Skip to content
Shyenova

Supply Chain & Analytics

Demand & Inventory Planning

An inventory policy without a demand signal is a guess. A planning cycle with no inventory consequence is theatre. They are one problem.

Every inventory reduction programme that works on the aggregate number fails within four quarters. Finance sets a days-of-inventory target, planners hit it by starving the fast movers because those turn quickest, service collapses two months later, and the buffer comes back larger than before with a stronger political defence.

Sales and Operations Planning has the mirror-image problem and an equally deserved reputation. In most organisations the monthly cycle has degenerated into a sequence of meetings where each function presents its own number, nobody's number reconciles, ninety minutes are consumed, and everyone leaves and executes the plan they arrived with. The process is running. It is simply not doing anything.

These are the same failure seen from two ends. The useful question about inventory is never how much, but which variability each unit of stock is buying protection against and whether it sits at the echelon where it does the most good, and that is unanswerable without a demand signal you can characterise. A planning cycle, equally, earns its cost only when it forces a trade-off to be made by someone with the authority to make it, and that trade-off is almost always about buffer: what to build, what to hold, what to promise. We run them as one engagement because they are one decision.

How the work runs

Four phases, and what comes out of each.

Durations are indicative and assume reasonable data availability. Where data is poor we say so in the first fortnight rather than modelling on sand.
01

3–4 weeks

Rebuild the demand signal

We sit in the existing cycle and trace what happens to a number, where it originates, who adjusts it, and whether any of those adjustments improve accuracy. Then we rebuild the statistical forecast underneath it with proper hierarchy, segment-appropriate models and honest measurement, including forecast value-add. In most organisations a meaningful share of manual overrides make the forecast worse and can simply be stopped.

Outputs

  • Decision-flow map of the current cycle with forecast value-add at each adjustment step
  • Segmented statistical forecast with model selection by demand pattern
  • Accuracy baseline, and the achievable floor, at the level decisions are actually made
02

2–3 weeks

Decompose the buffer against that signal

ABC by revenue is the least informative cut available, so we segment on demand variability, forecastability, lead-time exposure, margin and criticality instead. Against that segmentation we split existing stock into what it is genuinely doing: cycle stock set by order frequency, safety stock covering demand and supply variability, pipeline stock in transit, pre-build for seasonality, and the residue protecting against nothing at all. That residue is usually between fifteen and thirty per cent of the total and is where the first tranche of cash comes from.

Outputs

  • Multi-dimensional segmentation with a policy class per SKU-location
  • Buffer decomposition by segment and echelon, isolating stock serving no measurable purpose
  • Lead-time variability analysis by supplier and lane
03

3–4 weeks

Set policy across echelons, not within them

Optimising each echelon independently guarantees a worse answer than optimising the chain, every level buffers against the level above, and the aggregate protection compounds far past what the service target requires. Multi-echelon optimisation decides where in the chain a unit of safety stock buys the most service per rupee, which is frequently further upstream and in a less finished form than intuition suggests.

Outputs

  • Multi-echelon safety stock policy by SKU-location
  • Postponement and form-pooling opportunities
  • Service-level curve: achievable service at each inventory level
04

2–3 weeks

Design the cadence that has to hold it

A policy that lives in a spreadsheet decays within two planning cycles, and a planning calendar with no parameters behind it decides nothing. So the two are built together: parameters written into the planning system with the exception logic that tells planners when to override and when not to, inside a cadence designed around the actual decision horizons of your constraints. The textbook five-step monthly cycle is a template, not a law, a business with a two-week replenishment cycle and a six-month capital lead time needs two loops at different frequencies rather than one compromise that serves neither.

Outputs

  • Planning calendar with role-level accountabilities and the decision each forum must make
  • Parameters loaded and validated in the planning system, with planner exception rules
  • Escalation rules with pre-agreed trigger thresholds, and a cadence for re-deriving parameters
05

3 cycles

Run it with us, then without us

We facilitate the first cycle, co-facilitate the second and observe the third, running the new inventory policy in parallel with the old until the numbers are trusted. This is the part most engagements skip and it is where both halves either take root or revert. Between cycles we tune the pack, cut what nobody reads, and hold the line on decisions being made in the room rather than deferred to a side conversation afterwards.

