Team D - Analysis of Chocolate Industry

Acknowledgement

This research work is being done as an academic project undertaken by our team in IIM lucknow. We would like to extend our sincere thanks to Prof. Sameer Mathur who provided insights and expertise that has greatly assisted our research work so far. We also extend our acknowledgement to our team members for their hard work and dedication.

Research Objective and Problem Statement

Objective -

  • To analyze the impact of various distribution parameters on sales revenue of Nestle and Mondelez.

Problem Statement

  • To identify the distribution parameters that significanctly impact sales revenue.
  • To identify the interactions between variables like company, city, month and category that impacts sales revenue.

Data Variable Definitions

  • Value in Rs. Mn: The total revenue earned from the sales of the units in millions of Rupees
  • MS Unit : This is the market share in terms of the units of chocolates sold.
  • PDO Val (In '000) : Per distributor offtake which is equal to Value in Rs. Mn/ Dealers.
  • SAH (Share among Handlers) : Measure of company's strength within the company's distribution. This is a measure of company's market share in the outlets where it is present.

Data Variable Definitions

  • PPU : Price per unit of the category being considered which is equal to Value in Rs. Mn/ Sales Unit (Mn).
  • NumD : This is representation of company's distribution. This is number of company's outlets/total number of outlets the category is present in.

Data Variable Definitions

  • WtdD Val% - MAX : This measures the quality of distribution i.e. in your distribution what is the category sale.This is equal to category sales in the outlets Mondelez is present in / Total category sales

Regression Models

Regression Model 0

Regression of Revenue with NumD and SAH

# Model 1
 Y variable : Revenue
 X varibles : NumD + SAH + PDO + WtdVal

Regression Model 1

Regression of Revenue with NumD,SAH, PDO, WtdVal

# Model 1
 Y variable : Revenue
 X varibles : NumD + SAH + PDO + WtdVal

Regression Model 2

Regression of Revenue with NumD,SAH, PDO, WtdVal, PPU

# Model 2
 Y variable : Revenue
 X varibles : NumD + SAH + PDO + WtdVal + PPU

Regression Model 3

Regression of Revenue with NumD,SAH, PDO, WtdVal, PPU, City, Month, Year, Company

# Model 3
 Y variable : Revenue
 X varibles : NumD + SAH + PDO + WtdVal + PPU + City + Month + Year + Company

Regression Model 4

Regression of Revenue with NumD,SAH, PDO, WtdVal, PPU, City, Month, Year, Company along with interactions with company

# Model 4
 Y variable : Revenue
 X varibles : NumD + SAH + PDO + WtdVal + PPU + City + Month + Year + Company
 Interactions with Company : NumD + SAH + PDO + WtdVal + PPU

Regression Model 5

Regression of Revenue with NumD,SAH, PDO, WtdVal, PPU, City, Month, Year, Company along with interactions with company

# Model 5 
 Y variable : Revenue
 X varibles : NumD + SAH + PDO + WtdVal + PPU + City + Month + Year + Company
 Interactions with Company : NumD + SAH + PDO + WtdVal + PPU + Month + Year

Regression Model 6

Regression of Revenue with NumD,SAH, PDO, WtdVal, PPU, City, Month, Year, Company along with interactions with company

# Model 6 
 Y variable : Revenue
 X varibles : NumD + SAH + PDO + WtdVal + PPU + City + Month + Year + Company
 Interactions with Company and Type : NumD + SAH + PDO + WtdVal + PPU + Month + Year

Check for Linear Regression Assumptions

  • Linearity
  • Normality
  • Multicollinearity
  • Heteroscedasticity

Linearity

Normality

Multicollinearity

Regression Model 6 shows evidence of Multicollinearity

Heteroscedasticity

Regression Model 6 shows evidence of Heteroscedasticity

Research Insights

  • From the regression summary as per the final model, we see that the significant ;paramters which woulld help in affecting revenue are- NumD, SAH, PDO, WtdD_Val. PPU doesn't hold much affect on the revenue.
  • The ranking of the parameters as per their impact on revenue in Delhi for Mondelez is- PDO SAH WtdD_Val NumD
  • For other cities, it is seen that the impact on revenue decreases when one moves from Delhi to other cities. Out of other cities, the decrement is more evident in the tier 2 and tier 3 cities whereas the tier one cities like Banglore closely resemble Delhi as per this model. Although one outlier is Mumbai which stands out in a way that decrement is in the range of tier 2 cities.

Research Insights

  • The Months of December (Christmas and new Year),August (Raksha Bandhan festival), March (Holi festival) and October-November (Dusshera and Diwali Festival) are most significant in terms of impact on revenue, with December being the highest, followed by October-November, and finally August and March having very less impact. Hence, efforts put in during these months would reap more benefits.
  • For Nestle, most significant parameter is SAH, followed by WtdD_Val, PPU, NumD and then PDO. This is natural to expect since its share is less, while the WtdD_Val is good, giving us conclusion that we need to focus more on increasing in store share in order to increase revenue. All other activities would incur acquisition cost of retailers.

Research Insights

  • For Nestle, One important aspect is that PPU holds a significance and that too on positive side, which implies that people first use small packs as testers for switching to Nestle. Hence Nestle should focus more on taste conversion by using small packets.
  • Revenue difference between Moulded and Assorted is significant whereas, we cant say about the other two categories