Inventory Planning and Replenishment
Inventory Planning and Replenishment (IPR) is a solution for inventory control
based on reorder points. The main features of the solution are:
- Classification of parts based on the history records of inventory turnover value, frequency and lifecycle stage.
- The classification can be used to assign inventory
planning policies which are inherited to all parts with a particular
classification. This makes it easy for the planner to deploy a
differentiated planning for a large number of parts in an efficient and a fair
way.
- IPR is integrated with IFS Demand Planning, but
can also be used as a stand-alone solution. In case demand planning is used,
the demand forecast and the estimated forecast error can be used to
calculate planning parameters.
- IPR contains a number of planning models which
allows for successful inventory planning of both high frequent fast movers
as well as slow movers such as spare parts.
Scope of Solution
The solution is aimed for inventory planning of parts
with independent internal or external demands. Independent demands are those
that are not just a function of a demand for another part. Typically this
implies the demand for
sales parts, requested by customers or for spare parts needed for repair. Also
supply should be decoupled since a reorder point system will plan each part
independently of other parts. It means that in case a part is supplied through
transformation of other parts, no
advanced signal will be given to supply also
those parts one level down in the bill of material.
In practice this means that the IPR solution should be used mainly in
distribution and spare parts management. The solution can also be used in businesses where a component is used in a large number of structures as the demand for that component can be seen as decoupled from the demand for the products the component is used in.
IPR can be used as a single planning solution within a company or in combination
with other components such as Kanban for rate-based planning and for master-scheduling and MRP for parts with dependent demands.
It can also be used to plan parts that are purchased, manufactured or distributed from internal suppliers such as upstream warehouses. The solution does not
include any particular support for repair, where parts are supplied through
repair of defect parts.
Define Basic Data
Like any planning solution the IPR is dependent of
accurate and complete basic data. The solution depends on a couple of basic
parameters and attributes such as lead-times, ordering cost and inventory
interest rate. Those parts that should be planned by IPR should have the
Planning Method on the inventory part set to B for reorder point based
planning.
One of the most important elements in the solution is the classification of
parts. The classification is based on historical transactions and will group
parts along four dimensions:
- Asset Class or Site – the classification is done for all parts within a site
or within a particular asset class and site. The classification can be done for
the entire site if all parts are similar from a planning point of view. The
parts can also be divided into different asset classes if they belong to
different categories within the site. This is useful when it is necessary to
distinguish the classification between, for example spare parts, raw materials
and finished goods.
- Volume value – which is the product of the inventory value of the part
and the
issued quantity. A part belongs to either of the classes A, B or C which by
default corresponds to 80%, 15% and 5% of the total inventory turnover value
within the asset class that the part belongs to.
- Frequency – where the number of issue transactions per month is compared with
the defined frequency limits. A part belongs to Fast Movers,
Medium Movers, Slow Movers or Very Slow Movers.
- Lifecycle stage - When the system makes the classification it considers
the lifecycle stage of the pats and classifies them in to individual groups. As parts
mature, decline and eventually become obsolete they will automatically move
between lifecycle stages and the inventory planning policies that applies for a
particular lifecycle stage will be automatically utilized.
The result can be seen as a matrix where for example, fast
moving A parts and
slow moving C parts are easily recognizable. The system will create one of those
matrixes for each lifecycle stage and combination of site and asset class.
In order for the classification to work some basic data
has to be defined:
- ABC classes. By default parts that belong to class A will in total correspond
to 80% of the volume value within that asset class or site, B parts 15 % and
finally C parts, 5%. It is possible to change those percentages. The ranges for
the ABC classification are global and the same values will be used across all
parts planned by the IPR.
- Frequency limits which are used to determine if a part is considered a
fast
mover, a medium mover, a slow mover or a very slow mover within its site, asset class and lifecycle
stage. The frequency limits are defined by site or if applicable on the asset
class.
- Seasonality, it is possible to indicate that the demand pattern for an asset
class should be considered seasonal. If a part is indicated as seasonal the
system will fetch its history a year back, going forward instead of fetching
the most recent history going backwards. Seasonality can be indicated on the
asset class, which means that separate asset classes should be created for parts
with a seasonal demand pattern.
