I have heard about the Translation Memory usage statistics, but I am not sure how this new feature works. Can you tell me more?

Beginning with version 11.6, WorldServer now tracks the number of segment matches provided by each translation memory (TM) for each task during the segmentation process. It also monitors the frequency with which each individual segment has been used as a match. This information provides a clearer understanding of which TMs and segments are the most valuable. This works both for files segmented with Legacy filters and with File Types.

With this new feature, when a file or task is segmented, the number of TM matches found in each TM is displayed in the Task History/Task Information. Below is an example from the Segment Asset record in the Task History:


 

In the example above, the project type used a single TM (Default), which provided 25 segment matches for the task in question. However, if you use a TM sequence, the number of matches is calculated for each TM, as shown in the following example:

 

Once the feature is enabled, the Task History will show the record of the Segment Asset step, which includes the number of TM Entries (not segments) that have been applied to the segments in an asset. These TM entries can be applied to multiple segments, so the number of segment matches mentioned in the Task History does not necessarily correspond to the number of segments they have been applied to.

For instance, you can have a TM entry that has been applied as an ICE Match to one segment and as a Fuzzy Match to another. The count of Segment Matches in the Task History will be 1, even if it has been applied to multiple segments. The TM leverage covers ICE and 100% Matches as well as Fuzzy Matches.

The Usage Count is updated based on the asset. If an asset contains 2 or more segments leveraging from the same TM entry, the Usage Count value added to the TM Entry will still be 1 because it comes from the same asset/task. The Usage Count is updated each time an asset is segmented and the TM entry is utilized.

The Usage Count of the TM Entry is displayed on the Translation Memory page if you enable the Usage Count column. You can also add the Last Usage Date column to your view: 



You can also view this information in the System Attributes section when viewing or modifying a TM entry:

Note: in Oracle-based WorldServer environments only, the usage statistics do not work correctly for Fuzzy Matches. A defect has been filed for this.

Please refer to this article for more details:

WorldServer - Usage Statistics for Fuzzy Match Segments do not work in Oracle-based Environments

This article might be interesting for you as well:

WorldServer - what is Cache TM and how is it used?
WorldServer - Usage Statistics for Fuzzy Match Segments do not work in Oracle-based Environments