When Disruption Becomes Normal

The real test of a planning system is what happens when the plan breaks

For decades, supply chain disruption was largely treated as an exception. A supplier failed. A factory lost capacity. A port closed. A major customer unexpectedly changed demand. The normal response was familiar: extract the data, assemble the planners, build a spreadsheet and work out what to do.

That approach may have been reasonable when major disruptions were unusual. It is increasingly difficult to defend today.

Supply chains now operate in an environment of persistent volatility. Geopolitical change, supplier instability, transport disruption, changing costs and increasingly complex global networks mean that organisations need to plan for uncertainty as part of everyday operations.

This changes the question we should ask of planning technology. It is no longer enough to ask whether the system can create a good plan. We also need to ask what happens when reality makes that plan obsolete.

Why the spreadsheet war room still exists

Most large companies have invested heavily in ERP and planning technology. Yet a major disruption can still result in planners being pulled into a room to answer questions such as: what happens if a supplier becomes unavailable, can production move to another plant, how long can existing inventory support demand, which customers will be affected first, what happens if transport lead time doubles, and where will the next capacity constraint emerge?

These are fundamentally scenario-modelling questions.

Traditional systems often struggle because they were designed around defined processes, established networks and relatively static planning parameters. When an organisation needs to change the structure of the problem itself, Excel becomes attractive because it is flexible, immediate and disposable.

The spreadsheet is therefore not necessarily the problem. It is often evidence of a capability missing from the planning system.

Planning against the ERP rather than the real world

Another weakness becomes particularly apparent during disruption: many planning systems still rely heavily on static parameters held in ERP or master data.

A supplier lead time might be defined as 30 days. But actual deliveries may have been 24, 29, 41, 32, 46, 35 and 38 days. The planning engine may nevertheless continue to use 30 days for every calculation.

Likewise, static values may exist for batch sizes, transit times, supplier capacity, manufacturing yields, replenishment cycles, minimum order quantities and production run rates.

These values are necessary as reference data, but they should not automatically be treated as an accurate representation of current reality. When operating behaviour changes but planning assumptions remain static, a gap develops between what the ERP says should happen and what the supply chain actually does.

During disruption, that difference becomes critical.

Machine learning should improve the assumptions behind the plan

Much of the current discussion around AI in supply chain planning focuses on generative AI, agents and conversational interfaces. These technologies will certainly change how users interact with planning systems. But one of the most immediate opportunities for machine learning is less visible and arguably more important: improving the information on which planning decisions are based.

Instead of simply accepting that supplier lead time equals 30 days, machine learning can analyse operational history and recognise that expected lead time is currently 34 days, variability has increased, and this supplier typically deteriorates around Chinese New Year.

Similar approaches can be applied to demand behaviour, production performance, transport reliability, supplier performance and replenishment patterns.

The objective is not to surrender decisions to an algorithm. The organisation can still retain deterministic business rules, governance and human decision authority.

The more useful architecture is probabilistic intelligence underneath and disciplined decision logic above it. The planning model becomes more adaptive while the organisation remains in control.

From disruption visibility to disruption simulation

Many organisations have invested heavily in supply chain visibility. That has unquestionably improved the ability to identify problems. But knowing that a shipment is late or a supplier is at risk is only part of the problem.

The next question is: what should we do about it?

That requires simulation.

Imagine a supplier becomes unavailable for eight weeks. A modern planning environment should enable the business to model that disruption and immediately understand the consequences.

For inventory, where will shortages occur, how quickly will buffers be consumed, and where might excess stock exist elsewhere in the network?

For capacity, can production be moved, what additional constraints appear at another site, and which resources become overloaded?

For service, which customers or products are affected, what happens to order reliability, and can scarce inventory be prioritised differently?

For financial performance, what does each alternative cost, how does expedited transportation affect margin, and what is the impact on working capital?

For the network, can another supplier, plant, distribution centre or transport lane be used?

The user should then be able to change an assumption and recalculate. Move 30% of production to Plant B. Recalculate. Use Supplier C for six weeks. Recalculate. Increase inventory ahead of a known shutdown. Recalculate. Change priority shipments from ocean to air. Recalculate.

This is very different from simply generating an alert. Visibility identifies the disruption. Simulation enables the response.

Planning needs freedom from the transactional system

ERP remains essential. It should remain the trusted system of record for transactions, master data and execution. But a simulation environment needs considerably more freedom.

