US9967603B2

US9967603B2 - Video viewer targeting based on preference similarity

Info

Publication number: US9967603B2
Application number: US15/348,780
Authority: US
Inventors: Jason LOPATECKI; Adam Rose; John Hughes; Brett Wilson
Original assignee: Adobe Systems Inc
Current assignee: Adobe Inc
Priority date: 2008-09-17
Filing date: 2016-11-10
Publication date: 2018-05-08
Anticipated expiration: 2028-09-17
Also published as: US20180227615A1; US10462504B2; US20110225608A1; US20170094331A1; US9612995B2

Abstract

Presentation of a video clip is made to persons having a high probability of viewing the clip. A database containing viewers of previously offered video clips is analyzed to determine similarities of preferences among viewers. When a new video clip has been watched by one or more viewers in the database, those viewers who have watched the new clip with positive results are compared with others in the database who have not yet seen it. Prospective viewers with similar preferences are identified as high likelihood candidates to watch the new clip when presented. Bids to offer the clip are based on the degree of likelihood. For one embodiment, a data collection agent (DCA) is loaded to a player and/or to a web page to collect viewing and behavior information to determine viewer preferences. Viewer behavior may be monitored passively by different disclosed methods.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is a continuation of U.S. patent application Ser. No. 13/111,705, filed on May 19, 2011 which is a continuation-in-part of U.S. patent application Ser. No. 12/904,552, filed Oct. 14, 2010, which is a continuation-in-part of U.S. patent application Ser. No. 12/212,556, filed Sep. 17, 2008, each of which is incorporated herein in its entirety by this reference thereto.

BACKGROUND OF THE INVENTION

Technical Field

The invention relates to the monitoring of a website and the interaction of a client with a website relative to web content. More particularly, the invention relates to passively monitoring online video viewing behavior to determine viewer preference similarities, and to determine the likelihood that a viewer chooses to view a particular video clip when offered.

Description of the Background Art

Video clips may be supplied to viewers from any website for information, entertainment, or advertising purposes. Some of these websites may be Web 2.0 websites where a user can create an account, upload, share, contribute, comment, vote, or read personal opinions of other users, all on the same site. When video clips are viewed purely for entertainment purposes, users may be more motivated to rate a particular video clip according to their preferences. However, not all viewers expend the effort to rate a video, even if they feel strongly about it.

Video sharing and online video services allow individuals or content publishers to upload video clips to Internet websites. The website stores the video clip on its server, and provides different types of functions to allow others to view that video clip. These websites may allow commenting and rating of a video clip. Many services have options for private sharing and other publication options. Video sharing services can be classified into several categories including, user generated video sharing websites, video sharing platform, white label providers, and web based video editing.

As video hosting websites become increasingly popular, such websites provide a platform for traditional publishers, such as television broadcasters, to use these websites as another medium to display media content which may be of a more informative nature. For example, CBS and CNN networks often publish video clips on YouTube. For such publishers, it is highly desirable to know the ratings of their published video clips. In television the ratings, e.g. Nielsen Ratings, estimate the audience size and composition of television programming, and thus determine advertising rates. This method is not applicable for the Internet. Again, given the opportunity to rate a particular video clip, not all viewers expend the effort to rate the video clip, even if they feel strongly about it.

Either independent of, or associated with, entertainment or informative video clips, advertisement video clips may also be supplied to online users. Websites that supply such advertisement video clips may or may not provide users a means to rate such clips. In circumstances where the advertisement is embedded as part of an entertainment or informative clip, such as a pre-roll advertisement, offering users a voluntary rating facility for the advertisement portion becomes difficult from a practicality standpoint.

In the related art there are different techniques to determine the popularity of a website. One technique known in the art refers to page-hit or page views. The page-hit refers to an event in which a server receives a request for a page and then serves up the page. A common measure of traffic at a website is the number of page hits, especially in an advertising context, for particular pages or sets of pages. Page-hit counts are a rough measure of the traffic of a website. Other techniques involve the analyzing of the traffic between a Web server and clients. Such prior art techniques work well when the traffic of interest relates to particular pages, but are generally not informative when traffic by topic is desired because one page may relate to multiple topics. Systems have been suggested for embedding script code in web pages for tracking user activity on a web page.

Another technique for determining the rating of video clips published on online video sites is based on viewership information provided by these sites. Typically, the sites count the cumulative number of users who view the clip. However, more refined measurements that include, for example, the quantity and characteristics of viewers, as well as detailed information about the duration and repetition of each view, are neither generated by video sharing websites nor by any other prior art technique. Furthermore, viewership information is easily manipulated by the use of, for example, scripts, browser refreshes, and other means that skew the results. As a result, ratings measurements that are based on the viewership information are inaccurate at best and often misleading.

Systems have been suggested for placement of advertising slots within or in proximity to hosted video clips. In such systems, methods are used to track the ad placement and viewing. Such methods require preparation of the video clips for ad placement.

It would be therefore advantageous to provide a solution for online video analytics for generally available video clips on the Internet.

To acquire user preference information for a particular advertisement video clip effectively, and to expand the base of user preference information for any video clip beyond those viewers who voluntarily rate a video clip, it would be useful to provide a solution for acquiring viewership information reflecting viewer preferences without requiring viewers to provide their preferences proactively. Also, when a particular video clip is in the form of an advertisement, it would be advantageous to determine the best target viewers to whom the video may be offered, especially if a bidding process is used to determine which advertising video clip is to be shown to a prospective viewer at a particular moment in time.

