International Journal of Scientific & Engineering Research Volume 8, Issue 8, August-2017 523

ISSN 2229-5518

Supply Chain Inventory Control Model for Nestle

India Ltd.
Ayush Sengupta

Abstract— Nestle India manufactures and sells food products which are associated with a shelf-life. On the other hand there are different
points of sales through the distribution network of the company such as distribution centres, distributors and retail outlets. If at any point,
the stock fails to move on to the next point beyond its shelf-life, it results in expiry of the goods which is a loss to the company in monetory
terms since the company directly or indirectly has to compensate the affected parties in the distribution system and ultimately incur the loss
itself. Nestle refers to the expired stock as Bad Goods, though the term also includes stocks damaged in transit, which forms a very
nominal part of total bad [Link] has been an increasing trend of Bad Goods value over the past few years. As a result, there is a
growing concern among the various stakeholders, especially the Demand and Supply Planning, Supply Chain Department and also the
sales team to some extent, regarding controlling and then reversing the trend of Bad Goods. This is because a factor for the occurrence of
Bad Goods which has been identified is the mismatch, mainly the excess between the planned sales forecasted by the Demand Planning
team of the SCM division, and the actual sales, resulting in unsold stocks and ultimately expired stocks or Bad [Link] paper presents
an Inventory Control model for perishable items in three-level supply chains using just-in-time logistics. The goal is to minimize the total
cost of the whole supply chain. This includes cost of production, cost of freight and cost of inventory holding perished goods. The model,
for the purpose of simplification, has been developed with the assumption that all goods are delivered to the customers prior to their expiry
date and hence there are no perished items. To develop a more realistic model, a scenario analysis can be carried out by considering
different inventory capacities of the warehouses, as different multiples of the sums of the capacities of the customers they cater to. A
sensitivity analysis can also be carried out by making changes to the demand plans. Premium Solver Pro of MS Excel is the software tool
used to solve and validate the model.

Index Terms— demand and supply planning, inventory, just-in-time logistics, Nestle, perishable items, sensitivity analysis, supply chain

1 INTRODUCTION
IJSER ——————————  ——————————

There has been an increasing trend of Bad Goods value over

the past few years. As a result, there is a growing concern
among the various stakeholders, especially the Demand and
Supply Planning, Supply Chain Department and also the sales
team to some extent, regarding controlling and then reversing
the trend of Bad Goods.
PROCEDURE FOR MODEL DEVELOPMENT
2.1 Literature Review
A detailed and extensive literature review was carried out.
Research papers in the relevant fields of supply chain man-
agement, inventory control and perishable goods among oth-
The globalized and competitive market in twenty first century ers were referred to.
requires appropriate supply chain management. In supply
chain management all of the processes (such as planning, 2.2 Field Study and Data Collection
supply, delivery, operating and assessment) present in the A two-month field study was conducted at the company Nes-
network of organizations have to be considered. The interac- tle India Ltd. for the purpose of data collection. The scope of
tions among different economic institutions (such as supplier, the study comprised the organization’s corporate head office
wholesaler, final retailer, etc.) have to be analyzed as well. One in Gurgaon, the Western India regional office in Mumbai and
major difficulty in supply chain is when dealing with perisha- site visits and surveys at warehouses, distributors and varied
ble goods (e.g., food). These goods, if not delivered prior to types of retail outlets at selected urban and rural localities of
their expiry date, are considered as lost sale. On the other Maharashtra and Gujarat. However, the base data used for the
hand, designing inventory control mechanisms to determine development of the Model has not been shared in this paper
the optimal size of orders, the frequency and time of orders, for confidentiality purposes.
and total cost of inventory for perishable products is very
2.3 Model Development
complex. Therefore, organizations dealing with these kinds of
products try to use modern ways to preserve them as much as After extensive studying and understanding of the concepts
possible and also employ ways to minimize the cost of the from the literature review, followed by the field study, the
decaying items while trying to meet the customer's expecta- model development has been carried out. The mathematical
tion. model has been described in the following sections. The list of
assumptions is mentioned in the beginning. The mathematical
equation is presented along with the constraint equations and
the list of notations. The software of Premium Solver Pro has
been used for the purpose of Optimization. The same software
IJSER © 2017
[Link]
International Journal of Scientific & Engineering Research Volume 8, Issue 8, August-2017 524
ISSN 2229-5518