A planner should be able to change suppliers, sourcing percentages, lead times, production locations, capacities, routes, costs, batch sizes, inventory policies and demand assumptions without first changing the live ERP environment.

These are hypotheses. Some may never be implemented. A modern planning architecture therefore needs to be connected to ERP without being constrained by ERP.

ERP can remain the system of record while the planning environment increasingly becomes the system of decision. Only when an organisation chooses an alternative should the relevant decisions move towards execution.

The network matters more than the individual SKU

Disruption exposes a limitation of conventional planning. The real supply chain is not simply a collection of products and warehouses. It is a network: supplier, transport, manufacturing, inventory, distribution and customer.

Each element contains different costs, capacities, lead times, variability, risks and constraints. A change at one point can have consequences throughout the network.

This is why the historical separation between network design and operational planning is becoming increasingly difficult to maintain. Network design was traditionally a strategic exercise performed every few years. Operational planning happened daily or weekly.

But if a tariff changes, a supplier closes, a transport lane becomes unreliable or a plant loses capacity, network design suddenly becomes an operational problem. Organisations increasingly need to ask what happens if we change the network now, and receive an answer quickly enough to act.

Why visual planning matters

Conventional planning technology tends to be planner-centric. The planner understands the model, runs the calculations, interprets the output and gives the wider organisation an answer.

During a disruption, however, decisions often involve supply chain, procurement, manufacturing, logistics, finance, sales and senior management.

A visual representation of the supply chain can dramatically improve those conversations. Instead of interpreting rows of planning data, decision-makers can see which node is affected, which routes are changing, where inventory will move and where constraints will emerge.

They can then compare scenarios using common measures such as capacity, inventory, service and cost. The planning system becomes a collaborative decision environment rather than simply a specialist calculation engine.

Why the financial view matters

Supply chain decisions rarely have a single correct answer. Consider three possible responses to a disruption. Scenario A maintains existing sourcing and accepts reduced service. Scenario B moves production to an alternative facility. Scenario C uses another supplier and expedites transport.

Scenario C might provide the highest service level but at unacceptable cost. Scenario A might protect margin but sacrifice strategically important customers.

The planning environment therefore needs to translate operational alternatives into financial consequences. That should include procurement cost, production cost, transportation, inventory, working capital, potential lost revenue and margin.

The objective is not necessarily to allow an algorithm to select one mathematically perfect answer. It is to make the trade-offs transparent.

Carbon cost is becoming part of the planning equation

There is another emerging cost dimension that planning systems increasingly need to accommodate: carbon.

The introduction of ETS2 will extend carbon pricing into fuel combustion in areas including road transport, buildings and additional sectors outside the existing EU ETS framework. For supply chain planners, the important point is not simply regulatory compliance. It is that carbon-related cost will increasingly appear inside normal operational economics.

A transport lane may therefore have more than a conventional freight cost. It may also carry a carbon-related cost. A production location may look economically attractive under conventional costing but become less attractive once energy or carbon-related costs are considered.

A sourcing decision may have different economics depending on distance travelled, mode of transport, fuel type, energy consumption, production location and carbon intensity.

This means carbon can become another scenario variable. A disruption response such as changing supplier, switching transport mode, rerouting product or moving manufacturing can affect both conventional cost and carbon exposure.

That creates a need to compare service, inventory, capacity, financial cost and carbon cost within the same planning scenario.

At O8, ETS2-related planning is part of our development roadmap for Organic Planner. Our intention is to support carbon-related transport costs as an additional network cost line and, where appropriate, allow those costs to flow through the bill of materials into finished-SKU economics.

This means a planner could eventually compare two network alternatives not simply on which option is cheapest, but on which option produces the best outcome once service, inventory, capacity, operating cost and carbon cost are considered together.

That is particularly important because carbon should not become another isolated reporting system. Where possible, it should become part of the same economic decision model used to plan the supply chain.

Why carbon planning belongs inside disruption planning

Carbon cost becomes even more interesting during disruption. Imagine a company normally imports product by ocean freight. A major disruption forces it to consider air freight.

The operational question is straightforward: will air freight protect customer service? The financial question is: at what cost? The emerging carbon question is: what additional carbon-related exposure does the alternative create?

Likewise, moving production from one facility to another may solve a capacity problem but alter energy consumption and logistics requirements. Changing supplier may reduce risk but increase transport distance. Building additional inventory may protect service but increase warehousing and working-capital cost.

