Quality page
The Quality page is used to check the quality of data in a CDP. For example, if you were about to trigger an email campaign to thousands of customers, you could make sure your data quality is acceptable by first checking this page.
In general, you can expect to some peaks and valleys on the Quality of Data graph (depending on overall ingestion timing, ingestion loads, time of year, and so on.
Hygiene section
Quality of Data section
This section displays a snapshot of data quality so that you can answer questions such as:
How are we doing?
Do we have good data, especially as it relates to contact data on our customers? (Address, phone, and email are the typical contact data fields.) See Data quality scores for a discussion of how data quality is calculated and what good or bad means in terms of your data.
Click a data field button to toggle between graphing and hiding field data.
Hover over a graphed field line to display a popup of field quality score on a specific date.
Match Data Quality section
This section presents the same data as the Match Data Quality section on the Data Observability dashboard.
The information displayed in this section is another quick way of answering the question "how good is my data?"
Match data quality is an indicator of the quality of PII data used for matching. See Data quality scores for an overview of how data quality scores are calculated.
In the preceding screenshot, the calculated match data quality is Medium-High (pretty good!) based on matching data from 10,000 individuals and 2,500 households.
Matching section
Processed matches section
This section gives you insights into your matches, every day. It answers questions such as:
How many matches have been processed?
What is the average household size?
Hover over a colored run box to display the run stats.
Individual Match Confidence section
This graph gives you a more nuanced view than the Match Data Quality section. Instead of a flat rating, this graph lets you dig deeper into the numbers and ask questions such as:
How many matches were of medium quality?
How many matches were of low quality?
If the Match Data Quality is currently
Medium-High, is the data quality skewing more towards the high or medium side?
Match confidence relates to the threshold defined for matching. If a match score is higher than the set matching threshold (for example, 90%) it is considered a match. At this point match quality is determined by ranges above the matching threshold and inside the matching range.
For example, one could imagine match confidence ranges (aka buckets) set as such:
High Confidence = match score of 99%+
Medium Confidence = 97% to 99%
Medium Confidence = 95% to 97%
Low Confidence = 90% to 95%
Address, Phone, Email sections
These sections delve into the individual PII (Personal Identifiable Information) component details.
Click View Detail to display an address hygiene report.
Correction of a PII component means updating the data as a part of the data cleansing and normalization process on ingestion. See format validity: address, format validity: phone, and format validity: email for discussions concerning the format of each of these PII components.
For example, an address correction could be the addition of an address component like Delivery Point Code.
As shown in the screenshot, Redpoint CDP has processed:
141.5k addresses
7.75% of those addresses were corrected.
96.18% of those addresses were verified.
467.9k phones
0% of those phone numbers were corrected.
99.99% of those phone numbers were determined to have a valid format.
200.8k emails
0.29% of those email addresses were corrected.
100% of those email addresses were determined to have a valid format.
Match Runs section
The Match Runs section displays a table of feed runs in which matching is a part of the feed computational process. Each row corresponds to a single match run for the associated feed run.
Click the Feed link to display the Feed Definition page.
Click the Instance link to display the Feed Instance page.
Click Filter by field to filter feeds via text string.
The table columns correspond to feed matching fields populated by the matching run.
