Inventory management, loading strategy and warehouse categorization

INVENTORY MANAGEMENT, LOADING STRATEGY AND WAREHOUSE
CATEGORIZATION - GOLCHHA LUBRICANTS
MIHIR SANGODKAR [B15029] | PULKIT AGARWAL [B15037]
SEPTEMBER 11, 2016

Table of Contents
Introduction .………………………………………………………………………………1
Project Approach …………………………………………………………………………1
Data Collection ..…………………………………………………………………………1
Data Cleaning and Formatting ..…………………………………………………………1
Methodology and Analysis for Inventory Management Strategy ……………………..2
Forecasting………………………………………………………………………...2
Safety Stock and Re-order Point ………………………………………………….3
Cost Analysis ……………………………………………………………………..4
User Interface – Summary………………………………………………………...5
Methodology and Analysis for Loading Mechanism and Warehouse Categorization 5
Problem ...…………………………………………………………………………6
Docking Strategy …………………………………………………………………6
Warehouse Categorization (SKU –Wise) ....……………………………………..7
Future Scope ……………………………………………………………………………...8
References ………………………………………………………………………………...9

Page | 1
Introduction
Golchha Lubricants is currently handling a portfolio
of 1600 -1700 KL of lubricants across Shell, Castrol,
Servo and Mak. The distributor is facing with issues
regarding excess inventory for different SKUs of
different brands. The current inventory management
strategy is constrained with regards to SKU level
planning and only provides a brand level strategy.
Golchha lubricants is shifting its storage to a new
warehouse in Adityapur, capable of supporting 120-
3500 KL of lubricants. The old warehouse housed a
serial queuing model for loading. The system caused
delays of 24-48 hours in delivery of lubricants
causing a loss of sale ranging from 5-8 KL a month
per brand.
The project aims at designing an inventory
management strategy that helps the distributor with
SKU level planning for a month order. The strategy
will also encompass product level categorization
with respect storage to ensure efficient loading. The
construction for the docking system will begin in late
October this year and the distributor wants a docking
system that is scalable and easy to implement with
the constraint of minimal product loss while loading.
Project Approach
The project will be divided into 4 stages.
Stage 1: Collection and analysis of inventory and
demand data for different SKUs for the last 2 years.
Stage 2: Detailed analysis of the warehouse facility
with respect to specifications. (This stage will
involve a feasibility study of the docking and
loading mechanisms that can be implemented)
Stage 3: Designing a product level categorization
strategy for storage at the new warehouse.
Stage 4: Designing a coherent SKU level inventory
management strategy and docking mechanism
meeting the prescribed criteria
Project Evaluation Metric
Docking and loading Strategy: Loss of sale should
be minimal: 80 % reduction in loss of sale due to
loading. (In terms of KL sale)
Inventory Management Strategy: No SKU should
remain at the warehouse for more than 3 months.
(Apart from cases where demand fluctuations are
more than 20% for the SKU)
Data Collection
The data collection involved gathering data through
the distributor owner, sales officer or distributor
manager. For the purpose of inventory management
the sales force data was collected. SKU wise files
were generated together. The order quantities and On
hand inventory reports were generated. Price for
cases was obtained from the brand brochures
provided to the distributor by the respective area
managers.
The ordering cost data, lead time and ordering time
period was obtained from the order portal for the
respective brands. The holding cost data was
obtained from hard records maintained by the
distributor manager.
The data for the current loading mechanism was
obtained from the loaders, inspectors and distributor
manager at the ware house. Certain warehouse
dimensions were readily available, others had to be
measured. The investment quote was obtained from
the usual contractor that was often employed by the
distributor.
Data Cleaning and formatting
The main issue with the data collected was the
structure. It was not available in a format that can be
easily used for analysis. The sales force report
generated a separate PDF for each month for every
SKU. The data for all the 18 months for all SKUs had
to be entered in an array fashion which was time
consuming. The pricing data had to be mapped to the
SKU order data. Demand data had to be calculated

