Business operation report

更新时间:
复制 MD 格式

Description

OpenSearch provides six categories of statistical metrics for user application data: core metrics, traffic metrics, behavioral metrics, transaction metrics, user analysis metrics, and query analysis metrics.

Core metrics

Metric

Definition

Formula

Dependencies

Significance

Search PV

Page View (PV): The number of successful search requests. This count includes paging requests.

Each successful search request counts as one.

Measures the total volume of successful searches per day.

Search UV

Unique Visitor (UV): The number of unique users who perform a search.

Counts each unique user only once per day, regardless of the number of searches.

Requires the user_id parameter in search requests sent via the SDK or API.

Measures the daily number of unique users engaging with the search feature.

Paging rate

The ratio of paging requests to the total Search PV.

Number of paging requests / Search PV

Helps evaluate the effectiveness of search retrieval and sorting.

No result rate

The percentage of searches that return 0 items.

(Searches returning 0 items) / Search PV

Evaluates the effectiveness of search retrieval and content richness.

PV-CTR

The click-through rate (CTR) per search.

IPV / Search PV, where IPV stands for Item Page View.

Requires enabling the data collection feature and uploading user click behavior data.

Indicates the click activity on search results. This metric helps evaluate the effectiveness of search retrieval and sorting.

UV-CTR

The unique user click-through rate.

IPV_UV / Search UV, where IPV_UV is the number of unique users who clicked an item.

Requires enabling the data collection feature and uploading user click behavior data.

Indicates the click activity by unique users on search results. This metric helps evaluate the effectiveness of search retrieval and sorting.

Search PV per user

The average number of successful searches per user.

Search PV / Search UV

Requires the user_id parameter in search requests sent via the SDK or API.

A high value can indicate strong user interest, but may also suggest that users cannot find desired results with fewer queries.

Next day retention rate

The percentage of users from a given day who return to search on the following day.

(UV(t-1) ∩ UV(t)) / UV(t-1)

Requires the user_id parameter in search requests sent via the SDK or API.

Indicates user stickiness for the search feature.

Traffic metrics

Metric

Definition

Calculation

Dependencies

Description

search PV

The total number of successful search requests, including pagination requests.

Each successful search request counts as 1.

The total number of successful searches performed on a given day.

search UV

The number of unique users who perform a search.

Counts each unique user once, regardless of the number of visits.

Requires setting the user_id parameter in search requests sent via SDK or API.

The number of unique users who performed a search on a given day.

search PV per user

The average number of successful search requests per user.

search PV / search UV

Requires setting the user_id parameter in search requests sent via SDK or API.

The average number of searches per user. A high value can indicate strong user interest, but may also suggest that users are not finding their desired results quickly.

item impressions

The number of times an item (such as a product or content on a search results page) is displayed. If an item reappears on a subsequent page due to pagination, it is counted as a new impression. Impressions are not deduplicated.

Counts each impression as 1 if data collection is enabled and impression data is uploaded. Otherwise, it defaults to counting the number of items returned in each search response.

Accurate tracking requires enabling data collection and uploading user impression data. Without this data, the metric's accuracy is reduced.

Total item impressions from search results on a given day.

number of queries

The total count of queries, deduplicated per user.

The sum of each user's unique queries. For example, if two different users search for the same query, the total count is 2.

Requires setting the user_id parameter in search requests sent via SDK or API. The query is sourced from the raw_query parameter if available; otherwise, it uses the value from the default index. If the default index does not exist, the query is considered empty.

Represents the total number of user search queries, adjusted to reduce the impact of repeated searches and pagination.

queries per user

The average number of unique queries per user.

number of queries / search UV

Requires setting the user_id parameter in search requests sent via SDK or API.

The average number of unique queries per user. This metric is not affected by repeated searches or pagination.

number of independent queries

The total count of distinct query strings that returned at least one result.

The total number of unique queries after deduplication across all users. For example, a query submitted by multiple users counts only once.