SUMMARY OF THE INVENTION

An embodiment of the invention enables presentation of a video clip to a potential viewer who has a high probability of viewing the clip. A database containing viewers of previously offered video clips is analyzed to determine similarities of preferences among viewers. When a new video clip has been watched by one or more viewers in the database, those viewers who have watched the new clip with positive results are compared with others in the database who have not yet seen it. Prospective viewers with similar preferences are identified as high likelihood candidates to watch the new clip when presented. Bids to offer the clip are based on the degree of likelihood. For one embodiment, a data collection agent (DCA) is loaded to a player and/or to a web page to collect viewing and behavior information to determine viewer preferences. Viewer behavior may be monitored passively by different disclosed methods.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a network used to describe the various embodiments disclosed in accordance with the invention;

FIG. 2 is a block diagram of the VAS disclosed in accordance with an embodiment of the invention;

FIGS. 3A-3C show exemplary charts generated by the VAS according to one embodiment of the invention;

FIG. 4 is a flowchart that describes the operation of the DCA in accordance with an embodiment of the invention;

FIG. 5 is a flowchart for capturing, processing, creating and recording a preference score for a video clip in accordance with an embodiment of the invention;

FIG. 6 is a block diagram showing possible communication alternatives for sharing a video clip with other persons in accordance with an embodiment of the invention;

FIG. 7 shows a table containing elements of viewing duration analysis for a sample of video clips and viewers in accordance with an embodiment of the invention;

FIG. 8 is a block diagram illustrating how viewers may be grouped according to videos viewed and similarity of preferences in an example analysis according to the invention; and

FIG. 9 is a flow chart showing how preference similarities are determined and analyzed to determine the likelihood that a viewer chooses to watch a particular video when it is presented in accordance with an embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

According to an embodiment of the invention, a profile is created of those viewers in a database of viewers who are deemed likely to watch a particular video clip when presented. This profile is based on the calculated likelihood for each viewer being over a specific likelihood level defined for the profile. Subsequently, when an opportunity to offer a video clip to a prospective viewer arises, a decision to offer the video may include that prospective viewer being on a list of viewers that matches the profile.

When video clips are used as advertisements, it is especially useful to be able to predict the degree of likelihood that a particular prospective viewer chooses to watch a video clip not seen before by the viewer. If the advertiser must bid for the opportunity to offer a video clip to a prospective viewer, the likelihood prediction is a critical component in computing the amount to bid, and the decision of how much to bid for presenting the video can vary from viewer to viewer based on a specific likelihood calculation for the particular viewer.

According to an embodiment of the invention, determining the likelihood that a prospective viewer chooses to view a particular video clip is accomplished by analyzing preference similarities between the particular prospective viewer and other viewers. This analysis determines which other viewers have similarities most closely aligned with the perspective viewer, and when these other viewers have viewed the particular video clip, it can be inferred from a preference analysis according to the invention that the particular perspective viewer has a certain likelihood of choosing to watch the video clip when presented. The stronger the association of preferences between the particular prospective viewer and the other viewers, the higher the likelihood that the particular prospective viewer chooses to watch the video clip if other viewers with similar preferences watched the video clip in a positive manner. Watching the video in a positive manner may include, for example and without limitation, watching the entire clip, sharing the clip, or purchasing an item or service after watching the clip, to mention a few. A more extensive list of viewership parameters that indicate a positive viewing experience are described herein. While the conventional method for determining user preferences requires a viewer to rate a video clip actively, an effective alternative or supplemental process involves capturing viewership information by passive means and subsequently determining and analyzing user preferences based on captured viewership information.

Monitoring viewer behavior by passive means is useful because users are not required to spend time rating a video clip, something which they are less likely to do. As such, many more viewers have their viewing behavior captured, and a larger amount of behavior information is collected, recorded, and processed than would be otherwise possible if only actively supplied information is used for this purpose, as suggested by the prior art. Also, by concentrating on rating the user rather than just the video clip, it is more likely to be able to match preference similarities with other viewers because more facets of viewing behavior are available for analysis.

As such, in addition to allowing users to rate videos actively, an embodiment of the invention passively monitors and records various user behaviors when a user/viewer interacts with a network video player, e.g. a web video player, while watching an online video clip. In one embodiment, a data collection agent (DCA) is loaded to the player and/or to a web page that displays the video clip. The DCA passively collects, detailed viewing and behavior information without requiring any specific input or actions on the part of the user. Indications of user preferences are inferred by user actions leading up to viewing the video, while viewing the video, and just after and still related to viewing the video. The DCA periodically sends this information to a central server where it is stored in a central database and where it is used to determine preference similarities among different users. Recorded user preference information may also be used to rate a video itself.

In one embodiment, the invention comprises a method and/or an apparatus for monitoring and recording when a user interacts with a video player while watching a video clip online. Specifically, a data collection agent (DCA) is loaded to the player or to a web page that displays the video clip. The DCA collects detailed viewing information and periodically sends this information to a central server. The viewing information is processed by the central server and the central server then generates any of a viewership attention span report, a viewership demographics report, a viewership geographic report, and a viewership sharing report. The attention span report and the sharing report provide inferred user preference information that is used to rate a video clip passively without requiring any specific input from the user/viewer. For purposes of the discussion herein, the terms watching, viewing, and playing are used as interchangeably. When a video clip is playing it is assumed that the viewer is viewing or watching the clip. However, the viewer may in fact be watching something else and not the video clip at a particular moment in time. There is no way to know for sure and, thus, the assumption is made.