- Lifecycle stage, a part will move between a couple of different, predefined
lifecycle stages. The stages are Introduction, Mature, Decline and
Expired. When
the system makes the classification it will consider the lifecycle stage of the
parts and classify them in individual groups. As parts mature, decline and
eventually become obsolete they will automatically move between lifecycle stages
and the inventory planning policies that applies for a particular lifecycle
stage will be automatically utilized. In order to determine the lifecycle stages
the system uses a couple of offsets that are defined either by site or by asset
class.
It is possible to distinguish how many months of history that should be used for
the classification using the field Classification Periods on the asset class. It is
also possible to indicate the number of periods to use when the classification
job is launched.
Perform ABC, Frequency and Lifecycle Classification
The classification of parts is done on basis of the
history of issue
transactions. In order to simplify the switchover, for example when IFS
Applications is replacing another, system data can be imported into a special
transaction table which will be used together with the transactions created in
IFS during the switchover period.
The classification is useful on its own in order to understand what the most
important parts are, to identify candidates for termination as well as parts
that require extra attention. The classification is also used to define
inventory planning policies as described in the next section.
The classification type that a part has received is shown on the
Inventory Part/General tab.
Define Planning Policies
The IPR calculates four planning parameters which are used to create replenishment
proposals. They are:
- Lot size, which is the quantity that is proposed when a part needs
replenishment.
- Safety stock, which is the quantity in stock that should be held in order to
cover for the variation in demand. The larger the demand variation is expected to
be, the larger the safety stock must be in order to meet a particular service
level.
- Order point, which is the quantity in stock that triggers a replenishment
proposal.
- Next order date, which is the next date a replenishment order should be raised
for the part assuming that the part is consumed in line with its forecast.
Planning Hierarchy
In order to calculate lot size, safety stock and order point a number of
parameters must be defined. These parameters can be defined on the individual
part, but the better way to do it is to use a hierarchy where the lowest level
is the actual parts. Any value defined in the hierarchy will be inherited
downwards. The levels in the hierarchy, starting from the top are:
1. Company
2. Site
3. ABC – Frequency – Lifecycle
4. Asset Class
5. Commodity Group
6. Supplier
A value defined on a lower level in the hierarchy always override a value
defined on a higher level.
The attributes that can be defined on each of the levels are:
• Inventory Interest Rate
• Ordering Cost
• Service Rate (%)
• Demand Model
• Safety Stock Model
• Lot Size Model
• Order Point Model
• Lot Size Cover Time
• Safety Stock Cover Time
• Max Order Cover Time
• Lead Time Factor
Together with available quantities, lead-times and demand this constitutes all
the information the system needs to calculate lot size, safety stock, order
point and next order date.
Demand Model
The value for demand model controls how the system will predict future demands
for a part. In order to calculate the planning parameters it is necessary to
have an estimate of demand and demand variation during the lead-time.
This
estimate can be calculated in different ways depending on circumstances. The
possible values for demand model are:
- Forecast- this value means that the forecast and the expected demand variation
are fetched from IFS Demand Planning. When a forecast is fetched from Demand
Planning, any future changes are considered. It means if the forecast
increases or decreases for future periods, this will automatically be taken into
account and the inventory planning parameters will dynamically change in line
with the forecast. This is very useful for parts with clear seasonal patterns,
trends or campaigns.
- Yearly Prediction - the value for future demand is manually entered on the
inventory part in the field Pred Year Cons Qty.
- History - the transaction history is used to estimate future demand and demand
variation. The result is a fixed value which is considered to be valid for all
future periods.
Note that different demand models can be used for different parts or group of
parts.
Safety Stock Model
The selection safety stock model decides which method that is being used to
calculate safety stock. The following options are available:
- Manual – the value for safety stock is entered manually on the inventory part.
- Time Coverage – the safety stock quantity is calculated as the current demand
forecast from today and the number of days into the future specified by the
value for Safety Stock Cover Time.
- Historical Uncertainty – this safety stock model calculates the optimal safety
stock quantity given a specific service rate. By service rate we mean the
likelihood that a part is available in inventory when it is demanded. For
example the service rate might be set to 97%. This means that if 100 customer
orders with a quantity of 1 are received, then 97 of those orders can be shipped
directly from stock, whilst 3 of them are backordered. The safety stock quantity
is in this case dependent of:
- Historical standard deviation – the higher the
variation is the higher the safety stock must be for a given service
rate. Historical inventory transactions are used to calculate the
standard deviation.
- Lead-time – the longer the lead-time is the
more safety stock is required.
- Lot Size – the higher the lot size is, the
longer the replenishment cycle becomes. It means that the inventory
reach critical levels more seldom, which in turn decrease the necessary
safety stock quantity for a given service level.
- Mean Absolute Error – this model uses the same calculation as Historical
Uncertainty, but the estimate of future demand variation is fetched from IFS
Demand Planning.
Lot Size Model
The selection of lot size model decides which method that is used to calculate
the lot size. The following options are available:
- Manual - the value for lot size is entered manually on the inventory part.
- Time Coverage - the lot size quantity is calculated as the current demand
forecast per day multiplied by the value for Lot Size Cover Time
- Economic Order Quantity (EOQ) which is also referred to as the Wilson formula.
This is a trade-off between inventory holding cost and ordering cost. The result
is dependent on:
- The demand forecast according to the Demand Model used. The lot size will increase
as the forecast increase.
- The part cost, the lot size will decrease as the part cost increase since the
inventory holding cost is higher for more expensive parts.
- The inventory interest rate, higher the
inventory interest rate is, more expensive it is to hold inventory; thus
the lot size will decrease when the inventory interest rate increase.
- The ordering cost which represents all
expenses incurred in placing an order. An increase in ordering cost will
increase the lot size.
Three additional parameters also control the lot size
- Max Order Cover Time can be defined to limit the lot size when EOQ is used.
Very cheap parts will get large lot sizes with EOQ which may cover an
unrealistic time into the future considering the risk of obsolescence etc.
- Durability, if entered, the durability of the part will be considered.
- Min, Max and Multiple Lot Size are considered.
Order Point Model
The selection of order point model decides which method that is used to
calculate the order point. The following options are available:
- Manual - the value for order point is entered manually on the inventory part.
- Lead Time Driven - the order point is calculated as the demand during the
lead-time plus the safety stock quantity. The demand during the lead-time is
calculated according to the valid demand model.
In addition to this four different models are available for slow moving parts,
for example spare parts. These models are based on the assumption that the
demand for the part is Poisson-distributed rather than Normal-distributed.
Typically the models for slow moving parts give more accurate results when the
demand variation is high in comparison to the average demand for the part.
Accuracy in this case is how well the actual service rate aligns with the
specified target service rate. A rule of thumb is that these models are
applicable when the historical standard deviation is larger than half of the
historical demand.
The models for slow movers are based on the likelihood that a demand for a
certain quantity occurs during the replenishment lead-time.
This is after having compared against the defined target service rate, an order point that gives a theoretical service rate that is equal to or exceeds the target service rate is assigned. This means that no explicit safety stock is calculated for these
parts.
The available options for slow movers are:
- Slow Movers – Lifecycle: This model uses the historical transactions to
determine historical demand frequency and quantity. On the basis of this and the
lot size the order point is calculated to meet the specified Service Rate during
the entire lifespan of the part.
- Slow Movers - Lead Time: This model works as Slow Movers – Lifecycle with the
exception that the lot size is not considered. The order point is on this case
calculated to meet the specified Service Rate during one order cycle.
- Croston – Lifecycle: The model is similar to Slow Movers – Lifecycle but
instead of using historical transaction the values for Expected Demand Size and
Inter Arrival Time from Demand Planning are used.
- Croston – Lead Time: The model works as Slow Movers – Lead Time, but instead
of using historical transaction the values for Expected Demand Size and Inter
Arrival Time from Demand Planning are used.