When you run an Autosegment, you might see the generic error:
Unable to create auto segment. Please contact support with the following information…
This message means the Autosegment job didn't complete successfully. The message itself doesn't show the exact root cause.
In many cases, the error is caused by one of these two setup issues:
The Autosegment uses a customer property, metric, aggregate, or segmentation that references events or properties that don't exist in the project.
There is insufficient or no usable data in the last 6 months for the events and filters used in the Autosegment.
This article explains how Autosegments work and the most common reasons they fail.
How Autosegments use data
Autosegments are powered by Loomi AI. They generate potential customer segments by combining customer attributes and metrics and identifying patterns in your data.
Key points about the data they use:
Autosegments use at most the last 6 months of event history.
Segments are only returned if they meet the minimum user count you define.
Metrics and segments are calculated from long-term data storage, not from real-time data.
Because of this, Autosegments need both of the following:
Correctly defined events, properties, metrics, and aggregates in your project.
Enough recent eligible data for the selected setup.
If either of these is missing or inconsistent, the Autosegment can fail.
Cause 1: References to non‑existent events or properties
The Autosegment setup includes at least one element that points to something your project doesn't actually have. Examples:
A metric based on an event (for example
order) that isn't tracked in this project.A metric or aggregate that uses an attribute (for example
category) that was renamed or removed.A customer property or segmentation cloned from another project that still references events or fields that exist only in the original project.
Typos or case mismatches in event names (for example,
Emailvsemail).
When Autosegments try to compute results, they can't find those references to evaluate, and the job fails with the generic error.
Cause 2: No or insufficient data in the last 6 months
Autosegments only consider up to 6 months of event history. Older historical data isn't used for this calculation.
A failure can happen when:
The events used in metrics or filters are rare or haven't occurred recently.
The project is new or doesn't yet have enough recent tracked data for the chosen setup.
The customer filter leaves no eligible profiles.
The minimum user count is too strict for the available recent data.
In these cases, there may be so little usable data that the job can't build stable segments, and it fails.
Other limitations
Autosegments also have additional documented limits, including limits on the number of properties, metrics, segments, considered events, considered customers, and supported property types.
For the current full list of limitations, read the Autosegments documentation: Autosegments > Limitations.