Abstract: Disclosed are various embodiments for searching for items in a catalog using normalized item attributes. A range of values on a normalized scale is determined for an option for an attribute associated with a representative item. The range of values of the representative item is correlated with a first unit on a first scale of item classification. Multiple items are identified based at least in part on the option for the attribute. Each of the items is associated with a respective option for the attribute that has a respective range of values on the normalized scale that overlaps the range of values of the representative item. The respective range of values of one or more of the items is correlated with a second unit on a second scale of item classification.

Abstract: Disclosed are various embodiments for selecting users to whom to promote an online offering, such as an auction or a fixed-price item offering, based at least in part on bidding history or purchase history. A data representation of multiple items offered online by a web merchant and a purchase history associated with the items and the web merchant for multiple users are maintained. A first item is identified in response to determining that the number of users who have purchased a second item and who have also purchased the first item meets a minimum threshold. A set of users who have both purchased the second item and also purchased the first item is identified. The first item is promoted to one or more of the identified users.

Abstract: A recommendation system divides users into a plurality of segments based on scores generated for each of a plurality of users of an electronic service that provides functionality for users to browse and purchase or select items. The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent. The recommendation system can further detect behavior-based correlations between particular segments and particular items and/or item attributes. The correlations can be based on purchase or selection histories of the users assigned to specific segments. The recommendation system can select items to recommend to particular users based at least partly on the detected behavior-based correlations.

Abstract: A user interface that facilitates the selection and purchase of an item may be presented on a display. The user interface includes a presentation section and one or more other sections, such as a title section, buy section, alternate image section, package art section, review section, rating section, or other section. The presentation section may extend a full-width of a display or window available for display, while maintaining a desired aspect ratio. Video, slideshows, or other visual information may be played within the presentation section. One or more sections may overlay the presentation section. For example, at least a portion of the buy section may overlay at least a portion of the presentation section.

Abstract: A computer-implemented service integrates with merchant sites to enable registered users of the service to perform purchase transactions on the merchant sites. In one embodiment, a merchant can incorporate a reference to a display object served by the service into coding of a page of the merchant's site. When the page is loaded by a user's browser, the reference causes the browser to request the display object from the service for display on the page. If the user is recognized by the service, the service may return a personalized display object that includes functionality for the user to perform a purchase transaction using the service. Also disclosed are embodiments that enable users to initiate transfers of personal account information to the merchant sites.

Abstract: Disclosed are various embodiments relating to generating profile-based item suggestions from a stylist when purchasing items through an electronic commerce system. A network page is rendered in a client computing device comprising an input field to input descriptions of items the user wishes to purchase and a stylist profile selection mechanism to select a suggested stylist profile. The user inputs descriptions of items and selects the suggested stylist profile. An item suggestions request comprising the item descriptions and the selected stylist profile is sent from the client computing device to a server computing device.

Abstract: A computer-implemented service integrates with merchant sites to enable registered users of the service to perform purchase transactions on the merchant sites. In one embodiment, a merchant can incorporate a reference to a display object served by the service into coding of a page of the merchant's site. When the page is loaded by a user's browser, the reference causes the browser to request the display object from the service for display on the page. If the user is recognized by the service, the service may return a personalized display object that includes functionality for the user to perform a purchase transaction using the service. Also disclosed are embodiments that enable users to initiate transfers of personal account information to the merchant sites.

Abstract: Disclosed are various embodiments relating to generating media item recommendations based upon determining a beats per minute value. A media recommendation application determines a beats per minute value based upon the activity of a user. The media recommendation application then identifies one or more media item candidates having a number of beats per minute within a predefined delta of the beats per minute value. The media recommendation application then filters out the media items possessed and information associated with the user to identify a subset of the media item candidates to surface as recommendations.

Abstract: An interactive system generates personalized item recommendations for users based partly or wholly on ratings assigned by the users to particular items. The system includes an item rating user interface that enables a user to view, prior to rating an item, information regarding the types of items that will be recommended to the user if the user assigns a particular rating or type of rating to the item. The user interface thereby enables users to refrain from performing rating actions that will tend to result in low utility or “poor quality” recommendations from the users' perspectives.

Abstract: A computing system for searching for and identifying data repository deficits includes an interactive database of items, a data repository that stores a first type of information related to items, a search engine configured to identify items that do not have an amount of the first type of information that meets a first threshold, a candidate selection interface that provides functionality for the user to select an item from recommended candidate items needing the first type of information, a user interface via which the user can provide the first type of information for an item selected by the user, a game system configured to calculate for a game scores and rankings based at least in part on data stored in the data repository.

Abstract: A system for presenting a set of recommended items corresponding to a first item is presented. The system comprises a processor, a memory; and a computer executable component. The computer executable component executes on the system, via the processor, and is configured to perform the following. An indication is received requesting the display of recommended items corresponding to the first item. A set of recommended items corresponding to the first item is then obtained. Display information is generated for displaying a list of items and the corresponding set of recommended items. The list of items includes the first item. Additionally, display information is provided, the display information for displaying the list of items and the corresponding set of recommended items on a consumer device.

