Beyond Lean Supply Chains: Optionality, Uncertainty, and the Economics of Decisions
For most of the last two decades, I have lived with a small irritation: the way we talk about supply chains and the way our supply chains actually behave rarely line up. We celebrate end‑to‑end visibility, flawless coordination, and elegant diagrams of “value streams.” Then Monday morning arrives, and we are back to firefighting, spreadsheet rituals, and hurried calls to suppliers who were supposed to be “just in time.”
Early in my career I tried to fix this irritation by following the mainstream playbook. I read the standard forecasting texts, implemented safety-stock formulas, and experimented with lean-inspired programs. The intellectual framework was tidy; the economic results were not. The gap between the promise and the ledger is what ultimately pushed me to write Introduction to Supply Chain, where I attempted to put into one coherent argument how I now look at the field and why that view diverges sharply from the usual “lean supply chain” narrative.
In that book I define supply chain as “mastery of options under variability in managing the flow of physical goods.” The wording is dense, but the idea is simple. Every decision you take—raising a purchase order, setting a price, allocating a container, scheduling a truck—commits scarce resources and excludes alternative uses. The more uncertain your future, the more those unused alternatives, those options, are worth. Supply chain, in my view, is the discipline of creating, valuing, and exercising those options so that, over time, the company earns more for the risks it takes.
Lean thinking comes from a different tradition. Since the 1990s it has been summarized in five principles: define value, map the value stream, create flow, let the customer pull, and pursue perfection. The north star is the systematic elimination of waste: excess inventory, unnecessary motion, waiting time, defects, and every other activity that does not directly create customer value. Just‑in‑time (JIT) production, with its promise of “only what is needed, when it’s needed, and in the quantity needed,” became the emblem of that philosophy when applied to supply chains.
At first glance my work appears compatible with this lean picture. I also care about flow. I also dislike waste. I also think shorter lead times and better reliability are powerful advantages. Yet beneath this superficial harmony there are deep disagreements, both about what supply chain is and about what it should be trying to achieve. The rest of this essay is my attempt to spell out those disagreements clearly.
Where lean and I genuinely agree
It is important to start with common ground, because I do not consider lean a villain in the story.
If you walk a factory where basic lean principles have been applied with care, the difference is visible. Workstations are easier to understand. Material moves with fewer detours. Problems surface more quickly because the process does not have enough slack to hide them forever. It is hard to argue with any approach that makes defects rarer, rework smaller, and cycle times shorter. Lean’s insistence on going to the shop floor, looking at the real flow of work, and involving the people who actually move the materials is one of the healthiest habits our profession has acquired.
If your current operation is chaotic, a well-run lean initiative typically improves both service and cost. You will reduce obvious waste long before touching clever mathematics or exotic software. In that sense, lean is an excellent corrective to sloppiness.
My disagreements begin when lean is promoted from “excellent corrective” to “governing philosophy” for supply chains, particularly once we move beyond the four walls of a plant and into networks spanning oceans, categories, and business models.
Supply chain as economics, not housekeeping
Lean supply chain programs usually present their goal as some mix of lower inventories, shorter lead times, fewer defects, and smoother flows. Waste is the enemy; efficiency is the hero. Warehousing, transportation, and inventory policies are assessed through this lens: less is almost always assumed to be better.
In my experience, this is the wrong starting point. The starting point should be: What risk‑adjusted economic return are we getting from our scarce resources—cash, capacity, time, and goodwill—once we take all the decisions labeled “supply chain” together?
Consider inventory. From a lean perspective, inventory is classed as waste: money tied up that does not actively create value. The ideal is to shrink it as far as possible without hurting service. From my perspective, inventory is a financial option. When you hold stock, you have bought the right to serve future demand quickly. That right has a cost—the carrying cost of inventory—but also a potential payoff: higher sales, fewer lost opportunities, better negotiation power with customers.