Page | 2
for all SKUs for each brand for the all the 18 months
using order data and On Hand Inventory data.
The data for wages was ambiguous and had to be
derived from salary quotes by inspectors and loaders.
The scaling factor for overtime was obtained from
the Factories Act 1948: Section 51, 55 to 56 and 59.
The utilization is assumed to be 100% as the breaks
are not measured and ad-hoc in nature. The time
taken for loading and inspection is an average time
quoted by the distributor manager. The warehouse
layout was not readily available. The layout was
created by the team using
http://planner.roomsketcher.com/. Overtime figures
were assumed to be constant.
Methodology and Analysis for
Inventory Management Strategy
The estimated order quantity and inventory levels
were based on sales estimate coming from the sales
personnel as there was no particular inventory
management strategy adopted by the lubricant
distributor of Shell, Castrol and Servo. Since the
inventory management was sales driven, it often lead
to stock outs of slow moving goods and often bull-
whip effect was observed due to minor seasonality in
demand. Therefore, there was a need to determine the
forecasting for the lubricants to better understand the
demands of the product. Also, it was essential to
determine the safety stock and re order point in order
to maintain desired inventory level to keep the in-
stock ratio and the fill rate at satisfactory levels. In
order to formulate the inventory management
strategy, following steps were taken:
Step 1: Data Selection – This included taking data of
ordered quantity, closing and opening stocks, price
of the lubricant from the data dump received from the
distributor
Step 2: Data cleaning and Formatting – The data
received was not in the desired format to perform
analysis and calculate the desired metrics. Therefore,
data was cleaned and was brought to required format
through transposing etc. to perform calculations
Step 3: Forecasting – Demand forecasting was
performed using three methods in order to accurately
estimate the demand of the product. This will enable
the distributor to take calculated judgements on the
order quantity
Step 4: Safety Stock and Reorder point – Since the
inventory management is based on fixed time period
model, there was no need to calculate the EOQ while
calculation of safety stock level and reorder point
was required. Fill rate and in stock ratio was also
calculated in order for both actual demand as well as
forecasted demand for comparative purposes.
Step 5: Cost Analysis – There are two costs involved
in the distributors operations i.e. ordering cost and
inventory holding cost. There is no transportation
cost involved of the distributor since the prices
margins in the data are discounted for the
transportation cost.
Since data for sales in volume was not easy to
calculate, average slab cost has been considered for
ordering cost calculation of Castrol branded
lubricants.
Forecasting
In order to forecast the demand for lubricant which
have continuous demand with low seasonality, three
methods were employed. The three methods used
were weighted moving average method, simple
exponential smoothing method and lastly linear
regression method. Quantity sold was taken as the
proxy for demand which was calculated by using
data of orders made and opening stock of this month
versus opening stock of next month
Quantity Sold = Opening stock of month 1 +
Orders made in month 1 – Opening stock of month 2
In order to determine the forecast accuracy, Mean
Square Error (MSE) was used for all the three
techniques. It was observed that linear regression
technique proved to be most accurate for all the three

Page | 3
brands and hence was chosen as the forecasting
technique to be used for the lubricant category. An
assumption here is that the demand distribution over
a period of time will not alter drastically. For SES
method, optimum alpha was calculated using the
Solver Analysis in Excel but it can be altered by the
distributor.
Safety Stock and Reorder Point
In order to calculate the Safety Stock and Reorder
Point, mean and standard deviation of the demand
across 17 months data point was considered. Here
approximations for the standard deviation were
utilized due to paucity of data as well as
unavailability of the demand distribution type.
Normal distribution was assumed for he lubricant
demand while standard deviation was calculated
using average daily demand for 17 month time
interval. The expected service level was considered
to be 99% and the lead time and order interval was
15 days and 30 days respectively for Castrol and

