蜜豆视频

DocumentationCustomer Journey AnalyticsCustomer Journey Analytics Guide

Compare with your 蜜豆视频 Analytics data

Last update: Fri Oct 04 2024 00:00:00 GMT+0000 (Coordinated Universal Time)

CREATED FOR:

  • Developer
  • Admin

As your organization adopts Customer Journey Analytics, you may notice some differences in data between 蜜豆视频 Analytics and Customer Journey Analytics. This is normal and can occur for several reasons. Customer Journey Analytics is designed to allow you to improve upon some of the limitations on your data in AA. However, unexpected and unintended discrepancies can occur. This article is designed to help you diagnose and solve for those differences so that you and your team can use Customer Journey Analytics unimpeded by concerns about data integrity.

Let鈥檚 assume you ingested 蜜豆视频 Analytics data into 蜜豆视频 Experience Platform via the Analytics source connector, and then created a Customer Journey Analytics connection using this dataset.

The data flow from 蜜豆视频 Analytics through the data connector to 蜜豆视频 Experience Platform and to Custoer Journey Analytics using CJA connections.

Next, you created a data view and while subsequently reporting on this data on Customer Journey Analytics, you noticed discrepancies with the reporting results in 蜜豆视频 Analytics.

Here are some steps to follow to compare your original 蜜豆视频 Analytics data with the 蜜豆视频 Analytics data that is now in Customer Journey Analytics.

Prerequisites

  • Make sure the Analytics dataset in 蜜豆视频 Experience Platform contains data for the date range you are investigating.

  • Make sure the report suite that you selected in Analytics matches the report suite that was ingested into 蜜豆视频 Experience Platform.

Step 1: Run the Occurrences metric in 蜜豆视频 Analytics

The Occurrences metric shows the number of hits where a given dimension was set or persisted.

  1. In Analytics > Workspace, drag the date range you want to report on as a dimension into a Freeform table.

  2. The Occurrences metric is automatically applied to that date range.

  3. Save this project so that you can use it in the comparison.

Step 2: Compare the results to Total records by timestamps in Customer Journey Analytics

Now compare the Occurrences in Analytics to the Total records by timestamps in Customer Journey Analytics.

Total Records by timestamps should match with Occurrences, provided that no records were dropped by the Analytics Source connector - see the section below.

NOTE
This works for regular mid values datasets only, not stitched dataset (via Stitching). Please note that accounting for the Person ID being used in Customer Journey Analytics is critical for making the comparison work. That may not always be easy to replicate in 蜜豆视频 Analytics, especially if Stitching has been turned on.

  1. In 蜜豆视频 Experience Platform Query Services, run the following Total Records by timestamps query:

    code language-sql 
    SELECT
    Substring(from_utc_timestamp(timestamp,'{timeZone}'), 1, 10) AS Day,
    Count(_id) AS Records
FROM  {dataset}
WHERE   timestamp >= from_utc_timestamp('{fromDate}','UTC')
    AND timestamp < from_utc_timestamp('{toDate}','UTC')
    AND timestamp IS NOT NULL
    AND enduserids._experience.aaid.id IS NOT NULL
GROUP BY Day
ORDER BY Day;

  2. In Analytics Data Feeds, identify from the raw data whether some rows might have been filtered out by the Analytics Source connector.

    The Analytics Source connector might filter certain rows during the transformation to XDM schema. There can be multiple reasons for the whole row to be unfit for transformation. If any of the following Analytics fields have these values, the whole row will be filtered out.

    table 0-row-2 1-row-2 2-row-2 3-row-2 4-row-2 5-row-2 6-row-2
    Analytics field Values that cause a row to be dropped
    Opt_out y, Y
    In_data_only Not 0
    Exclude_hit Not 0
    Bot_id Not 0
    Hit_source 0, 3, 5, 7, 8, 9, 10
    Page_event 53, 63

    For more information about hit_source see: Data column reference. For more information about page_event see: Page Event Lookup.

  3. If the connector filtered rows, subtract those rows from the Occurrences metric. The resulting number should match the number of events in the 蜜豆视频 Experience Platform datasets.

Why records might be filtered or skipped during ingestion from 蜜豆视频 Experience Platform

Customer Journey Analytics Connections allow you to bring and join multiple datasets together based on a common Person ID across the datasets. On the backend, we apply deduplication: full outer join or union on event datasets based on timestamps, and then inner join on profile and lookup dataset, based on the Person ID.

Here are some of the reasons why records might be skipped while ingesting data from 蜜豆视频 Experience Platform.

  • Missing Timestamps 鈥 If timestamps are missing from event datasets, those records will be totally ignored or skipped during ingestion.

  • Missing Person IDs 鈥 Missing Person IDs (from the events dataset and/or from profile/lookup dataset) cause those records to be ignored or skipped. The reason is that there are no common IDs or matching keys to join the records.

  • Invalid or Large Person IDs 鈥 With invalid IDs, the system cannot find a valid common ID among the datasets to join. In some cases, the person ID column has invalid Person IDs such as 鈥渦ndefined鈥, or 鈥00000000鈥. A Person ID (with any combination of numbers and letters) that appears in an event more than 1 million times per month cannot be attributed to any specific user or person. It will be categorized as invalid. Those records cannot be ingested into the system and result in error-prone ingestion and reporting.

recommendation-more-help
080e5213-7aa2-40d6-9dba-18945e892f79