FIG. 1 is a block schematic diagram of an arrangement of elements 100 that may be used in connection with the invention. These elements 100 include at least a web server 110 for hosting video sharing websites. The websites include, but are not limited to, YouTube™, MetaCafe™, Yahoo™ video, Myspace.com™, users' blogs, etc.

A viewership analytics server (VAS) 120 is configured to connect to each web server 110 through a network 130, for example, but not limited to, a wide area network (WAN), which enables connectivity such as Internet connectivity. The VAS 120 executes the tasks related to gathering of viewership information for web servers 110, analyzing the gathered information, and generating reports on the quantity and characteristics of viewers, as well as providing information about the duration and repetition of each view. These tasks are described in greater detail below. The VAS 120 is connected to a database 140 in which the collected and generated viewership data is saved.

Clients 150-1 through 150-M communicate with web servers 110 through the network 130. A client 150 comprises at least a web browser, such as Microsoft™ Internet Explorer, that allows the user to view and navigate through web pages downloaded from one or more servers 110. Each client 150 is capable of downloading, playing, and displaying video clips provided by the servers 110. With this aim, each client 150 is capable of running a video player (not shown), which is typically integrated into a web page. The video player may be, but is not limited to, a Flash-based web player, DivX web player, HTML5 player, Microsoft Media Player, etc.

In accordance with the principles of the invention, a data collector agent (DCA) is loaded to video sharing websites that are hosted on servers 110 to capture information about the interactions of the viewers with web players. The DCA may be a script code, e.g. JavaScript, hosted by the VAS 120 and loaded to web pages hosted on servers 110. The DCA may be also in a form of a plug-in installed in the video players provided by video content providers.

The DCA collects and sends metadata and detailed viewing information to the VAS 120. The metadata comprises at least a video identification (ID), a publisher ID, a website ID that is derived from the uniform resource locator (URL), a length of the video clip being viewed, and the current time. The detailed viewing information includes the actions performed on the player and a timestamp. The recorded actions may be, for example, playing, pausing, rewinding, forwarding, and so on. The timestamp start and end times are expressed, for example, in seconds from the beginning of the video clip. For instance, the pair <play, 20-35> means that a user viewed the clip for only for 15 seconds starting at the 20.sup.th second from the beginning. The pair <pause, 30> means that the user paused 30 seconds after the beginning of the clip. The data gathered by the DCA is used by the VAS 120. In one embodiment, these requests are sent to the VAS 120 in the form of a hypertext transfer protocol (HTTP) request. An HTTP request that includes the metadata is sent to the VAS 120 once a web page, including the DCA, has completely uploaded to a client's 150 browser. The detailed viewing information, including the pairs of actions and timestamps, is periodically sent to the VAS 120. The VAS 120 extracts the data encapsulated in the received requests and saves the data in the database 140.

In accordance with one embodiment of the invention users, e.g. advertisers and content publishers, can access the VAS 120 through, for example, a client 150. This process is similar to that used when a user viewing the content accesses the VAS 120. Advertisers and content publishers can designate which websites, publishers, and video clips to trace. In one embodiment of the invention, the user views generated data from the VAS 120 by logging onto a website.

FIG. 2 is a block diagram showing the VAS 120 implemented in accordance with one embodiment of the invention. The VAS 120 includes an information collection module 210, an analyzer 220, and a graphical user interface (GUI) module 230. The collection module 210 communicates with a DCA on a client 150 for the purpose of receiving HTTP requests and responding thereto.

Specifically, the module 210 generates HTTP responses containing the script code of the DCA. The information collection module 210 further receives the HTTP requests, including the data collected by the DCA, extracts the information from the requests, and saves the information in the database 140. This information includes detailed viewing information and content metadata, which is saved together with tracking data including, but not limited to, the Internet protocol (IP) address, as well as the operating system and browser type of the client 150. The detailed viewing information is saved in an entry associated with the video ID. In an exemplary embodiment, the database 140 includes a table having the following fields: video_ID, website_ID, publisher_ID, date, IP, OS, browser type, and <action, timestamp> pairs.

The analyzer 220 processes the information saved in the database 140 to generate viewership-related analytics data, an attention span report, and viewership demographics. Viewership-related analytics data includes, but is not limited to, the number of viewers during any period of time, e.g. last three days, last week, last months, etc. for a video clip, for a publisher, or for a group of video clips over different periods of time. This information can be generated for a single website or across a plurality of websites. To generate the analytics data mentioned herein, the analyzer 220 first computes the number of viewers in each day, or any other time interval, from the gathered information. The process for generating the viewership-related analytics data is further discussed in U.S. patent application Ser. No. 11/871,880, A Method and System for Monitoring Online Video, assigned to a common assignee, the entirety of which is incorporated herein by this reference thereto.