Page | 4
Shell lubricants while it was 15 days and 15 days for
Servo lubricants. The formulae that were utilized to
calculate the Safety Stock and Reorder Point are as
follows:
ROP = Average Daily Demand * (Lead time + Order
Interval)
Safety Stock = z* Std. Dev. of daily demand *
√(Lead time +Order interval)
The quantity available for a particular month was
calculated by summing up the ordered quantities of a
particular month with the opening stock of the same
month. This along with quantity sold was used to
calculate the in-stock ratio and the fill rate. The fill
rate and in-stock ratio was calculated considering
both the actual quantity sold and forecasted demand.
It was observed that for slow moving goods
(SLOBS) the in stock ratio improved when we
considered the forecasted demand in comparison to
the actual quantity sold while for the continuous
moving goods, the fill rate as well as in-stock ratio
was quite similar.
For in-stock ratio, if the quantity available was less
than the quantity sold, it was considered as an out-of-
stock situation. For the fill rate, average of the ratio
of quantity sold to quantity available of all the
months was considered.
Cost Analysis
In order to calculate the costs, data extracted from the
dump included ordering cost and holding cost
components along with the price per case of each
product. The data for price per unit was mentioned
which was converted to price per case using the
number of units per case for a lubricant. Ordering
cost was fixed for Shell and Servo but it was variable
on both product category and ordered quantity front
for Castrol. For simplicity in calculation and owing
to data unavailability, the slabs for ordered quantity
were discounted using an average price for a
particular product category. The holding cost had
two primary components i.e. cost of damaged goods
and insurance cost which were obtained as a
percentage of price of a lubricant. For calculation of
holding cost, linear consumption of goods across the
time period was considered and thus average
inventory was quantity available in a month divided

Page | 5
by two. Since the inventory was managed through
fixed order time interval model, the number of
annual orders were fixed to 12 in case of Shell and
Castro and 24 in case of servo lubricants.
Total inventory cost was calculated by summing up
the ordering cost and holding cost for a particular
month. Mean annual total cost was determined. The
safety stock holding cost for each product was
calculated separately in order to understand the cost
associated with keeping the buffer stock at the
warehouse.
User Interface - Summary
A user interface has been created in which the user
needs to enter the SKU code of the respective brand
and the sheets of that brand will be automatically
calculated to give the summary of the results in the
user interface page itself. Some of the summary
statistics include reorder point, safety stock, mean
total cost etc. This sheet will enable the distributor to
have a look at the inventory management strategy at
an overall level and maintain stocks at the desired
level. Only the incremental data needs to be updated
in the working sheets.
Methodology and Analysis for
Loading Mechanism and Warehouse
Categorization
The current loading mechanism at the Golchha
Lubricants warehouse operates through 2 loading
bays. The loading bay dimensions are 5m*4m. The
distributor uses Ashok Leyland Boss 1212 LE
Distribution trucks. The trucks have a width of 2.24
m. The docking width assigned for these trucks
currently is 2.5 m. Consequently only 2 trucks can
dock at a single bay with restricted turning circle.
The front and rear loading at the warehouse prevents
any further entry of vehicles while the trucks are
docked. Consequently the other 4 trucks along with
visiting vehicles have to be parked outside the
compound causing traffic disruptions. The
warehouse also has similar gates at the sides. The
ramp and other docking facilities are not present at
the side of the warehouse.

Page | 6
Each truck is currently loaded by 2 loaders. The
loaders operate on a 10 hour shift. Each truck on an
average takes 3 hours to be loaded. The trucks and
the products at each bay are inspected by an
inspector. The inspection process takes on an average
30 minutes per bay per loading batch (comprising of
the 2 trucks). During the data collection, the
distributor stated that the inspection process for up to
5 trucks can be completed in 30 minutes.
Problem
The major cost center in the loading operations is
overtime. The distributor wants a faster loading
mechanism for loading to save on overtime costs.
The distributor has an investment cap of 10 Lakhs.
Docking Strategy
The docking strategies considered for the analysis
were:
 Flush Dock: The vertical face of the dock is
flush with the outside wall of the warehouse.
To prevent wall damage, often the
foundation/dock bumper extends 4” beyond the
outside wall. Risk of building damage is still high
and hence was rejected.
 Enclosed Dock: These are used for climate
control, product protection, security and
overhead lift capabilities are required. The
truck is parked inside the warehouse during
loading. The space is limited and requires
high construction cost. Since this does not
solve any problems and has high risks and
costs, this is rejected.
195 m 195 m
108 m 108 m
108 m 108 m
15 m
150 m
65 m
5 m
4 m
25 m 25 m
7 m
Max Loading: 1 Truck at a time
Max Loading: 1 Truck at a time
Max Loading:
8 Truck at a time
Max Loading:
8 Truck at a time

