A Reflection on Lora Cecere’s Work
Over the last few decades, “supply chain” has been drafted into almost every corporate conversation. It now covers everything from container shortages to board scorecards, from factory automation to AI pilots. Yet when you look at the financial statements of most large companies, the picture is sobering: despite all the noise, many still struggle to turn working capital, capacity, and organizational complexity into superior economic returns.
In that landscape, I have long appreciated the persistence of Lora Cecere. Through Supply Chain Insights and her work on market-driven, outside-in value networks, she has tried to ground the discussion in hard numbers: growth, operating margin, inventory turns, and returns on capital over long periods of time, compared across peers. Her “Supply Chains to Admire” benchmark is built exactly around this idea of identifying firms that manage to balance growth, profitability, and resilience better than their competitors.
In a recent essay titled Supply Chain as Economic Bets in a Market-Driven World, I argued that the daily decisions to buy, make, move, and price are, in effect, a very large portfolio of economic bets under uncertainty. Some of those bets turn out well, some fail, and the aggregate of millions of such decisions is what eventually appears in the income statement and balance sheet. This follow-up looks at the same reality from another angle: how this bets perspective relates to Cecere’s own definition of an effective supply chain, and where our respective views diverge and complement one another.
I developed my own perspective in more systematic form in my book Introduction to Supply Chain, especially in Chapter 1 and Chapter 4. There I describe supply chain as the craft of steering the flow of physical goods by allocating scarce resources—cash, capacity, inventory, time, attention—so that, over time, the firm earns a superior risk-adjusted return on those resources. Cecere, by contrast, starts from a network view: she defines an effective supply chain as bi-directional decision making on flows from the customer’s customer to the supplier’s supplier, where value shows up as sustained improvements in market capitalization through a balanced scorecard of growth, operating margins, inventory turns, and return on capital. Her empirical work, often in collaboration with academics, takes that scorecard seriously.
We are looking at the same animal, but from different vantage points.
Two vantage points on the same discipline
Cecere likes to talk about market-driven value networks. Rather than starting from internal transactions—purchase orders, production orders, transfer orders—she insists that an effective supply chain must see the market itself: consumption at the shelf, channel inventory, promotion calendars, social and macro signals, supplier constraints, and disruptions in logistics capacity. Planning, in this view, should be outside-in: it starts with what is happening from the customer’s customer backward and then orchestrates responses across sourcing, manufacturing, logistics, and commercial decisions.
On top of that, she overlays the long-term financial lens. Her “Supply Chains to Admire” methodology evaluates companies against their peer groups on three dimensions: improvement over time, current performance, and market value. The scorecard itself is deliberately simple: revenue growth, operating margin, inventory turns, and return on invested capital. When a company manages to improve on that set of metrics faster than two-thirds of its peers and reach a high absolute level, she calls it a leader. In other words, an effective supply chain is one that, over many years, leaves its competitors behind on this multidimensional frontier.
My vantage point is closer to the ground. I start with the moment of decision. Should we buy one more pallet of this SKU for that warehouse, expecting it to arrive six weeks from now? Should we bring a production run forward, delay it, or cancel it? Should we lower a price in a particular channel tomorrow, or hold firm? Each of these decisions consumes something scarce—cash, capacity, shelf space, goodwill—and creates exposure to a range of possible futures. Most of them are small on their own, but in aggregate they define the firm’s risk profile and its economic outcome.
From that angle, a supply chain is essentially a mechanism for turning uncertainty into decisions and decisions into financial consequences. I am less interested in the beauty of a process map and more interested in whether the next decision—and the one after that, repeated thousands of times per day—makes economic sense given what we know and what we do not know.
Cecere’s vantage point is from above: she looks at how a firm moves relative to its peers on the scorecard over ten-year horizons. Mine is from inside: I look at whether the marginal decision, given the uncertainty and the trade-offs, is a good use of scarce resources. These perspectives are not in opposition. A firm cannot stay on the frontier for long if its millions of daily decisions are systematically poor. Conversely, a firm can have elegant decision methods that are badly aligned with its strategy and never climb the frontier at all.
Uncertainty, signals, and the illusion of the single plan
Where Cecere and I are perhaps closest is in our irritation with traditional planning.