The analyzer 220 also generates an attention span report that includes detailed information about the duration and repetition of each view. This report includes, per each video clip, the total number of viewers, and the number of viewers that viewed the complete video clip. This report is produced by processing the data stored in the database 140. In addition, the analyzer 220 produces a viewership-geographic report. This report includes the number of viewers of a video clip in each country around the globe. The report is generated by correlating the number of views with IP addresses of the different viewers. Furthermore, a viewership demographics report is generated by analyzer 220. This report correlates the number of viewers with demographics including race, age, income, educational attainment, employment status, etc. The demographics are retrieved from the users' profiles as saved in the online video websites, if and when available.

In accordance with another embodiment, the analyzer 220 can detect fraud attempts. Such attempts are typically performed by browser refreshes or scripting intended to inflate the view count artificially. With this aim, the analyzer 220 maintains a history file of the video IDs that have been viewed in the past during a predefined period of time, e.g. video clips viewed in last two hours, by each IP address. If the analyzer 220 detects multiple views above a threshold from the same IP address within a predefined period time, the analyzer 220 discards the data regarding the subsequent views or any views. The analyzer 220 also validates that the database 140 does not contain timestamp entries with duration longer than the length of the video clip. This check protects against scripting attacks intended to record repeated video views under a single view count.

The GUI 230 displays the viewership-related analytics data produced by the analyzer 220 as charts or text-based reports. In one embodiment, the charts are dynamic. That is, the GUI 230 dynamically changes the displayed content of the chart as the user changes the chart's time scale. FIGS. 3A-3C show examples of charts of the various reports as generated by the GUI 230 according to several embodiments of the invention. FIG. 3A is a chart that shows an attention span report. FIG. 3B is a chart representing the viewership by geography. FIG. 3C shows charts of viewership demographics, specifically, the age distribution and gender distribution of viewers.

FIG. 4 is a flowchart 400 that shows the steps for operating the DCA in accordance with one embodiment of the invention. When a web page that includes a video clip is loaded, the DCA is inserted S410 on the page and sets a third party cookie in the browser. The third party cookie is used to track the video viewing activity of each unique user across all video providers. The DCA is inserted on the web page using an HTTP response from the server 110 and contains a script code. The DCA generates S420 an HTTP request that includes metadata and sends the request to the VAS 120. This request contains the provider site in the URL path and the ID of the video being viewed, the local client's current time, the client time zone offset, and the non-personally identifiable provider user ID. The VAS 120, upon receiving this request, extracts the metadata and saves it in database 140. Once the video clip is internally loaded in the player, the DCA generates S430 HTTP requests that include the detailed viewing information, for example in the format described above. Thereafter, these HTTP requests are periodically sent S440 to the VAS 120. Once the web, page is closed or the browser window location is changed while a video clip is being viewed, the DCA transmits an HTTP request that includes the final data pair that ends at the current viewing time-point.

The earlier discussion regarding FIG. 1 refers to collecting detailed information about the duration and repetition for viewing a particular video clip. The DCA functionality residing in the viewer's computer 150 or supplying website 110 collects this information passively and may include detailed information to enable further analysis of parameters related, for instance, to viewing duration and repetition. For instance, the DCA may passively record whether a viewer watched a particular video clip at all. It may also record what specific portions of a video clip the viewer watched and the sequence in which they watched those portions. Viewers are known to rewind entire videos, rewind portions of a video, or skip around within a particular video, watching different portions in a regular or irregular sequence. In fact, many other actions of a viewer may be passively observed and used to rate the viewer according to their preferences as well as rate a video clip itself. FIG. 5 is a flowchart 500 that shows the steps for passively collecting and processing behavior or viewership information related to a video clip. Viewer actions that reflect their preferences with respect to a particular video clip may occur before they play the clip, given some foreknowledge of the clip content; while they watch the clip; and after they watch the clip. In step S510, parameters respective of a viewer's behavior and actions leading up to their playing a video clip are recorded. Examples of such actions include but are not limited to:

(i) Enduring a pre-roll advertisement, thus showing an affinity and/or tolerance for ads;

(ii) Accessing the video clip by a particular method:

(a) Authenticating or logging-in, thus showing higher interest, to view the video clip;

(b) Being incentivized to see the video, such as obtaining points in an online game; and

(iii) Adjusting the bit rate for streaming or downloading the video prior to playing it, and in which direction they adjusted the bit rate, e.g. faster or slower.

(iv) Clicking Play in the event that the video isn't auto-playing.

In step S520, parameters respective of a viewer's behavior and actions during their playing of a video clip are recorded. Examples of such actions include but are not limited to:

(i) Adjusting the volume level, muting, and un-muting;

(ii) Pausing and un-pausing;

(iii) Fast-forwarding and rewinding;

(iv) Replaying the video clip, how many times it was replayed, and what portion was replayed;

(v) Viewing duration and % completion;

(vi) Viewing in full-screen mode;

(vii) Adjusting the bit rate for streaming or downloading the video while playing it, and in which direction did they adjust the bit rate, e.g. faster or slower;

(viii) Clicking-through on an overlay image or the video itself to be taken to another webpage or website;

(ix) Spending time viewing certain segments of the video clip that are judged to have higher value based on actions of previous viewers; and

(x) Enduring a mid-roll advertisement, thus showing an affinity and/or tolerance for ads.

