Amazon Case Study interview question on optimization of price
Question
We have a restaurant and we are only into Biryani.
We have an issue that on some days the biryanis get over by 7 pm, sometimes at 10pm. But one some days we are left with 60-100 kgs of biryani.
I want to come up with a model which creates a dynamic price depending on the inventory. If we have a lot of biryani left to sell, then the price should decrease, etc.
What factors should you consider to determine this problem?
Restaurant is only dine-in with a capacity of 50 people at a time.
Assume all the sales data
in progress
0
Case Study
55 years
5 Answers
1010 views
Grand Master 0
Answers ( 5 )
we assume that we 30 days sales data where we found out that
1. 15 days of month biryani is over at 7pm
2.10 days of month it is over at 10pm
3. only on 5 days we have 60 to 100 kg left over biryani
–consideration
1.here we are in India so most of the people it non veg on the weekend
so all 8 days biryani over at 7 days. and some random 7 days also is there
2. on Tuesdays we India’s don’t eat non veg so all four days we left over with 60 to 100 kg
and one Radom days also
3. rest 10 days are normal days people and eat till 10 pm
–solution
1.so we can increase the price on the weekend so we can cover the the lose of left over
2.we can provide some discount one Tuesdays people come and eat.
Factors to consider:
1. Weekday/weekend pattern: Check avg biryani sold for each day
2. Biryani sold on holidays: If holidays affect the amount of biryani sold
3. Consider the avg sales at different time intervals through out the day. Check if trending times on weekdays are different from weekends.
4. If there are any discounts given in the past, analyze how did it affect the sales
5. Since it is only a dine in restaurant, we need to consider how the time discount information will spread to others.
6. People might understand the pattern and will buy at those time of day where it gets cheap. Since there is a sitting capacity of only 50 people, that might be a problem
1. Increase biryani rate over weekend.
2. Over weekdays do provide coupons or discount .
3. We can even try one on one free on thusrday when usually most people dont eat non veg
I want to come up with a model which creates a dynamic price depending on the inventory. If we have a lot of biryani left to sell, then the price should decrease, etc.
What factors should you consider to determine this problem?
Restaurant is only dine-in with a capacity of 50 people at a time.
Assume all the sales data
First, let us define the kind of data we would need for this problem-
Hourly data, which can be aggregated to a daily, weekly and monthly level.
1. Starting inventory of the day (say it is 100 kg) and the price we open our shop at.
2. Hourly updated inventory and the price according to it (as the inventory reduces, our price can increase)
3. Number of orders per hour.
4. Order size – Was it a bulk order or a normal one?
5. Ending inventory of the day.
Now, let us look at the patterns to check the data for, while building an optimization model-
1. Patterns across hours – afternoon (lunch) vs evening(snacks) vs night(dinner) vs late night(snacks)
2. Patterns across days – weekdays vs weekends
3. Patterns across occasions – Some public holiday (other than weekend) vs a normal day
4. Pattern of order size across day type – more bulk orders during weekends?
5. Price variation at a day level.
6. Impact of promotion campaigns on the sales and inventory.
The factors that should be considered for determining this problem are:
– the average consumption of biryani every hour
– the average number of people visiting the dine-in restaurant
– pattern of arrival of customers
– average number of orders in a day
– average order size of each order
– previous responses and arrival pattern of the customers when giving discounts and offers
All these factors need to be looked upon based on weekdays and weekends individually.
Since the dine-in is limited to the capacity of 50, we need to consider other factors such that the duration for which the dine-in stays open, it should have biryani to serve over to its customers.
This being said, the optimal dynamic pricing should be done based on the consideration of the left-over biryani. Taking a record of the left-overs every hour or every few minutes (real-time) could help us develop a better idea of how to get the stock finished only by the end of the day.