ISSUE

If you use the Acrolinx Sidebar together with automated or batch checking, you might occasionally notice inconsistencies in the Acrolinx Scores for the same content. You’ll notice this behavior in CCMS-like applications where Acrolinx integrates both in the frontend (Sidebar) and backend (automated or batch check).

For example, in the Sidebar, you might get an Acrolinx Score of 81, which appears green.

If Acrolinx checks the same article as an automated check on save, you might see a score of 79, which appears yellow.

ENVIRONMENT

CAUSE: How Acrolinx Reads and Scores Content

Depending on your application and how content is stored in the database, it might not be possible to get the scores to match exactly across different checking features. Although the document is the same content to the eyes of the viewer, the underlying HTML or XML structure could differ depending on the API.

Document Extraction in the Integration

An Acrolinx Integration calls an application API to retrieve the content and structure of a piece of content in its entirety. The document includes both the words that the reader sees and any structural content and meta information that the reader doesn’t see. For example, this might be HTML or XML.

Document Extraction in Sidebar

When you use the Acrolinx Sidebar to check your content, the Acrolinx Integration relies on your application’s frontend API to extract the document in the source format.

Document Extraction in Backend

When the Acrolinx automated or batch checks run, the Acrolinx Integration relies on your application’s backend API to extract the document in the source format.

Document Extraction (Platform-Side)

Acrolinx Integrations handle the document extraction in the host application. But the Acrolinx Platform also uses extraction settings to filter out the natural language of the content that Acrolinx ultimately reads. You might be able to compensate for some of the differences in HTML or XML by fiddling with the Content Profile extraction settings.

How Acrolinx Scores Are Calculated

The Acrolinx Score is normalized based on the number of words in your document. If the Sidebar check shows a word count of X and the automated check shows Y, then the word count alone could cause the minuscule differences in Acrolinx Scores.

Particularly for very short documents, the smallest difference in the underlying document structure can lead to a slightly different score. Typically, the difference in scores is only by a few points and the issues that Acrolinx finds are exactly the same.

RESOLUTION

Make sure that all checking features use:

• the same Target
• the same Content Profile

It's possible that Acrolinx is using a different Target for the different checkings features.

If you’re still getting different Acrolinx Scores.

Acrolinx Integrations rely on an application’s APIs for fetching, or extracting, the content. Automated checks often use different APIs than the interactive Sidebar uses. Batch or automated checks need to work in the background and for larger sets of content. Many applications provide different APIs for that capability.

Neither developers nor Acrolinx can influence or change the APIs of applications where Acrolinx runs. The one thing we can influence is the way Acrolinx reads the extracted document.

Additional Information

• How Acrolinx Scores are Calculated

• How Acrolinx Reads Content

• Content Profiles let you configure how Acrolinx reads your content.