Outputs

  • Three facilitated cycles with structured retrospectives
  • Parallel run of new against old policy until the numbers are trusted
  • Handover to the internal process owner with a capability assessment

What good looks like

The ranges we design toward.

These are the improvement ranges these levers are known to produce where the underlying practice has not been examined recently. They are targets to test against your baseline, not promises.

15–30%

Inventory released

Typical reduction at constant or improved service where policy has been rule-of-thumb based.

5–15 pts

Forecast accuracy

MAPE improvement typical where statistical baselines were absent or unsegmented.

1 number

Operating plan

The point of the exercise: demand, supply, inventory and financial plans reconciled to a single committed set.

Deliverables

What you are left holding.

All of it handed over, documented, and with a session for whoever inherits it.
  1. 01Redesigned planning cycle with named accountabilities and decision rights
  2. 02Statistical forecasting baseline with segment-appropriate models in production
  3. 03Forecast value-add measurement so override discipline becomes evidence-based
  4. 04SKU-location level inventory policy, loaded into your planning system and running
  5. 05Service-versus-inventory curve so the next target is set with evidence rather than ambition
  6. 06Working capital release plan phased by quarter, with an obsolescence exposure register
  7. 07Three facilitated cycles and a documented handover to your process owner

What an engagement looks like

A representative shape.

Illustrative. This is the shape of work in supply chain & analytics, not an account of a specific client. We do not publish named results, and we would be wary of a firm that publishes them without the client in the room.

12 weeks

  1. 01Situation

    A consumer goods manufacturer was opening a third distribution centre on the strength of a regional sales forecast, with freight cost per unit rising and no cost-to-serve view below channel level.

  2. 02What we did

    Reconstructed eighteen months of shipment data into a true origin–destination model, built cost-to-serve to customer level, and tested the proposed footprint against demand growth, rate shock and node-loss scenarios.

  3. 03Where it landed

    The third site turned out to be robust only under the most optimistic growth case. Redrawing catchments and re-tendering two lanes captured most of the projected saving without the capital, and the site decision was deferred rather than cancelled.

Questions

The ones we get asked.

What is the real difference between S&OP and IBP?

Scope and altitude. Classical S&OP reconciles demand and supply volumetrically. Integrated Business Planning extends the same cadence to financials, new product introduction and strategic initiatives, and expresses the outcome in money as well as units, which is what makes it a management process rather than a supply chain one. The mechanics are similar; the difference is whether the CFO's plan and the operating plan are the same plan.

Our forecast accuracy is poor because our demand is genuinely volatile. Can this help?

Partly, and the honest answer matters here. Irreducible volatility is real and no process removes it. But it is routinely used to excuse avoidable error, wrong aggregation level, no promotional decomposition, ignored cannibalisation, models unsuited to intermittent demand. We measure the achievable floor first. If your accuracy is already near it, the answer is not better forecasting but an operation designed to need less of it (buffers repositioned, postponement, shorter lead times) which is the other half of this same engagement.

Will service drop while we take inventory out?

It should not, and if the sequencing is right it improves first. The initial tranche of reduction comes from stock that is demonstrably protecting against nothing, dead cover, duplicated buffers across echelons, pre-build for a seasonality that no longer exists. Service-relevant stock is often increased in the same programme, just at a different echelon and in a less finished form.

We already have an APS. Why is our planning still wrong?

Advanced planning systems compute correct answers from the parameters they are given, and those parameters are usually inherited, uniform and unexamined, a single service-level target applied across a portfolio with wildly different variability, or a lead time that reflects the contract rather than the observed distribution. The software is rarely the problem. The inputs, and the cadence that is supposed to review them, almost always are.

Who from our side needs to be involved?

A demand planning lead, a supply planning lead, a commercial representative with authority to commit to a number, and an executive sponsor who will attend every cycle. That last one is not optional. Every failed S&OP implementation we have reviewed shares the same root cause: the forum had no one in it who could settle a trade-off.

Start a conversation

Talk to us about demand & inventory planning.

Most useful first conversations are thirty minutes and involve no proposal. Describe the problem you are looking at and we will tell you honestly whether it is one we can help with.

Or write directly to hello@shyenova.com