Her criticism targets what she calls inside-out planning. Most companies still treat their own orders and shipments as if they were clean representations of demand. Yet orders are filtered and distorted by promotions, allocation rules, stockouts upstream, and internal politics. The resulting plans are late, biased, and often blind to what is actually happening in the market. Cecere’s answer is to redesign planning as a set of outside-in processes: demand sensing from market data, demand translation into what the network can do, and then demand orchestration across functions.
In her recent Outside-In Planning Handbook, she reports on several years of pilots that reposition planning around external signals and market-to-market flows rather than around internal ERP transactions. The reported benefits include shorter time to know about shifts in demand and supply, and a material reduction in bullwhip effects, as companies stop amplifying noise through their own processes.
My own irritation is with the way forecasts and plans are typically treated. In many organizations, the forecast is a single number per period and per item: the most likely quantity to be sold. That number becomes the anchor for production plans, purchasing plans, transfer plans, capacity reservations, and so on. Deviations are treated as errors to be explained away after the fact.
This approach discards the very information that matters most: not the central estimate, but the range of plausible futures and their probabilities. For any given item, the questions that matter are: What is the chance that sales next month will be half the usual level? Twice the usual level? Three times? What do the tails look like? The answers are rarely symmetric, and they are rarely well behaved.
Once you accept this, the idea of a single consensus plan becomes less compelling. Instead of asking “What is the forecast?” and then negotiating a plan around it, we should be asking “Given this distribution of possible futures and these financial consequences of stockouts, excess, and delay, what decisions make sense?” The same distribution can justify very different inventory or production choices depending on the margin structure, the availability of substitutes, and the lead times involved.
In this sense, Cecere and I are both rebelling against the same illusion: that the future can be reduced to a single line in a spreadsheet, and that the main task of planning is to reconcile departmental views until everyone agrees on that line. Her solution is to challenge where the information comes from and how it flows—outside-in rather than inside-out. Mine is to challenge the representation of uncertainty itself and to insist that decisions be computed against full probability distributions and explicit financial trade-offs.
In a healthy practice, these two concerns meet. You want better signals from the market and from your network, and you want models that treat uncertainty honestly. Outside-in flows become the raw material for probabilistic, economically grounded decision making.
Technology, roles, and what planners should actually do
Cecere often points out that the technology stack most companies rely on was built first and foremost for transactional efficiency. ERP systems are very good at recording orders, shipments, invoices, and receipts; they are much less good at helping you decide what to do next. Adding more user interfaces on top of such a stack does not fix the underlying problem. She argues for rethinking architecture around flows: integrating external demand and supply data, building better taxonomies, improving near real-time visibility, and giving business users more flexible, self-service analytical tools.
In recent years she has extended that argument into an agenda for native-AI supply chains. The idea is not simply to bolt machine-learning models onto existing processes, but to redesign data foundations, semantic layers, and workbenches in such a way that new forms of AI can actually help orchestrate flows from customer’s customer to supplier’s supplier. At the same time, she questions the traditional planner role, noting the mismatch between responsibility and authority: planners carry the blame for service failures and excess stock yet lack the power to change pricing, promotion, or product strategy. In her vision, planners become orchestrators, operating in cross-functional, outside-in processes rather than as clerks feeding disconnected systems.
I agree that the inherited technology stack is a major obstacle. My emphasis, however, lands elsewhere. For me, the core missing capability is not another layer of integration or another dashboard; it is a decision engine.
By this I mean a piece of software that, on a recurring basis, takes all relevant data and constraints, applies an explicit economic model of costs and opportunities, and then proposes or executes concrete actions: which purchase orders to place, which production orders to schedule, which stock transfers to organize, which prices to adjust. This engine must be programmable by people who understand the business, auditable in the sense that it can explain past decisions, and fast enough to handle large decision volumes in the time windows imposed by physical lead times. It must also be capable of working with distributions rather than point forecasts.
Outside-in architectures are valuable because they feed such an engine with richer, timelier information. But without the engine, it is easy to end up with very sophisticated reporting and very traditional human decision making. You see more, but you still decide as you did before: in meetings, with spreadsheets, under time pressure.
The same contrast appears when we talk about organization and governance. Cecere spends a lot of time on sales and operations planning and similar cross-functional forums. Her maturity models describe how S&OP can evolve from a basic capacity check into a profit-focused, demand-driven, and eventually market-driven orchestration process. In her stories, S&OP is where trade-offs across functions are surfaced and resolved, and where the outside-in perspective finds a human home.
I share her frustration with functional silos and with the way local metrics—service levels here, utilization there, forecast accuracy somewhere else—can destroy value at system level. Where I differ is in how central S&OP, as a planning meeting, should remain once technology is used to its full potential. In my view, if we build robust decision engines, most of the operational planning work currently done in S&OP should disappear into software. What should remain is governance.
In that world, the main task of executives and cross-functional teams is not to edit quantities in a spreadsheet, but to tune the rules of the game: the relative cost we assign to stockouts versus excess for each family of products, the value we place on lead-time reductions versus capacity utilization, the limits we impose on risk-taking in specific markets, the constraints we accept in contracts. They should review how the decision engine has behaved, where it has created or destroyed value, and then adjust the economic parameters and assumptions accordingly.
This is a shift from making the plan to governing the system. The daily plan becomes the emergent result of many small, automated decisions; the human focus moves to ensuring that the economic logic of those decisions matches the company’s strategy and risk appetite.
Here again, Cecere’s and my perspectives are not incompatible but differ in emphasis. She focuses on architecture and cross-functional processes; I focus on the economic and computational core that those processes should steer.
Toward a practical synthesis
If you are responsible for a sizeable supply chain, it can be tempting to treat such differences as a choice between schools of thought. Should you follow Cecere and invest in outside-in processes, market-driven metrics, and reimagined S&OP? Or should you follow my bets under uncertainty framing and invest in probabilistic modeling and decision engines?
I would urge you not to frame it this way.
Cecere’s work is most powerful at the strategic and diagnostic level. It forces uncomfortable questions. Are we, in fact, improving faster than our peers on growth, margin, inventory turns, and return on capital, or are we congratulating ourselves on internal KPIs that do not show up in shareholder value? Are our processes still driven by the inertia of ERP transactions and departmental views, or do we genuinely start from the market and work backward? When she says that a supply chain cannot be built through inside-out processes, that it needs to be market-driven and built from the outside-in, she is summarizing a decade of data that shows how many sectors have actually regressed.
My own work is more operational and computational. It lives in the uncomfortable question: given what we know, and what we do not know, about demand and supply, and given the financial consequences of different kinds of mistakes, are the decisions we take every day a good use of scarce resources? If not, can we redesign the logic of those decisions, implement it in software, and let that software handle most of the routine work?
Put together, a practical synthesis might look like this.
At the board and executive level, you adopt a scorecard not very different from Cecere’s. You watch your firm’s trajectory against peers on growth, operating margins, inventory turns, and returns on capital, and you treat that trajectory as an external check on whether your supply chain is actually delivering economic value over time. You accept that efficient transactions and pretty dashboards are not the same as excellence.
At the architectural level, you organize data and processes outside-in. You invest in seeing real consumption, real constraints, and real variability as early as possible, and you design processes that move information bi-directionally across your network rather than linearly within functions.
At the decision level, you gradually replace handcrafted plans and local heuristics with explicit models of uncertainty and value, embedded in software that can take millions of small decisions consistently, explain itself, and be improved over time. You judge those models not by how elegant they look in a slide deck, but by how they affect the firm’s risk-adjusted economic returns.
Finally, at the organizational level, you stop asking planners to do impossible jobs with incomplete authority. You retrain some as designers and custodians of the decision logic itself, and you retrain others as orchestrators who oversee exceptions, structural changes, and cross-functional trade-offs. Governance forums evolve from plan-building rituals into review sessions for a living, automated system.
From that perspective, the apparent disagreement between my views and Cecere’s is mostly a disagreement about where to put the spotlight. She insists we look up and out—at markets, networks, and long-term comparative performance. I insist we look down and in—at the quality of the next decision, and the one after that, as shaped by software.
Both lights are needed. Without the outside-in, market-driven lens, it is easy to optimize local decisions and still lose the competitive race. Without the economic and probabilistic lens, it is easy to build beautiful architectures that do not actually improve the bets the company is placing every day.
If I had to compress my own position into a single line, it would be this:
Supply chain, properly practiced, is an economic discipline that uses software to place better bets under uncertainty.
Everything else—architectures, processes, dashboards, even scorecards—should be judged by whether they help or hinder that very specific craft.