Sometimes the option is overpriced and should be trimmed; sometimes it is a bargain and should be bought in size. There are many categories where carrying more stock than a lean target would recommend is economically sound. A retailer facing a sharp, weather‑driven spring spike in a category like women’s razors may be better off ordering early, living with several quiet months of stock, and entering the hot weekend with shelves full, rather than trying to be “just in time” and discovering—along with every competitor—that replenishment capacity is suddenly scarce. The monetary pain of a stockout during the spike dwarfs the carrying cost of extra inventory in March.
The same reasoning applies to transport capacity, manufacturing flexibility, dual sourcing, and almost any other buffer or option. What looks like waste from a narrow process view often turns out to be cheap insurance—or even a profitable bet—once you price the underlying uncertainty correctly.
So I do not ask whether a supply chain is lean. I ask whether it is profitable for the risk it takes. Lean tools can help, but they are not an objective function.
Two ways of imagining the future
A second fault line between my view and lean has to do with how we imagine the future.
Lean’s great achievements took place inside factories, where sources of variability are mostly physical and can be driven down through better engineering. If you can stabilize machine behavior, standardize work, and smooth production schedules, the future inside the plant becomes much more predictable. The lean stance is that variability is a problem to be eliminated, or at least aggressively reduced, because it interrupts flow and creates waste.
Supply chains do not enjoy that luxury. Every meaningful decision in a supply chain reaches into a future where customer behavior, competitive moves, macroeconomic shocks, and regulatory shifts are all in play. You can and should stabilize your own processes, but you cannot legislate away demand spikes, fashion fads, pandemics, port congestion, or sudden currency swings. Those variations are not bugs in the system; they are the system.
For that reason, I find it dangerous to think of the future as something we design and then impose through discipline—what I sometimes call a “plan‑first” mindset. In that mindset you forecast a single demand number, build a plan around it, and then measure success by adherence to the plan. The plan becomes something to defend against messy reality.
I prefer to think of the future as a rugged landscape we navigate with a portfolio of options. There is no single true forecast to uncover, only a distribution of possibilities. The job of supply chain is to position the company so that it does acceptably well across a range of plausible futures, rather than maximally well in a single, fragile scenario. That, in turn, requires probabilistic forecasts and decision rules that explicitly account for upside and downside under uncertainty, not just point forecasts and local safety stocks.
Lean is extremely effective at improving the mechanics of execution once you have decided what should happen. It says relatively little about how to choose between competing futures when uncertainty is high and most of the leverage lies in those choices.
Pricing and assortment are supply chain decisions
Mainstream supply chain teams are often told to focus on physical flows—purchasing, manufacturing, transportation, warehousing—while prices, promotions, and assortment are labeled “commercial” matters and assigned to marketing or sales.
If you adopt my definition, that division no longer makes sense. A decision that changes what moves, where, when, and in what quantity belongs to supply chain, regardless of whether the lever sits in the ERP or in a pricing system. Setting a price is as much a supply chain decision as booking a truck: the price shapes demand, and demand shapes every downstream movement of goods.
In the book I argue that exiling pricing and assortment from supply chain has crippled the field. The result is a discipline that optimizes flows around given commercial decisions instead of co‑designing commercial and operational choices as a single economic problem. We end up with elegant pull systems, perfectly levelled, for products whose price and range decisions were wrong before the first pallet left the warehouse.
Lean thinking, as commonly practiced in supply chains, tends to preserve this split. It optimizes how a given product, at a given price, moves through the network. My own work treats price, range, and placement as first‑class levers in the same optimization.
From workshops to decision engines
Another contrast concerns where we expect intelligence to live.
Lean’s cultural center of gravity is the workshop: mapping the value stream on a wall, walking the gemba, running kaizen events, teaching people to see waste, and empowering teams to fix problems. The tools are deliberately simple and visible. The assumption is that, with the right habits, frontline experts will make good choices and continuously improve the system.
I share the respect for operators and for direct observation. At Lokad we have never improved a supply chain without extensive conversations with people who actually run it. However, I do not believe that workshops and visual boards can remain the primary vehicle for decision‑making in modern supply chains.