The query is sourced from the raw_query parameter if available; otherwise, it uses the value from the default index. If the default index does not exist, the query is considered empty. This metric does not require a user_id.

Indicates the richness and diversity of query terms used by the user base.

pagination rate

The percentage of total search PV accounted for by pagination requests.

Pagination request PV / search PV

Measures the proportion of search activity that involves pagination. This helps evaluate search result quality and ranking.

no-result rate

The percentage of searches that return zero results.

(Number of searches with zero results) / (Total search PV)

Helps evaluate the quality of search results and content coverage.

low-result rate

The percentage of searches that return 5 or fewer items.

(Number of searches with ≤5 results) / (Total search PV)

Helps evaluate the quality of search results and content coverage.

number of request errors

The total number of failed search requests, for reasons such as throttling.

Counts each failed request as 1.

Indicates if the application requires scaling out or other adjustments.

request error rate

The percentage of search requests that failed, for reasons such as throttling.

number of request errors / (search PV + number of request errors)

Indicates if the application requires scaling out or other adjustments.

Behavioral metrics

Metric

Definition

Calculation

Dependencies

Significance

IPV

The total number of clicks on items.

Each click is counted as 1.

Requires enabling the data collection feature and uploading user click behavioral data.

Measures the total click volume on search results.

IPV_UV

The number of unique users who clicked an item on a search results page.

A user is counted only once, regardless of the number of clicks.

Requires enabling the data collection feature and uploading user click behavioral data. Note: You must specify the user_id parameter when uploading data.

Measures the number of unique users interacting with search results.

IPV per capita

The average number of item clicks per user on the search results page.

IPV / Search UV

Requires enabling the data collection feature and uploading user click behavioral data. Note: You must specify the user_id parameter when uploading data.

Measures the average user engagement level with search results.

PV-CTR

The click-through rate (CTR) based on search page views.

IPV / Search PV

Requires enabling the data collection feature and uploading user click behavioral data.

Measures click activity on search results to evaluate the effectiveness of search retrieval and sorting.

UV-CTR

The click-through rate (CTR) based on unique visitors.

IPV_UV / Search UV

Requires enabling the data collection feature and uploading user click behavioral data.

Measures click activity by unique users on search results to evaluate the effectiveness of search retrieval and sorting.

Item-CTR

The click-through rate (CTR) for items (such as products or content) on the search results page.

IPV / Number of item impressions

Requires enabling the data collection feature and uploading user click behavioral data.

Measures the click performance of individual items to evaluate the effectiveness of search retrieval and sorting.

Clicked PV percentage

The percentage of search page views (PVs) that receive at least one click.

Search PV with clicks / Total Search PV

Requires enabling the data collection feature and uploading user click behavioral data.

Indicates the percentage of searches that result in a click, gauging overall result page effectiveness.

TOP3 PV-CTR

The click-through rate (CTR) of the top three items (clicks on subsequent pages are not counted).

IPV of top 3 items / Search PV

Requires enabling the data collection feature and uploading user click behavioral data.

Measures click activity on the top three items to evaluate the effectiveness of search sorting.

TOP5 PV-CTR

The click-through rate (CTR) of the top five items (clicks on subsequent pages are not counted).

IPV of top 5 items / Search PV

Requires enabling the data collection feature and uploading user click behavioral data.

Measures click activity on the top five items to evaluate the effectiveness of search sorting.

TOP10 PV-CTR

The click-through rate (CTR) of the top ten items (clicks on subsequent pages are not counted).

IPV of top 10 items / Search PV

Requires enabling the data collection feature and uploading user click behavioral data.

Measures click activity on the top ten items to evaluate the effectiveness of search sorting.

Number of favorites

The total number of times users add an item to their favorites directly from search results.

Each favorite action is counted as 1.

Requires enabling the data collection feature and uploading behavioral data for bhv_type="collect" events.

Measures how effectively search results drive users to favorite items.

Favorite conversion rate

The rate at which search page views convert to "favorite" actions.