In step S530, parameters respective of a viewer's behavior and actions after playing of a video clip are recorded. Examples of such actions include but are not limited to:

(i) Sharing the video via an embed code. Generally, with online video a user can copy a small HTML tag out and paste it into their blog to share the video. The system according to the invention tracks the action of copying out that HTML code. Such action may be the operation of a button or simply the act of highlighting the text in a textbox;

(ii) Sharing the video via E-mail;

(iii) Sharing the video via a social network;

(iv) Sharing the video in multiple separate events and how many separate sharing events the viewer initiated to share the video clip;

(v) Waiting for a recorded time duration between viewing the video clip and initiating a sharing event;

(vi) Bookmarking the video clip for later viewing;

(vii) Downloading and saving the video for later viewing; and

(viii) Subscribing to a channel or feed for the video content producer or artist to become better aware of new videos.

In step S540, a viewer preference score is created for the video clip based on the particular viewer's behavior and the score is associated with the user as metadata. In step S550, a video clip score is created for the video clip based on the particular viewer's behavior and the score is associated with the video clip as metadata. When a score is calculated for either the user's preference or the video clip itself, the different factors above that contribute to the score are optionally weighted in the score calculation. For instance, the fact that a video clip is shared may have more value in calculating a preference score than the fact that they paused the video clip. The weighting may also be different for calculating the score for a user's preference relative to a particular clip as opposed to the score for the clip itself. For instance, the fact that the viewer watched the pre-roll advertisement may be more valuable for scoring the user than for scoring the video. Another example of weighting occurs where an un-mute or full-screen action is considered to be a highly-valuable action, whereas simple viewing duration for a video that plays automatically may not be, as it may simply be playing in the user's browser without their attention.

In step S560, the viewer preference score and the video clip score are recorded in a database.

Network

600 in FIG. 6 shows a flow of information within an embodiment of the invention with regard to sharing of information related to video clips between users. Sharing typically occurs by way of a wide-area network 130 normally comprising the Internet or a portion thereof. The sharing process is initiated by the terminal belonging to a primary viewer/user 150 who decides that a particular video clip is interesting enough or entertaining enough to share with a friend. A terminal or communication device 640 as shown would typically be used by a friend or a group of friends. Many types of communication devices or terminals are known in the art to be used for communication within social networks and include PCs and smart phones. The process is initiated when the primary viewer at terminal 150 is supplied a video clip 610 from a webpage 110. This video clip may be supplied automatically without any initiation by the user, or it may be supplied to the user upon some action the user takes to view the video clip. When the primary viewer decides they wish to share a video clip they may be offered the ability to share by way of webpage 110, or alternately by providing the link to the particular video clip to a third-party link sharing website 660. When a sharing request 620 is made by viewer 150 to website 110, website 110 can notify a friend at terminal 640 via email or directly if the friend at terminal 640 is logged into website 110.

User

150 may also share the link to a video clip through communication path 650 and a third-party website 660 where link sharing is supported. Examples of link sharing websites include, but are not limited to, digg, del.icio.us, and reddit. Links to video clips can also be shared in a similar manner via social networking sites such as Facebook™ and twitter. Sharing behaviors can be monitored by DCA functions located in the viewer's computer 150, located on a server at the supplying website 110, or both. Data gathered by these DCA functions is periodically transferred to VAS 120 for logging, analysis, and reporting. The link sharing website may communicate with a friend using terminal 640 via email or directly if terminal 640 is logged into website 660. In addition to the sharing mechanisms described in FIG. 6, sharing may also be accomplished via an embed code as previously described for step S530 of FIG. 5.

Passively monitored sharing activity may include at least one or more of the following:

i. If a viewer shared a particular video clip;

ii. How many times they shared the video clip;

iii. How many persons they shared the clip with;

iv. Which sharing mechanisms and/or websites they chose to use to share the video clip; and

v. The time duration between when they watched the clip and when they shared it.

While immediately sharing a video clip certainly shows interest, sharing the video clip after a period of time has passed shows that the video clip left a lasting impression on the viewer. Also, should the user initiate a large number of separate sharing events or sharing actions for a particular video clip, this may provide an indication that the video clip deserves a higher score for the sharing-based component of its score.

Thus, passively observed user actions and behaviors are used to rate a particular user's preferences relative to the video clip in question, as well as rate the video clip itself.

One parameter that may be passively acquired and then used as a component of a viewing preference analysis is viewing duration. FIG. 7 shows an exemplary and non-limiting table 700 containing elements of a viewing duration analysis for a sample of video clips. Each of video clips A through G have been previously viewed by at least two of the three viewers shown: Joe, Bill, and Jane. For this example, it is assumed for the sake of simplicity that a judgment is made based solely on viewing duration analysis. In reality, the judgment relative to the likelihood of a viewer choosing to view a new video is far more complex and may include any or all of the viewership parameters discussed previously that may be passively captured prior, during, or after the viewing of a video clip, or alternately proactively supplied by a viewer.

For the analysis of FIG. 7, note that for video A the viewing durations for Joe and Bill were both high, indicating a similar preference level with regard to this parameter. Likewise for video B, the viewing duration for all three viewers was low, indicating again a similar preference level. Note that preference similarities may be positive or negative, and the degree of similarity is used by analysis algorithms according to the invention to determine a likeliness score, that score indicating the likelihood that a prospective viewer chooses to watch a video they have not seen before. For videos C, F, and G, there does not exist enough commonality of viewing to be useful in making a judgment. Thus, the greatest similarity of preferences among the viewers shown in FIG. 7 with regard to previously viewed videos is between Joe and Bill. Later, an opportunity arises to present to Bill a new video 710 called video Z. Bill has not seen video Z before.