The reason is scale and combinatorics. A large retailer or manufacturer faces millions of tiny resource‑allocation problems every day: which SKUs to buy, in which quantities, to which locations, at which prices, through which routes, using which capacities, under which constraints. Human working memory and attention are simply not designed to evaluate millions of options with delayed and uncertain payoffs. Computers, on the other hand, are remarkably good at exactly that.
In the book I argue that we should treat supply chain primarily as a software and economics problem: design specialized decision engines that read the company’s records, use probabilistic models to assess options, arbitrate between those options based on expected, risk‑adjusted return, and then write decisions—orders, allocations, prices—back into the system, unattended, day after day. The role of the human planner becomes that of a designer, supervisor, and critic of these engines, not their manual backup.
Lean practitioners sometimes object that this breaks the “respect for people” pillar. I see it differently. Offloading the combinatorial drudgery to software is the highest form of respect for human cognition: it acknowledges what we are good at—understanding context, questioning assumptions, spotting missing constraints—and what we are not. Asking planners to act as live solvers for thousands of SKUs, every day, is not respect; it is wishful thinking.
Just‑in‑time and the price of fragility
Nothing illustrates the divergence between my view and the lean supply chain orthodoxy better than JIT.
In its original manufacturing context, JIT was a brilliant idea: reduce inventories between process steps, expose problems, and drive continuous improvement. When your demand is stable, your suppliers are reliable, and your lead times are short, JIT creates both efficiency and learning.
The trouble starts when JIT is treated as a moral ideal rather than as a context‑dependent tactic. In extended supply chains, many of the assumptions that made JIT work in a car plant do not hold. Demand may be seasonal or promotional; lead times may be long and volatile; suppliers may be few and politically fragile; transportation lanes may be congested or vulnerable to disruption. Under those conditions, pushing inventory “back upstream” in the name of leanness does not eliminate risk; it merely changes its shape and location.
From my option‑centric perspective, the question is never “How do I get closer to zero inventory?” It is “Given the volatility of demand and supply, and the economics of stockouts versus carrying cost, how much inventory should I hold, where, and in which form, to maximize long‑run, risk‑adjusted return?” The answer will sometimes look very lean indeed. At other times it will look positively extravagant to a JIT purist.
Recent global events have made this painfully clear. Firms that had been praised for running “beautifully lean” supply chains discovered that the buffers they had shaved away were exactly the ones they needed when borders closed, ports jammed, and demand patterns inverted. A philosophy that treats all buffers as waste is poorly equipped to distinguish between prudent resilience and actual excess.
What I keep from lean, and what I discard
So where does this leave us?
I keep lean’s insistence that reality, not theory, is the ultimate judge. If your supposedly clever algorithm produces purchase orders that warehouse managers laugh at, the algorithm is wrong. If a process hides defects and delays under layers of inventory and batching, the process is wrong. Lean’s habit of walking the flow, talking to operators, and making problems visible is one of the best safeguards against ivory‑tower modeling.
I also keep lean’s intuition that smoother, more reliable processes tend to be better. Shorter and more predictable lead times widen your option set; they make better decisions possible. From an economic standpoint, they are worth pursuing.
What I discard is the idea that waste elimination and inventory reduction, taken in isolation, define success. I discard the notion that variability is always an enemy rather than a source of both risk and opportunity. I discard the assumption that daily decisions should stay in human heads and spreadsheets instead of being encoded, tested, and refined in software that can act consistently at scale.
In my practice, lean is one source of good heuristics and one source of cautionary tales. The heuristics become candidates for encoding in decision engines, then face the same question as any other policy: over time, under uncertainty, do they improve the firm’s risk‑adjusted economic performance? If they do, they stay. If they do not, they are retired, no matter how virtuous they might look on a poster.
Seen from this angle, my work is not anti‑lean. It is orthogonal. Lean asks, “Where is the waste in this process, and how can we remove it?” I ask, “Given the options available and the uncertainties ahead, how should we allocate our scarce resources so that, five years from now, we are glad we did?”
The modern supply chain deserves both questions. But if I must choose a single organizing principle for the field, I will always choose the second.