has also been used to carry out the Sensitivity Analysis for the • The demand in each period is constant and determi-
Model. nistic.
• Shortage of goods is permissible.
2.4 Conclusion and future work • The time when goods are sent from factories is consi-
The paper concludes by presenting the scope of future devel- dered the beginning of the expiration period
opment of the model in this paper. • Consumption of goods from warehouses (removed to
deliver to customers) is in first-in-first-out (FIFO) or-
3 LITERATURE REVIEW der.
• The transportation time is neglected as it is a short
The literature review carried out for this paper has been sum- time compared to the total time.
marized and presented in the following paragraphs. One can • The capacity of each supplier is limited.
find here the authors, their paper referred to and the metho- • The customer's demand in each period has been esti-
dology which each of the papers uses for developing their re- mated.
spective models. • Warehouses have limited capacity.
Ghasimi and Ghodsi in their paper “Inventory Optimization • Each warehouse allocates only a certain space to each
for Perishable Items” use the concepts of Mixed Integer Linear type of good.
Programming with tools like Genetic Algorithm and CPLEX. • There is an inventory limit at the end of each period
Tyagi in his paper,”Inventory Optimization of ageing stock by • Production cost worked backwards, using sales price
declining market demand and variable holding cost”deploys Opti- and profit-margins at various echelons
mization considering Weibull distributed deterioration and • Milk-runs considered for transportation from ware-
using tools like Mathematica. house to customers, circular path assumed
Van Elzakker, Zondervan, Raikar, Hoogland, Grossmanin • Identical storage capacity of all distributors
their publication,”Tactical Planning Optimization for FMCG • Inventory holding costs, transportation costs accord-
items considering shelf-life restrictions” implements Mixed Integ- ing to industry standards
er Linear Programming using CPLEX. Simatupang, Wright, • Production plan for considered zones divided accord-
Sridharan in”Theory of Constraints in Supply Chain Collabora- ing to contribution of zones to total sales

IJSER
tion” performs survey, performance metric, theoretic discus- • Production capacities for considered zones divided
sion. according to contribution of zones to total sales
Adebanjo and Mann in their paper,”Identifying problems in • Inventory holding capacities identical for all consi-
demand forecasting in FMCG industry” executes the Integrated dered products
Definition (IDEF) model. Abernathy, Dunlop, Hammond, • Back-order and expiry before final sales not consi-
Weil in their “Information age of retail supply chains” makes us dered
of Surveys, simulation, theoretical discussion
Waller, Johnson, Davis develops their publication, “Vendor- The notations used are as follows:
managed Inventory in retail supply chain” based on Surveys, si- • Factory index i = 1, …, I
mulation, theoretical discussion. Forslund and Jonsson discuss • Warehouses index j =1, …, J
in their paper,”Impact of forecast information quality on supply • Retailer index k =1, …, K
chain performance”the tools of FIQ definition, devising • Period index t =1, …, T
metrics, survey of Swedish companies • Product (Goods) index l = 1, …, L
Reichhart and Holweg in “Creating a customer-responsive
supply chain” talk about the Synthesis of existing literature on A two-month field study was conducted at the company Nes-
manufacturing and supply chain responsiveness. Webby and tle India Ltd. for the purpose of data collection. The scope of
O’Connor in “Judgmental and statistical time series forecasting” the study comprised the organization’s corporate head office
elaborates on theLiterature review of empirical studies on in Gurgaon, the Western India regional office in Mumbai and
comparison between and comparison of judgmental and sta- site visits and surveys at warehouses, distributors and varied
tistical forecast methods types of retail outlets at selected urban and rural localities of
Maharashtra and Gujarat. The inventory levels data at the 3
4 PROJECT MODELLING different levels of the supply chain has been collected in terms
of Indian Rupee (INR) vale. The variables are used in the cal-
The following assumptions have been considered at the be-
culation are the different product categories and the different
ginning of the project modelling:
geographical regions.
• The three levels of supply chain consist of multiple
producers (also called factories or suppliers), multiple
The final equation which has been developed as the result of
intermediate warehouses and multiple distributors
this study is as follows, subject to all the constraints equations
(also called customers). Several customers are
mentioned subsequently.
grouped together into zones, and are allocated to each
of the warehouses
(1) Minimize Z= ∑t∑l∑i∑j (Xijlt. aijlt) + ∑t∑l∑i (pilt. ∑j Xijlt)+
• Road is the only type (mode) of transportation
∑t∑l∑i∑j (bjklt .Yjklt) + ∑t∑l∑j hjlt.( ∑i Xijlt - ∑k Yjklt)
• Multiple products
IJSER © 2017
[Link]
International Journal of Scientific & Engineering Research Volume 8, Issue 8, August-2017 525
ISSN 2229-5518

Subject to: shouldn’t be more than the transportation capacity.