An analysis of the database containing preferences relative to previously viewed videos shows that because Joe had viewed video Z with a 95% viewing duration, there is a high likelihood that Bill chooses to watch video Z when it is offered to him. If video Z is an advertisement video clip where an advertiser bids for the opportunity to present the clip, the likeliness score for Bill to watch video Z may be used to determine the amount of the bid, i.e. the higher the likeliness score, the higher the bid.

FIG. 8 is a block diagram 800 describing one possible organization of viewers in a database according to the invention. For simplicity and in keeping with the simplified example of FIG. 7, diagram 800 of FIG. 8 refers to an organization in keeping with table 700 of FIG. 7 and, in particular, refers to the grouping of viewers in the database with respect to video clip Z. Group (1) 810 comprises viewers of previous video clips who have not viewed video clip Z. With respect to FIG. 7, this group contains Bill and Jane. Group (2) 820 comprises viewers who have viewed clip Z in a positive manner. With respect to FIG. 7, this group contains Joe. Group (3) 830 comprises viewers who have not yet viewed video clip Z, however have preference similarities most closely aligned to viewers in Group (2) 820 who have previously viewed clip Z in a positive manner. With respect to FIG. 7, Group (3) 830 contains Bill. Thus, Bill is deemed to have a high likelihood of choosing to watch video clip Z if and when it is presented to him.

FIG. 9 shows a flowchart 900 whereby preference similarities are determined and analyzed to determine the likelihood that a viewer chooses to watch a particular video clip when presented.

In step 910, a database is constructed based on captured viewership information respective of a first set of video clips. The capturing process for viewership information may be active whereby the viewer proactively supplies preference information, or passive whereby viewership information is captured without requiring any action by the viewer in accordance with one or more methods previously described for the invention.

In step 920, the database is analyzed to determine preference similarities among viewers who have previously viewed the first set of video clips.

In step 930, viewer characteristics respective of a second video clip are captured and recorded in the database. Typically, some of the viewers of the second video clip have previously watched the first video clip.

Subsequently in step 940, viewership characteristics respective of the second video clip are analyzed with respect to the database to produced a list of viewers in the database who are deemed most likely to watch the second video if and when it is offered to them. As described for FIG. 7, a prospective viewer who has not seen the second video, but has strong preference similarities with someone who has responded in a positive manner to viewing the second video, is determined to be a likely candidate to watch the second video when presented. The degree of likelihood is computed according to the present invention by analyzing each preference-related parameter of the viewership information, weighting the parameters as appropriate, and developing a likelihood score as a result. Should the second video clip be an advertisement, and should the advertiser be required to place a bid to offer the video to the prospective viewer, the amount of the bid is determined based at least in part on the likelihood score.

In step 950, the second video clip is presented to the prospective viewer who was previously determined to be part of a list or group of viewers likely to choose to watch the second video clip.

It should be appreciated by a person skilled in the art that methods, processes and systems described herein can be implemented in software, hardware, firmware, or any combination thereof. The implementation may include the use of a computer system having a processor and a memory under the control of the processor, the memory storing instructions adapted to enable the processor to carry out operations as described hereinabove. The implementation may be realized, in a concrete manner, as a computer program product that includes a tangible computer readable medium holding instructions adapted to enable a computer system to perform the operations as described above.

Although the invention is described herein with reference to the preferred embodiment, one skilled in the art will readily appreciate that other applications may be substituted for those set forth herein without departing from the spirit and scope of the present invention. Accordingly, the invention should only be limited by the Claims included below.

Claims

What is claimed is:

1. A computer implemented method for outputting a video clip to a display screen of at least one prospective viewer in a target list of prospective viewers, comprising:

generating and storing a database of viewer preferences related to a first video clip;

analyzing the database of viewer preferences related to the first video clip to determine preference similarities among viewers in a first list of viewers who watched the first video clip;

recording, in the database, viewer preferences related to a second video clip;

analyzing the viewer preferences related to the second video clip to:

identify a second list of viewers who watched the second video clip in at least one of a plurality of positive manners;

generate the target list of prospective viewers in the database who have not yet viewed the second video clip and are likely to watch the second video clip based on preference similarities with viewers in both (i) the first list of viewers who watched the first video clip and (ii) the second list of viewers who watched the second video clip in the at least one of the plurality of positive manners, wherein a likelihood determination for prospective viewers in the target list is made by:

analyzing at least one preference-related parameter associated with viewers,

weighing the at least one preference-related parameter,

developing a likelihood score based on the weighed at least one preference-related parameter, and

comparing the likelihood score with a defined likelihood level; and

causing the second video clip to be output to the display screen of the at least one prospective viewer in the target list of prospective viewers.

2. The method of claim 1 wherein generating and storing the database of viewer preferences related to the first video clip comprises collecting the viewer preferences passively without requesting a specific viewer input for rating the first video clip.

3. The method of claim 2, wherein viewer preferences related to a corresponding video clip reflect a viewer action occurring at least one of prior to playing the corresponding video clip, while playing the corresponding video clip, or after playing the corresponding video clip, wherein the corresponding video clip comprises the first video clip or the second video clip.

