Pricing and Demand: Hind Oil
A demand model for an edible oil brand, built with a classmate for Customer Data Analysis (Honors), spring 2026.
The question
Hind Oil wanted to understand what actually moves demand for its product, how sensitive its customers are to price, and how much the company could expect to sell in the coming months.
What we did
Using 54 months of sales data, we modeled monthly demand against the brand's own price, competitor prices, income per person, promotion spending, and the month of the year. We first built a linear model and then rebuilt it on a logarithmic scale, which turns each coefficient into an elasticity: the percentage change in demand that follows a 1% change in each factor. The final model explained about 84% of the variation in monthly demand.
What it showed
A 1% increase in Hind Oil's own price reduced demand by about 0.85%, which means demand moves less than price does and a modest price increase would raise revenue. Competitor prices worked almost one for one in the other direction, with demand rising about 0.91% for every 1% increase in competitor prices, and promotion spending was the most reliable lever the company itself controls, with demand rising about 0.71% for every 1% increase in promotion. Using the model, we forecast demand of roughly 12,500 units for August and 12,700 for September.
What we recommended
We recommended that Hind Oil raise its price gradually. The brand was priced below its competitors in every month of the data, by about 14% on average, and demand moved less than price did, so there was room to close part of that gap. We advised testing each step before the next, because the range around our estimate allowed for the chance that customers would react more strongly at higher prices. We also recommended increasing promotion in measured steps, since it was the most reliable lever under the company's own control, and timing any price change around competitor pricing, since competitors' price moves shaped Hind Oil's demand about as much as its own.
This project gave me a practical way to answer one of the most common questions in any business that sells to the public, which is how far a price can move before customers start to buy less, and how much of that answer depends on what competitors are charging at the same time.
The full Excel model is available on request at tkachr@bu.edu.