From Factory Planning System to Decision Engine
Most manufacturing companies who knock on Lokad’s door introduce themselves the same way:
“We need a better planning system.”
Sometimes it is “a modern MRP,” sometimes “an end‑to‑end APS,” but the expectation is consistent: a piece of software that will finally produce a stable, reliable plan for the plant and its suppliers.
You are not wrong to ask for this. If you run a factory, you are inundated with dates, quantities, and constraints. You want a plan you can trust so that machines keep running, customers stay happy, and capital is not wasted.
In my book Introduction to Supply Chain, I tried to give operators a compact language to think about these questions from first principles, with today’s computing reality in mind. What follows is the shorter, manufacturing‑focused version of that story, written for companies meeting Lokad for the first time.
What you think you are buying vs. what you are actually buying
When a manufacturer comes to Lokad, the implicit request is usually: “Give us a system that computes a better plan than the one we get today from our MRP and spreadsheets.”
Under that view, the software’s job is to take your master production schedule, explode your bills of material, time‑phase the requirements, and generate planned orders. This is the classic material requirements planning picture: using BOMs, inventory records, and a master schedule to calculate what is needed, in what quantity, and when.
Lokad does ingest that same data, but the purpose is different.
We are not trying to produce the plan and then protect it at all costs. We are trying to compute, every day, a ranked list of decisions—purchase, production, allocation, sometimes even pricing—each one evaluated for its expected financial impact under many possible futures. Our software is less an electronic Gantt chart and more a decision engine.
The moment you see it that way, the entire conversation changes. It is no longer about “which plan is correct,” but about “which next action buys us the most future profit for the risk we are taking.”
The plant and the supply chain are not playing the same game
Inside your factory walls, variability is the enemy. You pay engineers and lean specialists to drive out unwanted fluctuations: tighter process control, shorter and more predictable set‑ups, fewer breakdowns, better maintenance. The goal is a process that behaves the same way every time, within tight tolerances.
Outside your factory walls, in the broader supply chain, the world refuses to be tamed like that. Demand moves. Suppliers slip. Ports clog. Competitors change prices. Weather misbehaves. No amount of capital will buy you a deterministic future.
This difference matters.
If you treat the supply chain like a big machine waiting for a perfect plan, you end up disappointed. Plans look clean on the projector, then collide with reality and fragment into manual overrides, expediting, and firefighting. If, instead, you accept that the future will remain uncertain, the problem becomes: how do we make good bets, repeatedly, in the presence of that uncertainty?
Your plant is where atoms are transformed; your supply chain is where wagers are made about where those atoms should go, when, and in what quantity. Both need software support, but they do not need the same kind of software.
What MRP gets right—and where it stops
Classical MRP has done an enormous amount of good for manufacturing over the last fifty years. It forces companies to make their bills of material explicit. It links demand for finished products to the components that must be available. It makes material constraints visible instead of implicit.
However, traditional MRP makes several assumptions that were reasonable in the 1970s, but are increasingly restrictive today:
It presumes a single demand plan, usually expressed as a master production schedule, and treats that plan as the reference truth. It assumes lead times are known and behave more or less as averages suggest. It works in coarse time buckets—often weekly—with planning runs scheduled on a similar rhythm. It calculates material requirements deterministically and leaves it to humans, or to separate modules, to worry about capacity conflicts and economic trade‑offs.
In textbooks and certification materials, MRP is rightly described as “a set of techniques” driven by BOMs, inventory status, and the master schedule. In practice, this is exactly how most plants still operate: MRP as a sophisticated calculator, planners as interpreters and negotiators of its output.
From Lokad’s vantage point, this is where the mainstream view stops—at the point where the system has computed feasible orders, but has not truly chosen among them.
How Lokad looks at your factory
When Lokad connects to a manufacturing company, we see three layers overlaid on the same physical reality: your flow of goods, your flow of information, and your flow of money.
The flow of goods is familiar: raw materials entering, WIP moving through lines and work centers, finished goods leaving toward warehouses or customers. The flow of information is the digital shadow of that reality: the tables and records in your ERP, MRP, WMS, MES. The flow of money is the one that usually gets the least explicit attention in planning, despite being the one the board ultimately cares about.