(2) ∑i Xijlt >= dklt for all k, l, t Equation (10) shows that in each period the customer’s re-
(3) ∑j Yjklt >= dklt for all k, l, t quired quantity for an item per period, the quantity of re-
(4) ∑i Xijlt - ∑k Yjklt >=0 for all j, l, t ceived items from each supplier for each warehouse, and the
(5) ∑i Xijlt - ∑k Yjklt <= sjlt for all j, l, t items delivered to customers should all have non-negative
(6) ∑j Yjklt - dklt <= rklt for all k, l, t values.
(7) ∑j Xijlt <= pilt for all i, l, t
(8) Xijlt <= mijlt for all i, j, l, t
5 RESULTS
(9) Yjklt <= njklt for all j, k, l, t
(10) Xijlt, Yjklt, dklt >=0 for all i, j, k, l, t 5.1 Sample problem
Because IJSER staff will do the final formatting of your paper,
Xijlt Quantity of product l transported from factory i to ware- I J K L T
house j in period t 5 2 4 6 8
Yjklt: Quantity of product l transported from warehouse j to
customer k in period t • I: Number of factories
pilt: Production cost of product l in factory i in period t • J: Number of warehouses
hjlt: Holding cost of product I in warehouse j in period t • K: Number of customers
aijlt: Transportation cost of product I from factory i to ware- • L: Number of products
house j in period t • T: Number of time periods (months)
bjklt: Transportation cost of product l from warehouse j to cus-
tomer k in period t Factories Warehouses Customers Products Time pe-
dklt: Demand of product l at customer k in period t riods
cilt: Production capacity of product l in factory i in period t Samalkha, Bhiwandi, Mumbai Nestum Jan-2013
sjlt: Holding capacity of product l in warehouse j in period t Haryana Maharashtra Zone 1 Baby
rklt: Holding capacity of product l at customer k in period t Food
mijlt: Transportation capacity of product l from factory i to Moga, Mumbai Everyday Feb-2013

IJSER
warehouse j in period t Punjab Zone 2 Whitener
njklt: Transportation capacity of product l from warehouse j to Ponda, Goa Aslali, Gujarat Neslac Mar-2013
customer k in period t Gujarat Zone 1
Nanjangud, Gujarat Choco Apr-2013
The purpose of the model is: Karnataka Zone 2 Eclairs
• Reduction in the production cost at the factories Bicholim, Maggi May-2013
• Reduction in the transportation cost of warehouses Goa Sauce
and customers Nescafe Jun-2013
• Minimizing inventory holding costs at the ware- Classic
houses Jul-2013
Equation (1) is objective function of the model, which in- Aug-2013
cludes the cost of producing the items at the factories, the
transportation cost of items from the factories to the ware-
houses and from the warehouses to the retailers and the cost 5.2 Premium Solver Pro Results
of preserving the items in the warehouses.
Equation (2) states that the total demand during the plan- Actual Optimized Difference
ning period should be secured, in other words, the total deli- Total INR INR Less INR 87,
vered items to the customers (distributors) should be greater Cost 2,18,86,301.7 1,31,14,183 72,118.7 (40.08%)
than or equal to the total received demands during the plan-
ning period. The following scenarios were attempted:
Equation (3) also states that the total demand during the Inventory capacity at the warehouses = 1.5 times the base case
planning period should be secured, in other words, the total Inventory capacity at the warehouses = 10 times the base case
produced items at the factories should be greater than or equal
to the total received demands at the customers during the Base Case Scenario Difference
planning period. Case 1 INR INR INR 17,652 in-
Equation (5) shows the limitation of warehouse capacities (1.5 times) 1,31,14,183 1,31,31,835 crease (0.135%)
i.e. the difference between received quantity and delivered Case 2 INR INR INR 43,313 in-
quantity of the items cannot be more than its capacity. (10 times) 1,31,14,183 1,31,57,496 crease (0.33%)
Equation (6) is the space constraints of the distributors and
determines the inventory level in each distributor. The following sensitivity analyses were also attempted
Equations (8) and (9) restrict the capacities for transporta- • Demand at customers 5% higher than base case
tion modes in each period. The amount of transported items • Demand at customers 5% lower than base case