4. The method of claim 3, wherein the viewer action prior to playing the corresponding video clip comprises at least one of:

viewing a pre-roll advertisement,

accessing the corresponding video clip by a particular method comprising at least one of: authenticating, logging-in or being incentivized to see the corresponding video clip,

adjusting a bit rate for streaming or downloading the corresponding video clip prior to playing the corresponding video clip, including a direction in which the viewer adjusted the bit rate, or

clicking play if the corresponding video clip is not an auto-playing video clip.

5. The method of claim 3, wherein the viewer action while playing the corresponding video clip comprises at least one of:

adjusting volume level, muting, or un-muting the corresponding video clip,

pausing or un-pausing the corresponding video clip,

fast-forwarding or rewinding the corresponding video clip,

replaying the corresponding video clip, how many times the corresponding video clip was replayed, and what portion of the corresponding video clip was replayed,

viewing the corresponding video clip for a recorded duration and percent completion,

viewing the corresponding video clip in full-screen mode,

adjusting a bit rate for streaming or downloading the corresponding video clip while playing it, including a direction in which the viewer adjusted the bit rate,

clicking-through on an overlay image or on the corresponding video clip itself to be taken to another webpage or web site,

spending time viewing certain segments of the corresponding video clip that are judged to have higher value based on actions of previous viewers, or

viewing a mid-roll advertisement.

6. The method of claim 3, wherein the viewer action after playing the corresponding video clip comprises at least one of:

sharing the corresponding video clip via an embedded code,

sharing the corresponding video clip via e-mail,

sharing the corresponding video clip via a social network,

sharing the corresponding video clip in multiple separate events, where a number representing separate sharing events initiated by the viewer to share the corresponding video clip are recorded,

waiting for a time duration between viewing the corresponding video clip and initiating a sharing event, wherein said time duration is recorded,

bookmarking the corresponding video clip for later viewing,

downloading and saving the corresponding video clip for later viewing, or

subscribing to a channel or feed for a corresponding video clip content producer or artist.

7. The method of claim 1, wherein analyzing the database of viewer preferences related to the first video clip to determine preference similarities among viewers in a first list of viewers who watched the first video clip comprises determining preference similarities at least in part by way of a collaborative filter.

8. The method of claim 1, wherein the second video clip is an advertisement.

9. The method of claim 8, further comprising:

assigning a bid for one or more ad impressions of the second video clip according to the likelihood score of an identified prospective viewer.

10. The method of claim 1, wherein causing the second video clip to be output to the display screen of the at least one prospective viewer in the target list of prospective viewers comprises automatically playing the second video clip when a webpage is viewed that offers the second video clip.

11. The method of claim 1, wherein causing the second video clip to be output to the display screen of the at least one prospective viewer in the target list of prospective viewers comprises providing the at least one prospective viewer an option to play the second video clip when a webpage is viewed that offers the second video clip.

12. The method of claim 1, further comprising:

adding, from the first list of viewers who watched the first video clip, subsequent viewer preferences related to the second video clip to the database to generate an updated second list.

13. The method of claim 1, wherein the viewer preferences related to the first video clip are derived by (a) loading a data collection agent to at least one of a player or a webpage and (b) monitoring viewing and behavior information of each viewer with the data collection agent.

14. A system for outputting a video clip to a display screen based on preference similarities comprising:

one or more servers; and

at least one non-transitory computer readable storage medium storing instructions thereon that, when executed by the one or more servers, cause the system to:

generate and store a database of viewer preferences related to a first video clip;

analyze the database of viewer preferences related to the first video clip to determine preference similarities among viewers in a first list of viewers who watched the first video clip;

record, in the database, viewer preferences related to a second video clip;

analyze the viewer preferences related to the second video clip to:

generate a third list of viewers in the database who have not yet viewed the second video clip and are likely to watch the second video clip based on preference similarities with viewers in both: (1) the first list of viewers who watched the first video clip and (2) the second list of viewers who watched the second video clip in the at least one of the plurality of positive manners;

wherein a likelihood determination for prospective viewers in the third list is made by:

analyzing at least one preference-related parameter associated with viewers,

weighing the at least one preference-related parameter,

comparing the likelihood score with a defined likelihood level; and

cause the second video clip to be output to a display screen of an electronic device associated with at least one viewer in the third list of viewers who are likely to watch the second video clip.

15. The system of claim 14, wherein the instructions, when executed, cause the system to generate and store the database of viewer preferences related to the first video clip by collecting viewer preferences at least one of before, during, or after the first video clip is played.

16. The system of claim 15, wherein the viewer preferences related to the first video clip reflect viewer actions taken prior to playing the first video clip comprising at least one of:

viewing a pre-roll advertisement,

accessing the first video clip by a particular method comprising at least one of: authenticating, logging-in, or being incentivized to see the first video clip,

adjusting a bit rate for streaming or downloading the first video clip prior to playing the first video clip, including a direction in which the viewer adjusted the bit rate of the first video clip, or

clicking play if the first video is not an auto-playing video clip.