Lokad’s job is to reconcile those three.
We start from the same historical data your planning team already has: orders, stock movements, bills of material, lead times, capacities. But instead of forecasting a single quantity for each SKU in each week, we forecast a range of possible outcomes with associated probabilities for demand, lead time, and sometimes yield.
Then we fold your economics into that picture. Holding costs, stock‑out penalties, overtime and changeover costs, minimum order quantities, container or truck constraints, service commitments to key customers; all of these become explicit ingredients in a numerical recipe whose output is a set of suggested decisions.
In other words, instead of saying “the forecast for this SKU next week is 120 units,” we say “here is the distribution of possible demand and supply outcomes; given that distribution and your economics, producing 130 units today and 70 units in three days is, on balance, the best bet.”
From plans to daily bets
The practical consequence of this view is that Lokad runs every day, not every month or every planning cycle. Historical data and current positions are extracted from your existing systems. Forecasting and optimization are executed as a single, unified computation in the cloud. The result is a series of decision lists: purchase orders to raise, production batches to launch, transfers to execute.
These lists are ranked. At the top, you see the decisions with the highest expected financial impact—typically, actions that prevent catastrophic line stoppages, high‑margin stock‑outs, or costly emergency transport. Further down, you see more marginal improvements: small reductions in excess stock, gentle rebalancing between sites, fine‑tuning of batches.
Your planners and production managers no longer have to reconcile an abstract plan with a messy reality. Instead, they start from a prioritized backlog of concrete actions that already takes into account uncertainty, constraints, and economics. Their time is spent examining the logic, challenging assumptions, and refining the model, rather than manually editing hundreds of orders one by one.
The notion of a “frozen” plan becomes less central. What matters is that, on any given day, the decisions you execute are the best ones you can make with the information and resources you have.
Lokad and MRP: not competitors, but different layers
A question I hear regularly is: “Are you replacing our MRP?”
The short answer is no.
Your ERP and MRP remain the systems of record. They continue to own master data, transaction processing, shop‑floor execution, and financial posting. Lokad operates as an added layer on top of those systems, in SaaS mode, producing optimized decisions that are then written back as purchase orders, production orders, or planning inputs.
In that sense, MRP and Lokad are complementary. MRP knows your bills of material, routings, and basic availability constraints. Lokad knows how to turn large amounts of data into economically sensible bets under uncertainty. MRP provides structure; Lokad provides quantitative judgment.
If you expect Lokad to be “a better MRP,” you will miss this distinction and judge us on the wrong criteria. If you expect Lokad to be the brain that tells your existing systems which actions to take, with money as the measuring stick, then you are much closer to how we actually work.
What to bring to a first conversation
If you are a manufacturing company considering Lokad, you do not need to arrive with perfect data or a fully formed vision. What you do need is clarity on a few simple questions.
First, how do you make money, concretely? Which products, lines, or customers matter most, not in terms of volume on a dashboard, but in terms of contribution margin and strategic importance. Second, where does it really hurt today? Chronic line stoppages for lack of cheap parts, excess inventory in some nodes and shortages in others, erratic service on critical items—these pain points are often where probabilistic decision‑making pays off fastest.
Third, are you willing to let go of the idea that there exists a single “correct” plan? If you can accept that the future will remain uncertain, and that good decisions are good because they handle a spread of scenarios gracefully, then the rest follows naturally.
In closing
From the outside, Lokad can look like yet another planning tool. Under the hood, and in daily practice, it is something else: a way to turn your plant and its supply chain into a collection of repeatable, financially sound bets under uncertainty.
MRP taught the industry to make material requirements explicit and to connect demand to components. For that, it deserves its place in history. But it stops at the point where the world is assumed to follow the plan.
The reality you live with in your factory does not. That gap is where Lokad operates.
If you come to us expecting a prettier Gantt chart, we will disappoint you. If you come to us looking for a way to improve the quality of your decisions—what to buy, what to produce, what to ship, and in what quantity—day after day, then we are having the right conversation.
