System, method and program product for predicting best/worst time to buy
In response to a user request for information on the best/worst days in an upcoming time period to buy a commodity, a vehicle data system may determine anticipated daily discounts applicable to the commodity. An example commodity may be a vehicle of a specific configuration. In one embodiment, characteristics of month, day of week, and day of month may be gathered and fed into a Best Day to Buy model to determine, for each day of the time period, a projected daily discount relative to a set price for the commodity. Additional input variables such as incentives and seasonal discounts may be included. From the computed daily discounts, the vehicle data system may determine the best day and/or the worst day to buy and report same to the user.
1. A method for predicting best/worst time to buy a commodity, comprising:
determining which month, day of week, and day of month a first day falls on a given year;
for each of a number of days following the first day in the given year, deriving a projected daily discount relative to a set price for the commodity, wherein software running on one or more server computers performs the deriving utilizing historical discounts associated with the month, the day of week, and the day of month in years prior to the given year;
identifying a second day from the number of days to be best day to buy the commodity, wherein the second day is associated with a maximum projected daily discount; and
communicating the best day to buy the commodity to a user over a network.
2. The method of claim 1 , wherein the deriving comprises feeding the set of characteristics into a predictive model.
3. The method of claim 2 , wherein the predictive model comprises a regression formulation Y(Day)=β 0 +Σ i=1 3 β i Z i (Day) where Day represents a day in the number of days, Z 1 represents a first set of variables for the month, Z 2 represents a second set of variables for the day of week, and Z 3 represents a third set of variables for the day of month, and wherein each coefficient β i represents a set of coefficients, each of which is associated with a corresponding variable from the first, second, or third set of variables.
4. The method of claim 2 , wherein the predictive model comprises a regression formulation having a plurality of indicator variables, a first of which represents a first set of variables for the month, a second of which represents a second set of variables for the day of week, a third of which represents a third set of variables for the day of month, a fourth of which represents a major holiday, and a fifth of which represents a minor holiday.
5. The method of claim 4 , wherein the regression formulation for the predictive model further comprises an input variable representing a discount on same day of last year.
6. The method of claim 4 , wherein the regression formulation for the predictive model further comprises at least two input variables representing different incentives.
7. The method of claim 6 , further comprising:
determining an expected value for each of the different incentives, wherein the software running on the one or more server computers performs the determining utilizing an incentive model.
8. The method of claim 1 , further comprising:
on the first day, receiving from a computing device associated with the user a request for information on the best/worst time to buy the commodity, wherein the request specifies the number of days.
9. The system of claim 1 , wherein the instructions are further translatable by the at least one processor to construct a predictive model, the predictive model comprising a regression formulation Y(Day)=β 0 +Σ i=1 3 β i Z i (Day) where Day represents a day in the number of days, Z 1 represents a first set of variables for the month, Z 2 represents a second set of variables for the day of week, and Z 3 represents a third set of variables for the day of month, and wherein each coefficient β i represents a set of coefficients, each of which is associated with a corresponding variable from the first, second, or third set of variables.
10. The system of claim 1 , wherein the instructions are further translatable by the at least one processor to construct a predictive model, the predictive model comprising a regression formulation having a plurality of indicator variables, a first of which represents a first set of variables for the month, a second of which represents a second set of variables for the day of week, a third of which represents a third set of variables for the day of month, a fourth of which represents a major holiday, and a fifth of which represents a minor holiday.
11. The system of claim 1 , wherein the instructions are further translatable by the at least one processor to perform:
determining an expected value for each of a plurality of incentives; and
providing the expected value as input to the predictive model.
12. A computer program product comprising at least one non-transitory computer readable medium storing instructions translatable by one or more machines to perform:
determining which month, day of week, and day of month a first day falls on a given year;
for each of a number of days following the first day in the given year, deriving a projected daily discount relative to a set price for the commodity, utilizing historical discounts associated with the month, the day of week, and the day of month in years prior to the given year;
identifying a second day from the number of days to be best day to buy the commodity, wherein the second day is associated with a maximum projected daily discount; and
communicating the best day to buy the commodity to a user over a network.
13. The computer program product of claim 12 , wherein the instructions are further translatable by the one or more machines to construct a predictive model, the predictive model comprising a regression formulation Y(Day)=β 0 +Σ i=1 3 β i Z i (Day) where Day represents a day in the number of days, Z 1 represents a first set of variables for the month, Z 2 represents a second set of variables for the day of week, and Z 3 represents a third set of variables for the day of month, and wherein each coefficient β i represents a set of coefficients, each of which is associated with a corresponding variable from the first, second, or third set of variables.
14. The computer program product of claim 12 , wherein the instructions are further translatable by the one or more machines to construct a predictive model, the predictive model comprising a regression formulation having a plurality of indicator variables, a first of which represents a first set of variables for the month, a second of which represents a second set of variables for the day of week, a third of which represents a third set of variables for the day of month, a fourth of which represents a major holiday, and a fifth of which represents a minor holiday.
15. The computer program product of claim 14 , wherein the instructions are further translatable by the one or more machines to perform:
determining an expected value for each of a plurality of incentives; and
providing the expected value as input to the predictive model.
16. The computer program product of claim 14 , wherein the instructions are further translatable by the one or more machines to perform:
on the first day, receiving from a computing device associated with the user a request for information on the best/worst time to buy the commodity, wherein the request specifies the number of days.
17. The computer program product of claim 16 , wherein the instructions are further translatable by the one or more machines to perform:
generating a Best Day to Buy report containing information on the best day to buy the commodity; and
delivering the Best Day to Buy report over the network to the computing device associated with the user.
18. The computer program product of claim 16 , wherein the instructions are further translatable by the one or more machines to perform:
obtaining at least a portion of the set of characteristics from a plurality of sources.
19. The computer program product of claim 16 , wherein the instructions are further translatable by the one or more machines to perform:
generating at least a portion of the set of characteristics.
20. A system for predicting best/worst time to buy a commodity, comprising:
at least one processor;
at least one non-transitory computer readable medium storing instructions translatable by the at least one processor to perform:
determining which month, day of week, and day of month a first day falls on a given year;
for each of a number of days following the first day in the given year, deriving a projected daily discount relative to a set price for the commodity, wherein software running on one or more server computers performs the deriving utilizing historical discounts associated with the month, the day of week, and the day of month in years prior to the given year;
identifying a second day from the number of days to be best day to buy the commodity, wherein the second day is associated with a maximum projected daily discount; and
communicating the best day to buy the commodity to a user over a network.