Introduce how to display the ranking of the number of purchases / purchase price on the site.
- It is assumed that the product master has been imported to Datahub.
- If you have n’t imported yethere.Please import the product master referring to the article.
Set up the query
- Datahub administration screen → The screen changes to "Query" screen.
- Click "Create"-"Create from Collection" on the top left
- やりたいことから探す→「ランキングを表示したい」から「商品購入数/購入金額のランキング」を選択
- Set the following from parameter setting
- Extraction period (default 7 days)
- Data set ID of product master
- Product master table ID
- Record lifetime (h) (default is 24 hours) when linking action tables
- reference:Delete a record from the action table
- Check whether the result is returned by executing “Query” in the upper right.
Make sure that the query results are within 200,000 records. (The maximum number of records in the action table is 200,000 records, so data cannot be exported to the action table if it exceeds 200,000 records.)
Checkpoints when query results are not returned
- Is the event name and schema sent in the KARTE tag set correctly?
- In realizing the recommendation,
view_item
、buy
The event must be sent to KARTE correctly.- Check if the event is sent correctly from the global menu “Knowing Customers” → “Events”.
- In realizing the recommendation,
- Does it conform to the product master format?
- Check if field names and data types are linked according to the format.
- The product master format ishere.
- Check if field names and data types are linked according to the format.
- Is the linked data value correct?
- You can check the imported data from “Preset”, “Table information”, etc. of Dataset → Imported table.
Create action table
- Click on "Create" → "Create empty table" at the top right of the action table of the global menu
- Make the following settings and click "Create"
- Table ID: Specify any table name (in the example, ranking_weekly)
- Schema setting: Specify the schema of the query result as follows
- Authority: Since this time only for reference, check "Reference from widget"
- mode:
- No filtering or filtering with a single condition: general
- When filtering by multiple conditions: flexible
- reference:Action table mode
Schema setting of action table
Field name | Type |
---|---|
item_id (primary key) | string |
name | string |
brand_name | string |
price | Number |
display_price | string |
image | string |
url | string |
l_category_cd | string |
s_category_cd | string |
brand_cd | string |
sex_cd | string |
quantity_rank | Number |
amount_rank | Number |
expired_at | Timestamp |
Export query results to action table
- Click "Create" in the upper right of the job flow screen on the global menu
- Set the following from the job setting screen, click "Save and complete" in the upper right
Job settings
item | Contents |
---|---|
Job name | Weekly ranking results to action table |
Job type | Import from data table |
Export source | The query result created earlier |
Export destination | Select the action table created earlier |
Select execution method
- Check "Enable schedule execution"
- Check "Enable repetitive execution"
- Set execution frequency
- We recommend weekly degree of coordination frequency.
Job Flow Settings
- "Archive after job execution is finished" is not checked
- Job flow name: optional
- If there is only one job, we recommend "diverting the first job name".
- Click "Run Now" at the top right to execute the job.
- ※ It will take about 5 minutes for the job to finish.
- Confirm that the job is "Executed".
Create action table queries
- Select the created action table and click "Create" in the upper right of the screen
- Perform the following settings
- Query name: optional
- Extraction condition:
- Limit: 20
- Filter: Please specify the condition you want to filter.
- Example: When filtering by brand only
brand_cd = # {brand_cd}
- Example: If you want to filter by brand, large category, small category, gender
brand_cd = # {brand_cd}
l_category_cd = # {l_category_cd}
s_category_cd = # {s_category_cd}
sex_cd = # {sex_cd}
- ※ If you want to filter by multiple conditions, please set the mode of the action table to "flexible".
- Example: When filtering by brand only
- Sort:
quantity_rank
oramount_rank
ascending order
This completes the creation of the action table query.
- Please check the following document for detailed specifications of action table queries.
Set up Scenario
Scenario creation screen and select a template for recommendation.
- Template (3 types)
* If the template is not displayed, contact your sales representative.- KARTE recommendation (ranking) rank-mr
- After displaying a certain number of products, use the “more” button to display more products at the bottom.
- KARTE recommendation (ranking) rank-slk
- After displaying a certain number of products, display more products on the slide (carousel)
- KARTE recommendation (ranking) rank-mdl
- Display products in modal.
- KARTE recommendation (ranking) rank-mr
- Template (3 types)
Set the following from "Basic" on the Scenario detail screen.
Action table:
- Action table name: Specify the action table created above
- Action table query name: Specify the action table query name created above
filtering:
Whether to directly enter a value for each item (brand, child category, parent category, gender) that you want to narrow down in the action table frameUser information variableEnter the variable set in step
Example: When narrowing down by brand value: brand12345
Example: When filtering by brand_cd of the _view_item event
User information variable
filter
When you want to change filtering items
- Above ④Create action table queriesChange the extraction condition> Filter set in.
- Example: Brand
brand_cd
TEST field referencetest_cd
If you want to changebrand_cd = # {test_cd}
- Example: Brand
- Above ④Create action table queriesChange the extraction condition> Filter set in.
Data management:
User information variable
Purchased items (variable name: boughtItems): Setting not to display purchased items in the recommendation
- All period> Purchase> Product ID> Last 30
- Default value:"[]"
- Output as JSON: ON
Cart products (variable name: cartItems): Settings that do not display the products in the cart in the recommendations
- Latest Session> Cart> item_ids> Last 30
- Default value:"[]"
- Output as JSON: ON
Filtering items are designed assuming that they are specified in the character string type. If you want to filter by a data type other than character string, directly rewrite the logic in the template.
The main settings are now complete.
Depending on the type of template, the number to be displayed (maximum display number, minimum display number, etc.) can be set. In addition, it is possible to change the design freely by customizing in the same way as other Scenario.
- Perform the test delivery and confirm that the data put in the action table is displayed.
- If it is not displayed, please edit from the Scenario edit screen -> customize.
- For details, please check the following documents
Please make delivery settings with the same flow as the conventional Scenario .