17. The system of claim 15, wherein the viewer preferences related to the first video clip reflect viewer actions taken while playing the first video clip comprising at least one of:

adjusting volume level, muting, or un-muting the first video clip,

pausing or un-pausing the first video clip,

fast-forwarding or rewinding the first video clip,

replaying the first video clip, how many times the first video clip was replayed, and what portion of the first video clip was replayed,

viewing the first video clip for a recorded duration and percent completion,

viewing the first video clip in full-screen mode,

adjusting a bit rate for streaming or downloading the first video clip while playing the first video clip, including a direction in which the viewer adjusted the bit rate of the first video clip,

clicking-through on an overlay image or on the first video clip itself to be taken to another webpage or web site,

spending time viewing certain segments of the first video clip that are judged to have higher value based on actions of previous viewers, or

viewing a mid-roll advertisement.

18. The system of claim 15, wherein the viewer preferences related to the first video clip reflect viewer actions taken after playing the first video clip comprising at least one of:

sharing the first video clip via an embedded code,

sharing the first video clip via e-mail,

sharing the first video clip via a social network,

sharing the first video clip in multiple separate events, where a number of separate sharing events initiated by the viewer to share the first video clip are recorded,

waiting for a time duration between viewing the first video clip and initiating a sharing event, wherein said time duration is recorded,

bookmarking the first video clip for later viewing,

downloading and saving the first video clip for later viewing, or

subscribing to a channel or feed for a first video clip content producer or artist.

19. The system of claim 14, further comprising instructions that, when executed by the one or more servers, cause the system to:

assign a bid for an opportunity to present an ad impression of the second video clip to the at least one viewer from the third list of viewers according to the likelihood score.

20. The system of claim 19, wherein the system is one of: a client-based system, a server-based system, or a network-based system.

21. A non-transitory computer readable storage medium including a set of instructions that, when executed by at least one processor, cause a computing device to:

generate and store, in the computer readable storage medium, a database of viewer preferences related to a first video clip;

record in the database of viewer preferences related to a second video clip;

analyze the viewer preferences related to the second video clip to:

identify a second list of viewers who watched the second video clip and who viewed the second video clip in at least one of a plurality of positive manners;

produce a third list of viewers in the database who have not yet viewed the second video clip and are likely to watch the second video clip based on preference similarities with viewers in both: (1) the first list of viewers of the first video clip, and (2) the second list of viewers of the second video clip who viewed the second video clip in the at least one of the plurality of positive manners; and

causing the second video clip to be output to a display screen of at least one viewer in the third list of viewers who are most likely to watch the second video clip,

wherein a likelihood determination for viewers in the third list who are likely to watch the second video clip is made by:

analyzing at least one preference-related parameter associated with viewers,

weighting the at least one preference-related parameter,

developing a likelihood score based on the weighted at least one preference-related parameter, and

comparing the likelihood score with a defined likelihood level.

22. The non-transitory computer readable storage medium of claim 21, wherein the set of instructions, when executed by the at least one processor, cause the computing device to generate and store the database of viewer preferences related to the first video clip by passively collecting viewer preferences without requesting a specific viewer input for rating the first video clip.

23. The non-transitory computer readable storage medium of claim 22, wherein viewer preferences related to a corresponding video clip, wherein the corresponding video clip comprises the first video clip or the second video clip, reflect viewer actions taken prior to playing the corresponding video clip comprising at least one of:

viewing a pre-roll advertisement,

accessing the corresponding video clip by a particular method comprising at least one of: authenticating, logging-in, or being incentivized to see the corresponding video clip,

24. The non-transitory computer readable storage medium of claim 22, wherein viewer preferences related to a corresponding video clip, wherein the corresponding video clip comprises the first video clip or the second video clip, reflect viewer actions taken while playing the corresponding video clip comprising at least one of:

adjusting volume level, muting, or un-muting the corresponding video clip,

pausing or un-pausing the corresponding video clip,

fast-forwarding or rewinding the corresponding video clip,

viewing the corresponding video clip in full-screen mode,

adjusting a bit rate for streaming or downloading the corresponding video clip while playing the corresponding video clip, including a direction in which the viewer adjusted the bit rate of the corresponding video clip,

viewing a mid-roll advertisement.

25. The non-transitory computer readable storage medium of claim 22, wherein viewer preferences related to a corresponding video clip, wherein the corresponding video clip comprises the first video clip or the second video clip, reflect viewer actions taken after playing the corresponding video clip comprising at least one of:

sharing the corresponding video clip via an embedded code,

sharing the corresponding video clip via e-mail,

sharing the corresponding video clip via a social network,

sharing the corresponding video clip in multiple separate events, where a number of separate sharing events initiated by the viewer to share the corresponding video clip are recorded,

bookmarking the corresponding video clip for later viewing,

downloading and saving the corresponding video clip for later viewing, or

26. The non-transitory computer readable storage medium of claim 21, wherein the set of instructions, when executed by the at least one processor, cause the computing device to analyze the database of viewer preferences to determine preference similarities by determining preference similarities at least in part with a collaborative filter.

27. The non-transitory computer readable storage medium of claim 21, wherein the second video clip is a video advertisement.

28. The non-transitory computer readable storage medium of claim 27, wherein the set of instructions, when executed by the at least one processor, further cause the computing device to:

assign a bid value for an ad impression of the video advertisement to be offered to an identified prospective viewer, the assignment of the bid value based on a relative likelihood that the identified prospective viewer will watch the video advertisement.