IJSER © 2017
[Link]
International Journal of Scientific & Engineering Research Volume 8, Issue 8, August-2017 526
ISSN 2229-5518

Grossman, I.E., “Optimizing the Tactical Planning in the FMCG In-

dustry Considering Shelf-Life Restrictions
Case Change in demand plan Change in total costs [4] Forslund, Helena and Jonsson, Patrik (2007), “The impact of forecast
3 5% increase 2.03% increase information quality on supply chain performance”, International
4 5% decrease 2.29% decrease Journal of Operations and Production Management, Vol 27, Iss: 1, pp.90-
107
The obtained objective function values show a decrease or [5] Ghasimi, Salah A. and Ghodsi, Dr. Reza (2009), “Improvement and
savings of Rs. 88, 53,415.7 (40.45 percent) as against the actual Solving Three New Supply Chain Inventory Control Models for Pe-
company data for production and inventory for the similar rishable Items using Just-in-time Logistic”, 11th International Confe-
products, factories, warehouses, customers and time periods. rence on Computer Modelling and Simulation
This is because, while at the company the production and lo- [6] Poirier, Charles C., “Forecasting, Demand Management and Capacity
gistics plan is carried out according to the demand forecasts; Planning”
in this model all the production, inventory and transportation [7] Reichhart, Andreas and Holweg, Matthias (2007), “Creating the cus-
costs are also considered alongside demand plans, and the tomer-responsive supply chain: a reconciliation of concepts”, Interna-
optimum solution is arrived at from the perspective of mini- tional Journal of Operations and Production Management Vol 27, No. 11,
mum total cost. pp. 1144-1172
[8] Simatupang, Togar M., Wright, Alan C. and Sridharan Ramaswami
(2004), “Applying the theory of constraints to supply chain collabora-
6 CONCLUSION AND SCOPE FOR FUTURE WORK
tion”, Supply Chain Management: An International Journal, Volume 9,
Number 1, pp. 57-70
The report presents production and inventory plans which
[9] Tyagi, Ankit Prakash (2013), “An optimization of an inventory model
aims at minimizing the total cost which comprises production,
of decaying-lot depleted by declining marker demand and extended
holding and transportation costs. A linear programming prob- with discretely variable holding costs”, International Journal of Indus-
lem is modelled and an optimized solution arrived at. The trial Engineering Computations 5 (2014) 71-86
model is built on the dimensions of 3 levels of the supply [10] Waller, Matt, Johnson, Eric M. and Davis, Tom, “Vendor-Managed In-
chain, namely factories, warehouses and customers; and other ventory in the Retail Supply Chain”, Journal of Business Logistics

IJSER
dimensions of products and time periods. The model makes [11] Webby, Richard and O’Connor, Marcus (1996), “Judgemental and
use of assumptions which have been previously listed, and statistical time series forecasting: a review of the literature”, Interna-
which probably form the limitations of the model. This model tional Journal of Forecasting 12 (1996) 91-118
has been created, keeping in mind products of a food and be-
verages company Nestle India Ltd., and the products are pe-
rishable and come with an expiry period. So in order to give
the model a more realistic form, the expiry of the goods and
the cost of expiry (loss or disposal cost) can also be considered.
Moreover, another improvement can be made in which the
discounting of near-expiry periods so as to prevent expiry and
the cost of doing so is also considered.

ACKNOWLEDGMENT
I take this opportunity to extend my sincere thanks to Nestle
India Limited for providing all the demand and supply plan
data required for the project.
I also thank Dr. Priyanka Verma, my faculty guide, who in-
spired me by our discussions and showed me the right course
to pursue. Prompt and elaborate feedback and suggestions
were always provided and it helped me a lot to stay on the
right track.

REFERENCES
[1] Abernathy, Frederick H., Dunlop, John T., Hammond, Janice H. and Weil,
David (2000) “Retailing and supply chains in the information age”, Technology
in Society 5-31
[2] Adebanjo, Dotun and Mann, Robin(2000), “Identifying problems in
forecasting consumer demand in the fast moving consumer goods
sector”, Benchmarking an International Journal Vol 7 No. 3 pp.223-230
[3] Elzakker, M.A.H. van, Zondervan, E., Raikar, N.B., Hoogland, H. and

IJSER © 2017
[Link]