized by its motto:
built for people not perfection;
While the philosophical stance towards humans in supply chains does not explain all the differences between QSCM and DDMRP, it sheds light on why these two perspectives are irreconcilable to some extent.
Scarcity vs abundance of supply chain practionners
Supply chain practitioners are recognized as a valuable resource for the company by both DDMRP and QSCM. However, the two approaches differ quite sharply in how this angle is factored into their respective methodology.
QSCM starts from the hypothesis that every mundane supply chain decision should be automated1. This perspective emphasizes that competent supply chain practitioners are considered too rare and too expensive to spend their time on generating stocking, purchasing or pricing decisions. All those decisions can and should be automated, so that the practitioners can focus on improving the numerical recipe itself. From a financial perspective, QSCM turns those salaries from OPEX where man-days are consumed to keep the system rolling to CAPEX where man-days are invested in the ongoing improvement of the system.
The DDMRP angle starts from the hypothesis that competent supply chain practitioners can trained en masse2, thus lowering both the cost for the employer, but also reducing the bus factor associated with the departure of any employee. Also, by adopting numerical recipes specifically tailored for human processing, the OPEX investments can themselves be reduced. DDMRP establishes a process to generate mundane supply chain decisions, but achieving full automation is mostly a non-goal3, although DDMRP isn’t averse to automation whenever the opportunity arises.
Interestingly, whether the industry is steering towards the QSCM perspective or the DDMRP should be observable to some extent. If the QSCM perspective gets adopted more broadly, then supply chain management teams will evolve to become more like other “talent” industries, e.g. finance with their quantitative traders where a few exceptionally talented individuals drive the performance of large companies. Conversely, if the DDMRP perspective is adopted more broadly, then supply chain management teams will evolve to become more like successful franchises - e.g. Starbucks store managers - where exceptional individuals have little effect on the system - but where a superior culture makes all the difference between companies.
Local vs global transparency
Both QSCM and DDMRP strive to avoid the black box effect which is inherent to any attempt at optimizing a complex system. Both approaches value the idea of achieving a degree of supply chain transparency; however due to divergent initial assumptions, both approaches end up with wildly different perspectives on what transparency entails.
From a QSCM perspective, transparency must be achieved first and foremost at the management level through explicit quantified economic drivers4. Each decision produced by the system should be backed by a series of drivers - measured in monetary amounts (e.g. dollars) - which motivate why this decision is put forward. For example, a purchasing decision is motivated by the extra margin that will be generated by having some extra inventory (rather than not) but also negatively impacted by carrying costs and an increased risk of inventory write-off . The management is in control of the economic drivers, and, at the system-level, QSCM is highly transparent: the system merely rolls out at scale the complex but mundane implications of those drivers. Yet, the downside of such a system-wide optimization is that deciphering the fine print of any given decision is complicated, precisely because every decision is a complex balance of many drivers assessed against many possible futures.
From a DDMRP perspective, the transparency is intended and delivered at the operational level. The simplicity of the numerical recipes ensure that every decision can easily be assessed as correct simply by “guesstimating” what the result should be. Also, replicating the calculations in a spreadsheet always remains possible. Furthermore, through priority lists, DDMRP mitigates the inherent complexity of supply chains, by providing an attention mechanism to supply chain practitioners, so that they don’t end-up manually revisiting all SKUs all the time. However, the downside of a local optimization strategy as offered by DDMRP is that the system-wide outcomes, when measured in monetary amounts, are opaque. For example, DDMRP offers no control to the management to adjust a systemic tradeoff such as resiliency vs growth5 when considering the risk of an abrupt loss of a large and growing client who had been ordering diverse products in diverse quantities so far.
It’s not possible to have both local and global transparency: either decisions are locally optimized (like in DDMRP) with simple numerical recipes, in which case, there is no control and no transparency on what happens at the system level; or decisions are optimized globally, in which case all decisions tend to be numerically entangled, as it happens with QSCM, complicating any attempt at achieving transparency while picking one decision in isolation.
The numerical recipes put forward in DDMRP are straightforward and accessible with little or no technical background required. In contrast, Quantitative Supply Chains rely on Supply Chain Scientists which comes with a rather demanding skill set combining both business acumen and programming skills. ↩︎
Numerically speaking, DDMRP follows a two-stage process: first, establish the decoupling points; second, trigger flows based on numerical prioritizations. If full automation was a goal of DDRMP, then decoupling points would be automatically computed. However, if decoupling points are automatically computed, there is no need to pay any attention to decoupling points, as those would only be a transient state of the overall computation. DDMRP reifies its decoupling points precisely because those decoupling points are not the strict result of a numerical recipe. ↩︎
One way to make the supply chain more resilient against the loss of a large client consists of lowering the stock purely driven by the orders of this one client. However, if this already large client is still growing, lowering stocks will put future growth at risk. In this situation, there is a fundamental tradeoff between favoring resilience or favoring growth. This tradeoff has ramifications on almost every single supply chain decision. ↩︎