Number of favorites / Search PV

Requires enabling the data collection feature and uploading behavioral data for bhv_type="collect" events.

Measures the effectiveness of search retrieval and sorting, and their impact on driving favorite actions.

Number of add-to-carts

The total number of times users add an item to a cart directly from search results.

Each add-to-cart action is counted as 1.

Requires enabling the data collection feature and uploading behavioral data for bhv_type="cart" events. This action can include adding to a shopping cart, bookshelf, or playlist.

Measures how effectively search results drive users to add items to their cart.

Add-to-cart conversion rate

The rate at which search page views convert to "add-to-cart" actions.

Number of add-to-carts / Search PV

Requires enabling the data collection feature and uploading behavioral data for bhv_type="cart" events. This action can include adding to a shopping cart, bookshelf, or playlist.

Measures the effectiveness of search retrieval and sorting, and their impact on driving add-to-cart actions.

Number of likes

The total number of times users like an item directly from search results.

Each like action is counted as 1.

Requires enabling the data collection feature and uploading behavioral data for bhv_type="like" events.

Measures how effectively search results drive users to like items.

Like conversion rate

The rate at which search page views convert to "like" actions.

Number of likes / Search PV

Requires enabling the data collection feature and uploading behavioral data for bhv_type="like" events.

Measures the effectiveness of search retrieval and sorting, and their impact on driving like actions.

Number of comments

The total number of times users post a comment on an item directly from search results.

Each comment is counted as 1.

Requires enabling the data collection feature and uploading behavioral data for bhv_type="comment" events.

Measures how effectively search results drive users to comment on items.

Comment conversion rate

The rate at which search page views convert to "comment" actions.

Number of comments / Search PV

Requires enabling the data collection feature and uploading behavioral data for bhv_type="comment" events.

Measures the effectiveness of search retrieval and sorting, and their impact on driving comment actions.

Transaction metrics

Metric

Definition

Formula

Dependency

Description

search-guided GMV

The total gross merchandise volume (GMV) from search-driven purchases.

Sum of transaction amounts.

Requires enabling the data collection feature and uploading behavioral data for bhv_type="buy" events.

Evaluates how effectively search retrieval and sorting drive purchases.

search-guided purchase UV

The number of unique users who made a search-driven purchase.

Each unique user is counted only once, regardless of the number of purchases made.

Requires enabling the data collection feature and uploading behavioral data for bhv_type="buy" events with the user_id parameter specified.

Evaluates how effectively search retrieval and sorting drive purchases.

search-guided average purchase value

The average purchase value per user for search-driven purchases.

search-guided GMV / search-guided purchase UV

Requires enabling the data collection feature and uploading behavioral data for bhv_type="buy" events.

Evaluates how effectively search retrieval and sorting drive purchases.

order conversion rate

The percentage of unique search users who make a purchase.

search-guided purchase UV / search UV

Requires enabling the data collection feature and uploading behavioral data for bhv_type="buy" events.

Evaluates how effectively search retrieval and sorting drive purchases.

click-to-purchase conversion rate

The percentage of users who click on a search result and then complete a purchase.

search-guided purchase UV / IPV_UV

Requires enabling the data collection feature and uploading behavioral data for bhv_type="buy" events.

Measures how effectively a clicked search result leads to a purchase, reflecting both item relevance and user intent.

User analytics

Metric

Definition

Formula

Dependency

Description

Next day retention rate

The percentage of users who search on one day and return to search the next day.

(UV(t-1) ∩ UV(t)) / UV(t-1)

Requires setting the user_id parameter in search requests sent via SDK/API.

Indicates user stickiness for the search feature.

Zero-click rate

The percentage of users who search but do not click any results.

(Search UV - IPV_UV) / Search UV

Requires enabling the data collection feature and uploading user click behavior data. Note: Data uploads must include the user_id parameter.

Measures the effectiveness of search retrieval and ranking.

