The 10 Key Elements for a Best Practice Supply Chain

IntroductionThe traditional approach is to calculate a moving
Over the past 10 years a great number ofsafety stock based upon a number of weeks
businesses ranging from SME’s to Blue Chipsforecast, sometimes “refined” by using
have been rolling out or updating their Enterpriseforecast variability against historical usage (in APS
Resource Planning (ERP), Distribution Requirementssystems). This method is fatally flawed in 2 ways; it
Planning (DRP) and Advanced Supply Chain Planningrelies exclusively on forecasts in order to calculate
solutions (APS). There are a variety of reasons forthe amount of safety stock required, and it actively
this intense activity, ranging from the need toplans in a level of ‘dead’ stock, with the
consolidate IT following an acquisition, through to theanticipated on-hand levels moving between the
desire to improve the IT capability in order tosafety stock level and safety stock plus the minimum
implement a particular supply chain strategy.order qty.
The 10 key elements for a Best Practice Supply chainThe new approach to inventory target calculation
1. Have a clearly understood and agreed service levelsets a maximum target level of inventory for each
agreement (SLA) with your customersSKU. This is made up of an element of inventory for
The SLA should be a detailed understanding of thethe replenishment time plus an element for demand
service to be offered, particularly in relation to leadvariability, which is statistically related to the required
time, minimum order qty and stock holdingservice level from the stocked item. This approach
requirements. It should also articulate the parametersmakes the entire inventory available for use, with
that define exceptional demand (e.g. a promotion)on-hand levels fluctuating between the inventory
from normal fluctuations in demand that can betarget and zero. It also builds in some sound statistical
accommodated as “business as usual”.probability of material availability based on historical
2. There should be a robust, regular channel ofdemand variability. As long as your S&OP process
communication with your customer, in order toflags up demand that falls outside of this agreed
measure and improve performance levels defined invariability, you will have a lean level of stock that
the SLA.supports your customer SLA at least cost.
Most enlightened businesses now have some kind of7. Completely separate Planning activity from
Sales & Operations Planning (S&OP) processes. ManyExecution activity
however are very inwardly focussed and don’tAnother curse of MRP is its ability to blur the line
include sufficient or any direct input from thebetween planning and execution. A planner is being
customer. This is the opportunity for the customerasked to replan and chase orders daily or even hourly
to communicate significant future demand changesas the MRP “shuffles the order pack” each
for which the supply chain needs to be re-calibrated.time it runs, requiring their constant attention. It is
3. Proper supply chain planning must consider totalvital to separate the activity associated with planning
business cost including Demand, Capacity, Supply &from that of daily order raising and execution. Best
Inventory Planning.practise requires that a plan be set, normally for a
Another common failing of many S&OP processes ismonth, in line with the frequency of the S&OP or
that they do not cover all the elements of cost.forecasting cycle, and then execution happens daily
Typically the debate can be around manufacturingagainst this plan, enabling a set of lower skilled or
efficiency and capacity and ignore the costsautomated actions to be taken daily. This normally
associated with poor customer service or resultantmeans a key change to the skills required by a
inventory. A good S&OP process understands theplanner, meaning considerably fewer but more highly
service model agreed and then determines the leastskilled individuals.
cost way of delivering this.8. Execution tools that allow orders to be raised in
4. Know when and when not to use a Forecastline with appropriate replenishment rule.
Forecasts, no matter how inaccurate, are the bestThere will invariably be the need to cater for a range
tool that we have to determine future capacityof replenishment rules when placing manufacturing or
requirements. Therefore we should have a toolsetpurchase orders, from fixed repeating schedules,
“A common mistake here is to confuse Demandthrough Kanbans and re-order cycle items, spares
Variability calculated entirely from the historicalrequirements, to pure Make to Order. Most ERP/DRP
demand pattern with Forecast Variability, which is thesystems support some but not all of the required
variance between history and forecast. The former istechniques. Therefore you will either need a new
correct the later is meaningless”that enables usorder generation tool that uses the required
to easily access this information. Forecasts areexecution technique to compliment the chosen
typically not bad at determining how much ofreplenishment rule, or you will need to imaginatively
something we will need, i.e. is demand increasing orconfigure your ERP/DRP systems to behave and
decreasing, but very poor a predicting exactly whenraise orders differently.
the demand will occur. Therefore never use a9. Forecasts must be completely eliminated from the
forecast for order generation, to do so flys in theordering/execution process.“There is a belief that
face of any Demand Driven Lean approach.by extending lead-times you increase your available
5. Segment SKUs based upon their demand volumecapacity. This is a myth”
and variability and then select the appropriate“Inaccurate Forecasts are the major cause of
replenishment rule for each segmentcost in all supply chains, and Forecasts are always
The same service level and/or replenishment rule isinaccurate!”
rarely appropriate for all SKUs. Normally there is aInaccurate Forecasts are the major cause of cost in
range of items from high volume, low variability itemsall supply chains, and Forecasts are always inaccurate!
that require a highly repetitive supply plan, through toThe aim should be to never execute an order against
those with sporadic requirements that should ideallya Forecast. Forecasts can however be used as an
be 'make to order'. This segmentation fits closelyindicator of forward demand volumes and linked
with the principles of Lean Manufacturing. A commoncorrectly with actual demand variability from history
mistake here is to confuse Demand Variabilitycan be used to set appropriate inventory policies and
calculated entirely from the historical demand patterntargets.
with Forecast Variability, which is the variance10. When planning, use the shortest possible planning
between history and forecast. The former is correcthorizon in order to minimize the likelihood of plan
the later is meaningless.change and to minimize the number of orders that
6. Use the correct replenishment rule to calculate theneed to be controlled.
correct stock level for each SKU level, to satisfy theThere is a belief that by extending lead-times you
agreed customer service level in the SLA.increase your available capacity. This is a myth.
Once all the levers of cost are understood and theExtending lead-times will upset your customers,
appropriate replenishment rule selected for each SKU,particularly if this violates an agreed SLA, and
an inventory and production plan can be built thatincrease the level of activity and cost required to
delivers the desired customer service Levels. In orderplan. Wherever possible, drive down lead-time, which
to get the correct balance of inventory andwill in turn drive cost out of the supply chain. This
manufacturing cost, a new way of calculatingapproach fits exactly with the requirements of lean
inventory holding is required, that flys in the face ofmanufacturing.
much of the conventional inventory planning wisdom.