Page | 7
 Depressed Dock: These are used when the
there is a need to eliminate basement/ dock
level floors. This does not meet our
requirement due to the sloped nature and
hence is rejected.
 Saw Tooth Dock: This comes in handy when
the space is limited. However there should be
a provision for the trucks to leave in the
direction of the angle of dock, which does not
meet our criteria and hence is rejected.
 Open Dock: The open dock has the truck
trailer open at the docking bay and provides
sufficient product protection. The dock is
ideal for multiple horizontal docks without
interference and hence is selected.
The new bays are proposed on the sides with the
current bays serving the purpose of emergencies and
breakdowns. The dimensions of the truck are 2.24 m.
Since the loading has to be done horizontally one
truck width has been given as tolerance to avoid
interference and product pile up. 1 m each from each
edge has been kept un-allotted due to congestion.
The total required width of the docking bay for 4
trucks comes out to be 19.92 m (Recommended 20
m). One truck width tolerance has been given for
further expansion which when utilized can support
20 trucks. The current suggestion will have 4 trucks
docking on each side. Since the distributor currently
has only 8 trucks, each horizontal docking wall will
see 2 trucks.
According to the analysis, the monthly overtime paid
for the current model amounts to ₹12,00,00 per
month. The new model reduces the overall time
required from 14 hours to 10 hours. The overtime to
be paid with this model essentially turns out to be
null. The construction of the new docking system
will cost ₹67,50,00. The breakeven period for this
investment is 5.625 months (~6 months).
Warehouse Categorization (SKU-Wise)
Currently the warehouse has 4 storage silos. The first
silo comprises of a space of 20m* 140m. The second
silo comprises of a space of 25m*140m. The other 2
silos, adjacent to the proposed loading bay2 are of
sizes 30m*13m each. The total average volume is
used to assign the SKUs to the storage silos.

Page | 8
Castrol has an average volume of 655 KL. Servo has
an average 469 KL. Shell has an average volume of
80.9 KL. The Storage silos 3 and 4 are incapable of
holding the inventory for Servo and Castrol and
hence by default are assigned to Shell products. The
larger volume brand is given the space closer to the
loading bay. Thus the storage area 1 has been
assigned to Castrol.
Storage Areas 1 and 2 will be divided into 2 halves.
Each half would be further divided into 9 categories.
Thus brands would be divided into 18 categories
based on the volume. The highest selling SKUs for
Castrol would be placed to the leftmost category near
the midpoint. The arrow heads in the direction
indicate the direction of ranks. Rank 1 being the
bestselling SKU and Rank 18 being the worst.
The Highest selling Servo SKUs will be placed at the
left-top-most category and the right-bottom most
category.
The Shell SKUs are stored in 2 separate silos in either
direction of the loading bay 2. Each storage silo is
sub- divided into 9 categories.
The bestselling SKUs are placed closest to the bay 2
entrance. The Silo 3 has a clockwise rank
progression and the silo 4 has an anti-clockwise rank
progression.
Future Scope
The project analysis was run for a period of two
months. The possible future commitments to the
project subject to the distributor willingness are:
 Training the loaders with new schemes
suggested in the report for optimum impact
on warehouse activities.
 Potential move to a fixed quantity – EOQ
model with a comparative study between
fixed quantity and fixed period inventory
models.
 Developing an automated warning level with
sales force and Excel model collaboration.
 Optimizing the delivery route for the trucks
using a Logware model.
 Identifying low performing SKUs using
appropriate margin levels.
 TQM application to the different
warehousing operations at Golcha
Distributor.

Page | 9
References
http://www.aalhysterforklifts.com.au/index.php/about/blog-
post/different_types_of_loading_docks [1]
http://www.asprova.jp/mrp/glossary/en/cat249/post-660.html
[2]
http://www.theoperationsmanagement.com/fixed-time-period-
model-4178 [3]
http://www.accountingtools.com/periodic-inventory-system
[4]
http://accountinginfocus.com/financial-
accounting/inventory/inventory-discounts/ [5]
Operations Management Along The Supply Chain, 6th Ed
By Robert S. Russell, Bernard W.Taylor-III [6]
http://mcu.edu.tw/~ychen/op_mgm/notes/inventory.html [7]
post/loading_dock_safety [8]
post/order_picking_in_the_warehouse

Inventory management, loading strategy and warehouse categorization

More Related Content

What's hot (19)

Viewers also liked (11)

Similar to Inventory management, loading strategy and warehouse categorization (20)

Recently uploaded (19)

Inventory management, loading strategy and warehouse categorization