Query analysis metrics

Dimension

Metric

Definition

Formula

Dependencies

Significance

High-frequency queries

Query

The top 500 queries by search frequency.

Selects the top 500 queries based on search frequency.

Depends on search requests from the SDK or API. If a request includes the raw_query parameter, its value is used. Otherwise, the value from the default index is used. If no default index exists, the query is empty.

Identifies the most popular user queries.

Search PV

The total number of searches for this query.

Each search is counted as 1.

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

Measures the search volume for the most popular queries.

Average number of exposed items

The average number of items exposed per search for this query.

Total number of items returned for this query / Search PV

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

Shows the item exposure for the most popular queries. This helps measure retrieval effectiveness and content richness.

PV-CTR

The average click-through rate (CTR) per search for this query.

Total clicks for this query / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Percentage of PV with Clicks

The percentage of searches for this query that result in at least one click.

PV with clicks / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

TOP3 PV-CTR

The CTR for the top three items on the first page of search results. Clicks on subsequent pages are not included.

Total clicks on the top three items / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on the top three items, which helps measure sorting effectiveness.

TOP5 PV-CTR

The CTR for the top five items on the first page of search results. Clicks on subsequent pages are not included.

Total clicks on the top five items / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on the top five items, which helps measure the effectiveness of sorting.

TOP10 PV-CTR

The CTR for the top ten items on the first page of search results. Clicks on subsequent pages are not included.

Total clicks on the top ten items / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on the top ten items, which helps measure the effectiveness of sorting.

High-frequency queries with low PV-CTR

Query

The 100 queries with the lowest PV-CTR among the top 1,000 most frequent queries.

From the 1,000 most frequent queries, this identifies the 100 with the lowest "Percentage of PV with Clicks".

Depends on search requests from the SDK or API, and requires enabling the data collection feature and uploading user click data.

Identifies popular queries with low user engagement.

Search PV

The total number of searches for this query.

Each search is counted as 1.

Depends on search requests from the SDK or API, and requires enabling the data collection feature and uploading user click data.

Identifies popular queries that have low user engagement with the search results.

Average number of exposed items

The average number of items exposed per search for this query.

Total number of items returned for this query / Search PV

Depends on search requests from the SDK or API, and requires enabling the data collection feature and uploading user click data.

Helps diagnose if low engagement is due to poor item exposure, measuring retrieval effectiveness and content richness.

PV-CTR

The average CTR per search for this query.

Total clicks for this query / Search PV

Depends on search requests from the SDK or API, and requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Percentage of PV with Clicks

The percentage of searches for this query that result in at least one click.

PV with clicks / Search PV

Depends on search requests from the SDK or API, and requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Queries with few results

Query

Queries that return five or fewer items.

The top 100 most frequent queries that return five or fewer items.

Depends on search requests from the SDK or API. If a request includes the raw_query parameter, its value is used. Otherwise, the value from the default index is used. If no default index exists, the query is empty.

Identifies queries with few results to help measure retrieval effectiveness and content richness.

Search PV

The total number of searches for this query.

Each search is counted as 1.

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

Measures the search volume for queries that return few results to help assess the severity of the issue.

Average number of exposed items

The average number of items exposed per search for this query.

Total number of items returned for this query / Search PV

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

Measures the exact item exposure for these low-result queries, which helps confirm retrieval issues.

PV-CTR

The average CTR per search for this query.

Total clicks for this query / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Percentage of PV with Clicks

The percentage of searches for this query that result in at least one click.

PV with clicks / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Queries with no results

Query

Queries that return zero items.

The top 100 most frequent queries that return zero items.

Depends on search requests from the SDK or API. If a request includes the raw_query parameter, its value is used. Otherwise, the value from the default index is used. If no default index exists, the query is empty.

Identifies queries with no results to help measure retrieval effectiveness and content richness.

Search PV

The total number of searches for this query.

Each search is counted as 1.

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

Measures the search volume for queries that return no results to help assess the severity of the issue.

