This analysis aims to provide the client, Dominick’s Finer Foods chain, with key insights and recommendations to their current pricing of three products in the orange juice category: Minute Maid (MM) 64oz, MM 96oz, and Private Label (PL) 64oz. My analysis is accordingly three-pronged in my approach to the questions of interest.

1. Vulnerability to private label competition

Key findings

PL 64oz poses a competitive threat to both MM 64oz and MM 96oz, as sales of both MM products are influenced by the price of PL 64oz.

Support for findings

I first confirmed that there is enough variation to reliably estimate the own-price and cross-price elasticities for the two MM products.

Indeed, the standard deviation of prices is reasonably large, and the histograms indicate significant variation in the price gaps across stores and weeks.

I then used a log linear demand model that includes time trend, store fixed effects, and pulls only from non-promoted weeks as the basis for my analysis. This model was selected to account for potential confounding variables affecting demand, such as:


Call:
   felm(formula = log(q_MM64) ~ log(p_MM64) + log(p_MM96) + log(p_PL64) +      week | store, data = .) 

Residuals:
     Min       1Q   Median       3Q      Max 
-0.29100 -0.05994  0.00051  0.05464  0.32171 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
log(p_MM64) -2.6011607  0.0322757  -80.59   <2e-16 ***
log(p_MM96)  1.1943832  0.0442094   27.02   <2e-16 ***
log(p_PL64)  1.1779572  0.0345040   34.14   <2e-16 ***
week        -0.0034292  0.0002363  -14.51   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.08983 on 838 degrees of freedom
Multiple R-squared(full model): 0.9479   Adjusted R-squared: 0.9425 
Multiple R-squared(proj model): 0.8949   Adjusted R-squared: 0.8841 
F-statistic(full model):177.2 on 86 and 838 DF, p-value: < 2.2e-16 
F-statistic(proj model):  1784 on 4 and 838 DF, p-value: < 2.2e-16 


Call:
   felm(formula = log(q_MM96) ~ log(p_MM64) + log(p_MM96) + log(p_PL64) +      week | store, data = .) 

Residuals:
     Min       1Q   Median       3Q      Max 
-0.29539 -0.05855 -0.00281  0.04917  0.31841 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
log(p_MM64)  0.9637830  0.0327083  29.466   <2e-16 ***
log(p_MM96) -2.9021210  0.0448020 -64.777   <2e-16 ***
log(p_PL64)  0.3443559  0.0349666   9.848   <2e-16 ***
week        -0.0021579  0.0002395  -9.011   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.09104 on 838 degrees of freedom
Multiple R-squared(full model): 0.9612   Adjusted R-squared: 0.9572 
Multiple R-squared(proj model): 0.8407   Adjusted R-squared: 0.8244 
F-statistic(full model):241.5 on 86 and 838 DF, p-value: < 2.2e-16 
F-statistic(proj model):  1106 on 4 and 838 DF, p-value: < 2.2e-16

From these regression tables, we see that the cross-price effect of PL 64oz on MM 64oz (1.1779572) and the cross-price effect of PL 64oz on MM 96oz (0.3443559) are both positive and statistically significant. This suggests that PL 64oz and MM 64oz are substitutes, as are PL 64oz and MM 96oz.

While the unit increase in log(sales) for MM 64oz is greater than that of MM 96oz in response to log(price) changes of PL 64oz, it remains that both are vulnerable to private label competition. It makes sense that MM 64oz would be ‘more’ vulnerable to price changes in PL 64oz, as MM 64oz and PL 64oz are almost certainly substitutes by nature of offering the same good at the same size (orange juice at 64oz).

2. Cannibalization within the MM product line

Key findings

There is evidence that MM 96oz cannibalizes the sales of MM 64oz, and that MM 64oz also cannibalizes the sales of MM 96oz.

Support for findings

The cross-price elasticity of MM 64oz with respect to MM 96oz is decently large (1.1943832) and statistically significant (Table A).

The cross-price elasticity of MM 96oz with respect to MM 64oz is decently large (0.9637830) and statistically significant (Table B).

3. Current pricing tactics & recommendations

Key findings

: Dominick’s Finer Foods can increase total MM product line profits by 15.57% if the base price of MM 64oz is decreased by 5% and the base price of MM 96oz is increased by 5%. This pricing strategy takes advantage of cannibalization currently within the MM category by further differentiating MM 64oz from MM 96oz.

: Alternatively, they can increase total MM product line profits by 2.29% if the base prices of both the 64 oz and the 96 oz products is increased by 5%.

Support for findings

First, I calculated the base prices using the data for the non-promoted store-weeks, and the base volume as average yearly chain-level volume sales. The given retail margin for the Minute Maid products is 25%; the unit cost of production (and distribution) for the Minute Maid 64oz product is $1.00, and the cost for the 96oz product is $1.40.

I then considered four possible scenarios for MM product line pricing, and calculated the resulting total new expected volume of MM:

Next, I constructed a data.frame with all the proposed price changes in rows. Using (a) a function, predictProfit, that takes base volume, the base prices, the proposed percentage price changes, and the margin and cost data as inputs, and (b) a for loop, I returned the predicted new volume levels for the two MM products and the predicted total profit as a list. See Appendix A for extended code.

I then calculated the predicted profits at the current base prices and predicted percent profit change for the proposed price changes. The main results: