Search relevancy
Getting the relevant results is critical for the success of the community. Here you can see what parameters impact the relevancy score for a piece of content and the rank it will get when you search for a specific search phrase.
The relevancy rank is calculated as follows:
Rank = (SimilarityScore + ProximityScore) * OutcomeType * ObjectType * Recency * SocialScore
The parameters that impact the rank of a content item are described below.
Similarity score
When searching for a phrase, the system looks at each word in the phrase and checks the match type and place of match for this word. Each match type and place has its own boost score.
Match types that the Cloud Search employs include:
- Raw: Exact matches of the search term.
- Analyzed: Matches created by language analysis. In this case, stemming is used, meaning it looks for the root of the word.
- Edgengram: Partial matches used for wildcard searches.
Places of matches the Cloud Search employs:
- Subject: Title field of content items
- Body: Content of content items
- Tags: Tags added to content items
| Match type | Subject | Body | Tags |
|---|---|---|---|
| Raw | 1.0 | 0.1 | 0.5 |
| Analyzed | 1.0 | 0.1 | |
| Edgengram | 1.0 | 0.1 | 0.5 |
Proximity score
The proximity score checks how close the searched term is to what appears in the content. It ensures that content with terms appearing together gets boosted more than those appearing far apart.
Types of proximity boosts include:
- Exact match: All search terms appear next to each other.
- Proximity match: All search terms appear less than three words apart.
| Place | Proximity boost | Exact match boost |
|---|---|---|
| Subject | 0.5 | 1.6 |
| Body | 0.5 | 1.0 |
| Tags* | 0.1 | 1.0 |
* Having a proximity score on Tags is unlikely to happen.
Object type (type_weight)
Object type boost is based on the type of content used. Documents and blogs are ranked higher in the search results as these are typically used for more comprehensive content that may be more relevant for the searching user. The settings are listed in Table 1.
| Object | Boost |
|---|---|
| Document | 1.4 |
| Blog | 1.4 |
| Poll | 1.0 |
| Idea | 1.0 |
| Discussion | 1.0 |
| Video | 1.0 |
| Question | 1.0 |
| Status Update | 1.0 |
Recency (recency_weight)
Recency (or time decay) lowers the score for older content. The impact is described in the following table:
| Period | Boost |
|---|---|
| Less than 9 weeks ago | 1.0 |
| 9-55 weeks ago | 0.75 + (0.25 * (56 - <week_number>)/48) |
| 56-223 weeks ago | 0.5 + (0.25 * (224 - <week_number>)/168) |
| More than 223 weeks ago | 0.5 |
The Recency score also includes parameters that determine how quickly the algorithm reduces a content score by age, the latest period content maintains its score without decay, and the score difference of very old documents compared to fresh content.
Outcome type (outcome_weight)
Content in Jive can be marked with structured outcomes. The search results are boosted based on outcome type.
The boosts given to content according to outcome type are listed in Table 3.
| Outcome | Boost | Outcome | Boost |
|---|---|---|---|
| Finalized | 1.4 | Official | 2.0 |
| Outdated | 0.1 | Default | 1.0 |
This score is being multiplied by the number of boosts as follows: base*(1+0.01*<number_of_outcomes>)
Note that a higher boost results in that content being ranked higher in the search results, so the 0.1 score for outdated documents significantly reduces its rank.
Social score
The Jive R2E2 service calculates a social score for the search phrase based on given user activities, follows, and other behavioral connections. This service provides improved search relevance by incorporating social information into search, tailoring rankings based on user dynamics.
As users engage with content within Jive, their activities generate data that enhances the relevance of search results. Items that are considered "close" to the user based on their activities or connections are given a boost in the search rankings, resulting in personalized search results.
Rank count examples
For example, the rank may be counted as follows:
Type | Age | Outcome | Rank |
|---|---|---|---|
Blog | 3 days | None |
|
Blog | 10 weeks | None |
|
Document | 10 weeks | None |
|
Document | 56 weeks | Official |
|
Document | 256 weeks | Official, Finalized |
|
Question | 3 weeks | None |
|
Event | 10 weeks | None |
|
Discussion | 56 weeks | None |
|
Additional considerations
- The boosts cannot be changed with normal methods.
- Social scoring and personalized results are not available in On-Premise Search service.