Customer Insight

Query Customer Insight Resource

You can download the prediction insight results from the API, in order to enrich your CRM customer data. Prediction insight is per user base - you will get a list of insights broken down by customer customuid.

To understand customer insight and the API:

• API Information
• How to Export Prediction Insight
• About Customer Insight

API Information

Name	Value
HTTP Method	GET
URL	https://${AIXON-API-HOST}/api_download/v1/output/${resourceId}

Currently we support two kinds of output format: CSV (default) and JSON.

Example:

1. CSV (default output)

Request

curl -k -v "https://aixon.appier.com/api_download/v1/output/${resourceId}" \
     -H "x-api-key: ${api-key}" \
     -o output.csv

2. JSON

Request

curl -k -v "https://aixon.appier.com/api_download/v1/output/${resourceId}" \
     -H "x-api-key: ${api-key}" \
     -H "file-format: json" \
     -o output.json

You will get the API response then process and import into your CRM customer records.

---

About Customer Insight

Some notes regarding the prediction insights:

• Predictions will be updated on a weekly basis

• Prediction and insight results will depend on which Brain you use to deploy this Robot

• For the JSON output under Analyze&Segment Users Brain, we also provide the keywords result under each interest.

---

Predict Conversions Output Structures

• customuid => Derived from the custom_uid in the CRM data you used for this Robot

• emailsha256

• idfa

• dmpid

• conversion score => On the granular level, AIXON will rank the users based on their probability to convert. AIXON will then give a score of 100 to user(s) in the group with highest probability. Please note that a score of 100 does NOT mean that the user has a probability of 100% to convert.

• prediction valid start time

• prediction valid end time

• use PC in last 7 days => Binary (0 or 1)

• use mobile in last 7 days => Binary (0 or 1)

CSV:

JSON:

---

Analyze&Segment Users Output Structures

• customuid => Derived from the customuid in the CRM data you used for this Robot

  • We will only export custom_uids that have general topics

• emailsha256

• idfa

• dmpid

• Out of network keywords => Keywords derived from websites (other than yours) that users visited

• In network keywords => Keywords derived from your websites

• Preset Interest => Groups of keywords that were pre-set by Appier

• My Interest => Groups of keywords customized by you

• Custom keywords => Keywords customized by you

• Keyword timeframe => how far into the past we extracted the user's topic interests

CSV:

JSON:

Note: After downloading the file, you will see the information displayed by HTTP with most of which is the standard HTTP headers.

Pleaser refer to this wiki for detailed definition of each field. Additionally, there are additional headers:

• total-export-count: number of rows in the exported file
• robot-finish-date: date of when robot finishes in YYYY-MM-DD
• robot-finish-timestamp: timestamp for when robot finishes in YYYY-MM-DD HH:MM
• used-db-date: latest date of Persona (Cassandra) version used by Robot This used-db-date result is queried when the robot robot-finish-date finishes processing and thus both of these dates are closely connected. For example, today is 0828 but we saw used-db-date as 0827, this does not mean that the persona db died on 0828 but is just that the latest persona DB query date was 0827. Since Robots does not have a fixed end processing time and could finish processing before persona DB finished deployment so we can’t take used-db-date to understand persona db status.
• aixon-db-deployment: timestamp for when Aixon persona DB deployment completes per region (YYYY-MM-DD HH:MM for specific region)
• file-status: This is to check if csv/json file output node is processing or finished today (“in-progress” or “finished”)

There are also more header options to retrieve Insights with:

• Files will be transferred by gz file
• The downloaded file by default will be kept as-is (input -H "file-format: json" in header, and downloaded file will be .json)
• We built additional 4 options in the header for your preferences;

1. -H "file-format: json" -H "Accept-Encoding: gzip": file format download in .json.gz
2. -H "file-format: csv" -H "Accept-Encoding: gzip": file format download in .csv.gz
3. -H "file-format: json" -H "Accept-Encoding: identity": file format download in .json
4. -H "file-format: csv" -H "Accept-Encoding: identity": file format download in .csv

• There is a checksum response as well that you can verify for data integrity