You can configure threshold weight values (e.g. minWeight or maxWeight) for filters with 'keyword', 'user-keyword' or 'user-interest' type(see traffic filters). For user segments these values will determine if the user should be included or not, according to the interest area of the user, which is based on the articles that the user has read. In regard to contextual segments, the weight indicates how relevant the respective 'keyword' is in relation to the context of the page. --- not sure about this: Here the user interest is out of scope as the content itself will be added to the contextual segment.
The presence of the Piano Audience JavaScript snippet on your pages will cause the Piano web robot to crawl those pages on a regular basis. The crawled pages are subsequently parsed, and the text extracted from each page is then semantically analyzed and summarized to produce a structured dataset, the content profile. A content profile can thus be thought of a kind of "fingerprint" that conveys the topic and theme of the page. Logically, a content profile can be thought of as a set of triples:
-
An item: A string of some kind that carries some interest. Often this is a normalized version of something actually found on the page, but it could equally well be a category name or something else.
-
A weight: A number that somehow reflects the relative prominence of the item on the page, as measured via well-defined concepts from information retrieval.
-
A group name: A string that somehow indicates the type of the item. This typically identifies how to semantically interpret the item. For example, an item john smith might have the group name person, indicating that it's a person name.
Content profiles are used to build user interest profiles. These profile are based on the user's interaction with the site. As the user consumes pages, their interest profile will be updated by incorporating a suitably weighted version of the content item that was consumed. This "suitably weighted" takes into account several factors. E.g., on a general basis the following holds:
-
Recently consumed items account for more than content items consumed a long time ago.
-
Content items that are generally popular and consumed by "everybody" accounts for less than content items that are consumed by fewer people.
-
The time spent on consuming a content item may influence the weighting.
-
Different profile groups carry different weights, e.g., some profile elements are forgotten quickly whereas others have a longer lifespan.
When building contextual segments, you can directly access the weighting reflected in the content profile by leveraging the minimum interest level in the segment builder.
-
The higher the weighting is set when creating segments, the more relevance the keyword in question has compared to the context of the given page.
-
In case you are using a higher weighting, it's expected that fewer pages are included in your contextual segments, as pages with lower relevance are excluded.