We Knew This Was Possible. Why Wasn’t It In the Plan?

There’s one meeting almost every planning leader knows by heart. Something slipped. A supplier missed a window, a batch failed a quality hold, or a promotion pulled harder than forecast. The room gathers to walk back through what happened, and somewhere in that walk-back someone says the quiet part out loud: “We knew this was possible; we just didn’t plan for it.”

It’s worth sitting with that sentence, because it points to something structural rather than personal. The information about what could go wrong was usually available before the disruption hit. Maybe a supplier’s on-time performance had been drifting for months, or a production line’s yield variability was well documented. The plan simply was not built to carry that information forward. It captured a single expected outcome and treated everything else as an exception to handle later, in a meeting, after the fact.

Planning exists for one reason: to prepare the supply chain to execute. Not just to produce a document that looks confident on a screen, but to give the people running plants, moving inventory, and scheduling changeovers a plan they can actually act on when conditions shift.

A plan that only works if nothing goes wrong isn’t really a plan. It’s a forecast wearing a plan’s clothes.

What It Looks Like to Plan This Way

Consider a raw material that has some supply variability, the kind almost every process manufacturer deals with somewhere in its network. The traditional approach sets a safety stock number based on average lead time and average demand, then hopes the averages hold. They rarely do. Lead times spike around holidays, port congestion, or a single-source supplier’s capacity constraints. Demand has its own volatility layered on top.

But there’s a different planning approach: Quantify risk as the plan is built. Position that raw material against a percentile of historical and anticipated lead time variability, not just the average, so the safety stock reflects what has actually happened in that supply line rather than what would happen in an ideal one. If the material has historically taken twelve days but occasionally stretched to twenty, the plan can be built around that distribution instead of the midpoint. The planner still decides where to set the threshold and how much working capital to commit to that cushion. The plan simply gives them the fuller picture to decide from, instead of a single number that quietly assumes the best case.

The same logic applies to changeover sequencing on a shared line. A schedule that treats changeover time as fixed will look clean until the actual sequence runs longer because of an allergen flush or a viscosity change between products. A plan that anticipates that variability (and prepares for a range of likely changeover durations) will give the scheduler room to adapt without the whole day’s plan collapsing the moment reality deviates from the average.

None of this requires guessing at the future. It requires treating the future as a range, instead of a point, and building that range into the plan from the start.

The Cost Side of The Story

Planning for a range isn’t free. Carrying safety stock against a wider distribution ties up working capital. Building schedule flexibility can mean underutilizing a line on paper so it has room to absorb variability in practice. Resilience always has a cost, and pretending otherwise is how planning teams end up with plans that look efficient right up until they are not.

The honest version of this work is not eliminating that cost. It’s making the cost visible and deliberate, so leaders can decide how much certainty they want to pay for, rather than discovering the price after the disruption has already happened.

Quantifying risk in this way doesn’t replace judgement. It gives planners a clearer, more honest set of options to apply their judgment to, so they can make confident decisions. Planners have spent years learning how a particular supplier behaves under pressure, or how a particular line handles a rushed changeover. A plan built to quantify risk simply gives that knowledge somewhere to live, instead of letting it live in someone’s head until a post-mortem draws it back out.

Where This Leaves Planning Teams

Most planning tools were built to answer a narrower question: what does the plan look like if everything goes as expected? That question is useful, but it is not the one that matters most for teams running complex supply chains. The more useful question is what the plan looks like across the range of things that could reasonably happen, and what the team needs to have ready for each scenario.

This is the thinking behind Arkieva’s approach to planning. Arkieva is an AI-native platform built for process manufacturers who need their plans to hold up under real variability. Risk won’t disappear. But the big shift happens when the team deals with it, from the post-mortem back to the plan itself.

We covered this idea in more depth in a recent session on quantifying risk in process manufacturing planning. Watch the recording here.