The modern planning environment therefore needs to show decision-makers the total consequence of the response, rather than solving each issue in isolation. This is another reason why disruption planning increasingly requires a network model rather than a simple replenishment calculation.

Organic Planner: O8’s approach to adaptive simulation

This is one of the areas O8 is developing through Organic Planner. Organic Planner is designed to allow organisations to represent, manipulate and simulate their supply networks visually while retaining the planning intelligence underneath.

Users can explore scenarios across capacity, inventory, financial, network and carbon perspectives. They can see where constraints are emerging, what happens when production moves, where shortages or excess inventory occur, what the economic consequences are, how suppliers and lanes interact, and how changing transport, sourcing or production decisions could alter emerging carbon-related costs.

Planning assumptions can be manipulated locally within the simulation environment without first changing the ERP host.

O8’s machine-learning capabilities are also being applied to operational data so that planning can increasingly use observed lead times, variability and behaviour rather than simply relying on historic static master-data values. This allows an organisation to compare both what the system expects and what the supply chain is actually doing.

From dynamic replenishment to adaptive planning

The same thinking applies to replenishment. Approaches such as DDMRP have already challenged traditional forecast-driven MRP thinking by recognising variability and using dynamically positioned inventory buffers.

Machine learning provides an opportunity to extend that concept. A planning system can observe actual consumption, supplier behaviour, lead-time distributions, seasonality, forecast error, service performance and replenishment history. That evidence can then influence future planning parameters and replenishment recommendations.

The progression is therefore from static parameters, to dynamic buffers, to machine-informed replenishment, to adaptive planning.

This is a practical application of AI and ML rather than AI for its own sake.

Why spreadsheets remain attractive

It is worth asking why spreadsheets continue to win during disruption. They have three characteristics traditional planning systems often lack: flexibility, immediacy and disposability.

A planner can change almost anything. No IT project is required simply to test an assumption. A scenario can be created, tested and discarded.

The next generation of planning software needs to preserve those advantages while removing the weaknesses: uncontrolled logic, version confusion, poor scalability, limited optimisation, manual data manipulation, weak governance and dependence on individual spreadsheet authors.

The goal should not be to tell planners to stop using spreadsheets. It should be to give them a planning environment that makes the emergency spreadsheet unnecessary.

What should businesses expect from their planning systems?

If disruption is becoming normal, disruption planning should become a core selection criterion for planning technology.

Businesses should ask whether their planning environment can model the supply network visually, create scenarios independently of live ERP data, change suppliers and routes quickly, use observed lead times and variability, compare capacity, inventory, service and financial outcomes, incorporate demand changes into supply decisions, learn from operational history, represent uncertainty, model emerging carbon-related costs, communicate consequences clearly to non-planners, and generate useful answers quickly enough for a real disruption.

These capabilities may ultimately matter far more than the traditional checklist of planning modules.

From planning system to decision environment

For decades, businesses have bought planning systems primarily to produce a plan. The future requirement is broader. The planning platform needs to help the organisation understand possible futures.

That means producing not one plan, but many. Testing them. Comparing them. Discarding them. Understanding why they differ. Deciding which one the organisation is prepared to execute.

The plan is therefore no longer the only product of the planning system. The greater value is the organisation’s ability to make a better decision when conditions change.

The ultimate test

There is a very simple test for any modern planning platform. Imagine that tomorrow your largest supplier closes, a port becomes unavailable, transport lead time doubles, demand moves unexpectedly, a factory loses capacity, fuel and transport costs rise, or a new carbon cost changes the economics of a key lane. What happens next?

If the answer is that everything is exported into Excel and the planning team is put in a room, then however sophisticated the existing technology may appear during normal operations, there is still a significant capability gap.

The future of supply chain planning should look very different. A planner should be able to open the network, identify the affected area, change the assumptions, simulate alternatives and immediately understand the impact across capacity, inventory, service, financial performance and increasingly carbon cost.

Machine learning should ensure that those simulations are based increasingly on how the supply chain actually behaves, rather than assumptions that have been sitting in master data for years. And the results should be understandable not just to the planner, but to the wider business responsible for making the decision.

Because disruption is no longer something planning systems can treat as exceptional. When disruption becomes normal, simulation has to become normal too.

O8 develops advanced supply chain planning, optimisation and AI/ML technology designed to help organisations make better decisions across complex supply networks. Organic Planner combines visual network simulation with demand, supply, inventory and machine-learning capabilities to provide an interactive environment for planning and decision support.