Advertisement scheduling in a packet-based media-delivery system
View Patent ↗In one embodiment, a scheme for the display of targeted and personalized advertisements in a packet-based media-delivery system, such as an Internet Protocol Television (IPTV) service. An Internet keyword-based advertisement-bidding model is used to place the most-appropriate IPTV advertisements for viewers depending on their interests as determined through the users' Internet activities, while maximizing advertising revenue for the IPTV service provider. One method for scheduling an advertisement for rendering in one or more time slots in packet-based media programming comprises: (a) obtaining at least one keyword from one or more Internet sessions corresponding to at least one user; (b) receiving a plurality of bid amounts corresponding to a plurality of available advertisements for the one or more time slots; and (c) scheduling, based on the at least one keyword and at least one of the bid amounts, the advertisement to be rendered to the at least one user in the one or more time slots.
1. A method for scheduling an advertisement for rendering in one or more time slots in packet-based media programming, the method comprising the steps of:
(a) obtaining, by a processor at least one keyword from one or more Internet sessions corresponding to at least one user;
(b) receiving, by the processor a plurality of bid amounts corresponding to a plurality of available advertisements for the one or more time slots; and
(c) the processor further adapted for scheduling, based on the at least one keyword and at least one of the bid amounts, the advertisement to be rendered to the at least one user in the one or more time slots by obtaining a selected advertisement for a group of users using a combinatorial algorithm that computes, for all combinations of advertisements i and channels d, a maximum value of θ id based on the equation:
θ
id
=
∑
j
:
c
(
j
)
=
d
max
{
v
ij
-
u
j
,
0
}
,
where:
i represents an advertisement,
j represents a viewer,
v ij represents the bid that an advertiser places on a given viewer j for a given advertisement i,
c(j) represents the channel currently being viewed by viewer j, and
u j represents a revenue value of assigning an advertisement to viewer j.
2. The invention of claim 1 , further comprising constructing or updating a profile for each user based on the keywords, wherein step (c) is performed based on the profile.
3. The invention of claim 1 , wherein the Internet session is a search engine session, and the at least one keyword is a keyword provided to a search engine by the at least one user.
4. The invention of claim 3 , wherein the at least one keyword is obtained by inspecting traffic destined for the search engine.
5. The invention of claim 1 , wherein the at least one keyword is obtained by inspecting the content of web pages viewed by the at least one user.
6. The invention of claim 1 , wherein the scheduling in step (c) is further based on historical data reflecting the viewing habits of the at least one user.
7. The invention of claim 1 , wherein the scheduling in step (c) is performed so that the scheduling maximizes bid revenue.
8. The invention of claim 1 , wherein the scheduling in step (c) is performed by obtaining a selected advertisement a(j) for each user based on the equation
a
(
j
)
=
arg
max
1
≤
i
≤
m
v
ij
,
where:
m represents the total number of advertisements,
i represents an advertisement,
j represents a viewer, and
v ij represents the bid that an advertiser places on a given viewer j for a given advertisement i.
9. The invention of claim 1 , wherein the scheduling in step (c) further comprises:
(c1) constructing a bipartite graph having two columns of nodes, with one column of nodes corresponding to each of the time slots and the other column of nodes corresponding to each of the advertisements; and
(c2) computing a maximum-weight matching of the bipartite graph.
10. The invention of claim 9 , wherein the scheduling in step (c) further comprises, if the numbers of time slots and advertisements are unequal, adding dummy nodes to the bipartite graph to equalize the numbers of nodes in each column.
11. The invention of claim 1 , wherein the scheduling in step (c) further comprises computing an expected revenue for each combination of advertisement and time slot, wherein the maximum-weight matching is performed based on the computed expected revenue.
12. The invention of claim 1 , wherein the scheduling in step (c) is performed so that an advertisement is scheduled only if the user is already receiving content when the time slot arrives.
13. The invention of claim 1 , wherein the packet-based media programming is a digital television system.
14. A system for scheduling an advertisement for rendering in one or more time slots in packet-based media programming, the system comprising:
a computer connected to the Internet; and
a processor adapted to:
(a) obtain at least one keyword from one or more Internet sessions on the computer corresponding to at least one user;
(b) receive a plurality of bid amounts corresponding to a plurality of available advertisements for the one or more time slots; and
(c) schedule, based on the at least one keyword and at least one of the bid amounts, the advertisement to be rendered to the at least one user in the one or more time slots,
wherein the processor is further adapted to schedule by obtaining a selected advertisement for a group of users using a combinatorial algorithm that computes, for all combinations of advertisements i and channels d, a maximum value of θ id based on the equation:
θ
id
=
∑
j
:
c
(
j
)
=
d
max
{
v
ij
-
u
j
,
0
}
,
where:
i represents an advertisement,
j represents a viewer,
v ij represents the bid that an advertiser places on a given viewer j for a given advertisement i,
c(j) represents the channel currently being viewed by viewer j, and u j represents a revenue value of assigning an advertisement to viewer j.
15. The invention of claim 14 , wherein the processor is further adapted to construct or update a profile for each user based on the keywords, wherein the scheduling is performed based on the profile.
16. The invention of claim 14 , wherein the Internet session is a search engine session, and the at least one keyword is a keyword provided to a search engine by the at least one user.
17. The invention of claim 16 , wherein the at least one keyword is obtained by inspecting traffic destined for the search engine.
18. The invention of claim 14 , wherein the at least one keyword is obtained by inspecting the content of web pages viewed by the at least one user.
19. The invention of claim 14 , wherein the processor is further adapted to schedule based on historical data reflecting the viewing habits of the at least one user.
20. The invention of claim 14 , wherein the processor is further adapted to schedule so that the scheduling maximizes bid revenue.
21. The invention of claim 14 , wherein the processor is further adapted to schedule by obtaining a selected advertisement a(j) for each user based on the equation
a
(
j
)
=
arg
max
1
≤
i
≤
m
v
ij
,
where:
in represents the total number of advertisements,
i represents an advertisement,
j represents a viewer, and
v ij represents the bid that an advertiser places on a given viewer j for a given advertisement i.
22. The invention of claim 14 , wherein the processor is further adapted to schedule by:
(c1) constructing a bipartite graph having two columns of nodes, with one column of nodes corresponding to each of the time slots and the other column of nodes corresponding to each of the advertisements; and
(c2) computing a maximum-weight matching of the bipartite graph.
23. The invention of claim 22 , wherein the processor is further adapted to schedule by, if the numbers of time slots and advertisements are unequal, adding dummy nodes to the bipartite graph to equalize the numbers of nodes in each column.
24. The invention of claim 22 , wherein the processor is further adapted to schedule by computing an expected revenue for each combination of advertisement and time slot, wherein the maximum-weight matching is performed based on the computed expected revenue.
25. The invention of claim 14 , wherein the processor is further adapted to schedule so that an advertisement is scheduled only if the user is already receiving content when the time slot arrives.
26. The invention of claim 14 , wherein the packet-based media programming is a digital television system.