This article provides troubleshooting tips for resolving issues with segments in Pendo when they don't work as expected.
Product area: People > Segments (web, mobile)
Missing or incomplete data
You might find that you have missing or incomplete data when your segment criteria are too specific or your segment rules are not properly applied. Review your segment logic and rules to ensure you're capturing all the necessary data points.
Analytics don't match what I expected
Inconsistent or unexpected results typically stem from how segment rules are defined. Below are some common issues.
Account versus visitor data
A common cause of unexpected results is confusion between account and visitor data. Visitor and account data come from different sources and must be used intentionally in segment rules. Accounts represent groups of associated users, typically the company they belong to. Visitors represent individual users of your application and typically belong to an account.
Accounts and visitors can have parallel fields when you add segment rules, like First Visit and Last Visit. When creating a segment, check that you're selecting a field under the desired data type (Account Data or Visitor Data).
Using multiple "does not contain" rules with "OR"
Using multiple does not contain
rules combined with OR
logic can lead to unexpected results. Each rule includes visitors if they meet either condition, which can result in visitors who should have been excluded.
For example, if you want to exclude all Visitor IDs that contain either “Apples” or “Oranges”, the following rules won’t work as expected:
Visitor ID does not contain Apples
This statement is saying to display the guide to everyone other than Apples, including Oranges.
OR
Visitor ID does not contain Oranges
This statement is saying to display the guide to everyone other than Oranges, including Apples.
This logic doesn't work because:
- A Visitor ID that contains "Apples" will match the second rule (does not contain Oranges), so it's included.
- A Visitor ID that contains "Oranges" will match the first rule (does not contain Apples), so it's also included.
As a result, both "Apples" and "Oranges" are included in the segment.
To correctly exclude Visitor IDs containing both Apples and Oranges, use AND
instead:
Visitor ID does not contain Apples
AND
Visitor ID does not contain Oranges
This ensures that only visitors whose IDs contain neither value are included in the segment.
Visit the Pendo Academy for a video on Segment Logic (And, Or, Not).
Segmenting by a boolean metadata field
Visitors who don't have a value for the metadata field are included in the "false" group for segments that evaluate metadata with the boolean datatype. If you see inconsistent or unexpected results, ensure that the boolean metadata field is consistently applied to all users and events that should be included in the segment.
Number of Eligible Visitors is different from what I expected
You can target a large number of known Visitor or Account IDs by uploading a .csv file containing the Visitor or Account IDs that you want to include or exclude from the segment. For more information, see the Segments article.
After you upload the file, Pendo shows the list of Eligible Visitors.
If the number of Eligible Visitors is different than expected, check that the format of the Visitor or Account IDs in the CSV file match the IDs in Pendo. Any formatting discrepancy (spacing, punctuation, or capitalization) can make a difference. If parent accounts are enabled for your subscription, check that the Account ID format in the CSV file matches the Account IDs in Pendo. For more information, see Managing multiple level accounts.
Additionally, only Visitor IDs with activity are included in the Eligible Visitors count. If you're pulling the IDs from another system, like Salesforce, Pendo might not have a record of them yet. This is especially likely with a new installation. When those users are identified by Pendo, they are added (or excluded) from the segment you've defined.
Filters are showing as "Unknown Segment"
If your filters are showing as "Unknown Segment", someone has deleted the segment currently set in Filters. This is indicated with the Unknown Segment label. Change the segment to update the page to show data as expected.
CSV failed to upload
The two most common causes for a CSV failing to upload are as follows:
- Double underscores. The presence of a double underscore at the beginning of a value, for example, "__username", generates an error upon upload. Ensure none of the values in your CSV file use this naming convention.
- Incorrect file type. CSV is the only supported file format. Ensure your file is saved as .csv.
My CSV segment is showing 0 eligible visitors when I apply it to a guide
Typically, this is due to the processing time associated with a CSV upload. The segment can be saved and used while Pendo processes the CSV and generates the eligible visitor total.
Sometimes this is due to Visitor IDs in the CSV not being known to Pendo. A visitor must use your app and generate a Pendo event for the Visitor ID to show up in your Pendo data. Pendo can't count unidentified visitors.
When those unidentified visitors sign in for the first time, Pendo receives the Visitor ID and the guide segment targets them as expected. The Eligible Visitor number on the guide updates to include the newly identified visitor.
For more information about segment issues, please see the following articles, Segment using Account ID returns visitors from other accounts, Eligible visitor count changes when applied to a guide, and video tutorial on Segments and Segments for Guide targeting.