Fastest-rising queries

Query

Queries with the highest growth rate compared to the previous day, sorted in descending order. The minimum Search PV from the previous day must be the greater of 10 or 0.1% of the total Search PV.

The top 100 queries with the largest increase in search frequency compared to the previous day.

Depends on search requests from the SDK or API. If a request includes the raw_query parameter, its value is used. Otherwise, the value from the default index is used. If no default index exists, the query is empty.

Identifies recent trending and emerging queries.

Previous Day's Search PV

The Search PV for this query on the previous day.

Each search is counted as 1.

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

The search volume from the previous day for trending or new queries.

Current Day's Search PV

The Search PV for this query on the current day.

Each search is counted as 1.

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

The search volume from the current day for trending or new queries.

Growth Rate

Measures the rate of increase in a query's popularity.

(Current Day's Search PV - Previous Day's Search PV) / Previous Day's Search PV

The growth rate for a trending query on the current day.

Average number of exposed items

The average number of items exposed per search for this query.

Total number of items returned for this query / Search PV

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

Shows the item exposure for recent trending and emerging queries. This helps measure retrieval effectiveness and content richness.

PV-CTR

The average CTR per search for this query.

Total clicks for this query / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Percentage of PV with Clicks

The percentage of searches for this query that result in at least one click.

PV with clicks / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Fastest-declining queries

Query

Queries with the highest decrease rate compared to the previous day, sorted in descending order. The minimum Search PV on the current day must be the greater of 10 or 0.1% of the total Search PV.

The top 100 queries with the largest decrease in search frequency compared to the previous day.

Depends on search requests from the SDK or API. If a request includes the raw_query parameter, its value is used. Otherwise, the value from the default index is used. If no default index exists, the query is empty.

Identifies queries that are rapidly decreasing in search frequency.

Previous Day's Search PV

The Search PV for this query on the previous day.

Each search is counted as 1.

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

The search volume from the previous day for queries that are rapidly decreasing in frequency.

Current Day's Search PV

The Search PV for this query on the current day.

Each search is counted as 1.

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

The search volume from the current day for queries that are rapidly decreasing in frequency.

Decrease Rate

Measures the rate of decrease in a query's popularity.

(Previous Day's Search PV - Current Day's Search PV) / Previous Day's Search PV

Average number of exposed items

The average number of items exposed per search for this query.

Total number of items returned for this query / Search PV

Depends on search requests from the SDK or API. Uses the raw_query parameter if provided.

Shows the item exposure for queries that are rapidly decreasing in frequency. This helps measure retrieval effectiveness and content richness.

PV-CTR

The average CTR per search for this query.

Total clicks for this query / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Percentage of PV with Clicks

The percentage of searches for this query that result in at least one click.

PV with clicks / Search PV

Requires enabling the data collection feature and uploading user click data.

Indicates click activity on search results, which helps measure the effectiveness of retrieval and sorting.

Note

Note:

1. If the data collection feature is disabled for your application, only traffic metrics and query analysis metrics are available.

2. Ensure the user_id has the same format in both your behavioral data and search requests.

Usage

  1. By default, the single-day view displays metrics from the previous day. Use the date picker control to select a different date. Reports are generated on a T+1 basis. This means that after you upload behavioral data, the report becomes available the next day.

  2. By default, a line chart displays the trend of each metric over the last seven days. Use the date picker control to select a different time period. To customize the chart, click a metric's icon to include or exclude it from the view.

  3. To view the report, go to Report Statistics > business operation report in the navigation pane. Select an application, business scene, metric category, and a time period to view the statistical charts.

The report page has the following tabs: Overall trend, Traffic, Behavior, Transactions, User analysis, and Query analysis. By default, the Overall trend tab displays the previous day's Core metrics card (search PV, search UV, pagination rate, no-result rate, PV-CTR, UV-CTR, and week-over-week change) and Top search terms list, along with a line chart of the 7-day search PV trend.