Abstract: Aspects of the present disclosure relate to machine learning techniques for identifying collections of items that are visually complementary. These techniques can relying on computer vision and item imagery. For example, a first portion of a machine learning system can be trained to extract aesthetic item qualities from pixel values of images of the items. A second portion of the machine learning system can learn correlations between these extracted aesthetic qualities and the level of visual coordination between items. Thus, the disclosed techniques use computer vision machine learning to programmatically determine whether items visually coordinate with one another based on pixel values of images of those items.

Abstract: Disclosed are various embodiments for presenting alternative shopping options to a customer. A collection of alternative shopping options is identified in a computing device based upon subject matter of a network page generated by a first one of a plurality of network sites to be presented to a customer. Each of the network sites is configured to sell a plurality of products, and at least one of the alternative shopping options comprises a purchase opportunity associated with a second one of the network sites. An estimate of a benefit for each of the alternative shopping options is determined with respect to the first one of the network sites, and a subset of the alternative shopping options is identified to be included in the network page based at least upon the estimates. The subset of the alternative shopping options is included in the first network page.

Abstract: Disclosed are various embodiments for presenting alternative shopping options to a customer. A collection of alternative shopping options is identified in a computing device based upon subject matter of a network page generated by a first one of a plurality of network sites to be presented to a customer. Each of the network sites is configured to sell a plurality of products, and at least one of the alternative shopping options comprises a purchase opportunity associated with a second one of the network sites. An estimate of a benefit for each of the alternative shopping options is determined with respect to the first one of the network sites, and a subset of the alternative shopping options is identified to be included in the network page based at least upon the estimates. The subset of the alternative shopping options is included in the first network page.

Abstract: Purchase phrases can be analyzed to infer user interests. A purchase phrase can include one or more user-defined textual words that a user employs to purchase items in an expedited manner. Many users' purchase phrases may include words that reflect the users' interests, such as “golf” or “rose.” Users with similar purchase phrases may therefore be interested in purchasing similar items. A recommendation system can therefore find items purchased with similar phrases to a target user's purchase phrase and then recommend these items to the target user.

Abstract: When a user selects an item at an online marketplace and subsequently provides a search query to the online marketplace, search results or recommendations received from the online marketplace in response to the search query may be displayed based on whether the search results or recommendations would satisfy a requirement established by a source of the selected item. The requirement may relate to a minimum cost threshold, or a minimum number of items. Any search results or recommendations that would, if selected, cause the requirement to be met or exceeded may be preferentially displayed to the user above, or in a more prominent manner, than search results or recommendations that would not, if selected, cause the requirement to be met or exceeded.

Abstract: Improved user interfaces and methods are provided for presenting item recommendations to a user when the user selects an item to add to an electronic shopping cart. In response to the user's selection, a page generation process generates and returns a page that includes a condensed shopping cart portion and a recommendations portion. The condensed shopping cart portion displays the contents of the shopping cart, but omits some of the item information and/or controls provided on a shopping cart page, thereby preserving more screen real estate for the display of item recommendations. The condensed shopping cart portion also includes a link to the shopping cart page.

Abstract: Disclosed are various embodiments that provide file system merchandising. A file service maintains file systems associated with file service accounts. When accessing content of the file system through a network page, the network page includes a merchandising recommendation that is based at least in part upon characteristic information associated with the requested content. The merchandising recommendation may also be based at least in part upon account behavior history associated with the file service account. A link may be included in the merchandising recommendation to initiate previewing of at least a portion of a recommended item and/or to initiate acquisition of the recommended item.

Abstract: Guided item selection based on profiles maintained dynamically. The profiles represent, for example, purchase, usage, or personality profiles corresponding to various segments of the user population. Each of the profiles has one or more items associated with the profile. The user selects one or more of the profiles while browsing or searching for items. The items associated with the selected profile are provided to the user. The user's interaction, or lack of interaction, with the items is monitored and analyzed. Based on the monitored interaction, a relevance of the selected profile relative to the other profiles is dynamically adjusted along with a relevance of particular items for the selected profile. For example, items may be added or removed from the list based on the monitored user interactions.

Abstract: Techniques and systems may be used to generate a list of interests of a user that interacts with a catalog of items, such as by purchasing the items from a host. The host may then generate a list of interests of the user using a taxonomy that is mapped to the catalog of items and the user interaction with the items. By generating the list of interests based on user interaction with the catalog, the list of interests may be generated based on factual data rather than user opinion. However, in some instances, the user may additionally provide a self-rating for an interest that is identified by the host. In various aspects, the list of interests may be associated with a reviewer of items of the catalog. In some aspects, a reader may identify reviewers that have similar interests by comparing reviewers' interests to the reader's interests.