Why looking at either side alone isn’t enough #
Most WordPress debugging starts from 1 signal. You look at the PHP error log, or you describe what happened in the browser, or you tell your developer what the error message said. Any 1 of those signals tells you something. None of them alone tells you enough.
A PHP error that fires on the server might cause a JavaScript error in the browser, which triggers a network request failure, which produces the broken UI you’re actually looking at. If you only see the PHP error, you don’t know how it expressed itself. If you only see the JavaScript error, you don’t know what caused it. If you only see the broken UI, you don’t know where in the chain the failure started.
Cross-signal correlation is how Loupely reads all 3 layers together and identifies the chain.
What correlation actually does #
After the browser capture and server capture complete, Loupely runs a set of correlation rules against the combined signal set. These rules look for known patterns that connect browser-side evidence to server-side causes.
Some examples of what that looks like in practice:
- A JavaScript AJAX error in the browser + a PHP Fatal Error in the same request on the server = the PHP error caused the AJAX call to fail. The browser-side error is a symptom. The PHP error is the cause. Without correlating the 2, you might spend hours debugging the wrong layer.
- A WooCommerce order confirmation in the browser + no corresponding database write on the server = a silent save failure. The browser thinks the order went through. The server never recorded it. That pattern is invisible unless you read both sides together.
- A network request returning a 200 status code + malformed JSON in the response body + a JavaScript parse error in the browser = a plugin conflict corrupting the REST API response before it reaches the browser. The 200 status code looks like success. It isn’t.
Why this is hard to do manually #
Correlating browser and server signals manually requires reading PHP error logs (which requires enabling debug mode and knowing where the log file lives), reading browser console output (which requires knowing how to use DevTools), cross-referencing timestamps across both sources, and knowing which patterns connect which signals. Most non-technical site owners can’t do this. Most technical site owners find it tedious even when they can.
Loupely automates the correlation step. The rules run in milliseconds. The result is a diagnosis that names the cause, not just a list of symptoms.
What the correlation layer produces #
The output of correlation is a failure classification: a specific category of problem (WooCommerce Save Pipeline failure, JavaScript conflict, PHP fatal, authentication failure, and so on) along with a confidence score and the specific signals that triggered the classification. That classification is what the AI diagnosis layer uses to generate the plain-language explanation and triage step you see in the popup.
When correlation finds a clear pattern, the diagnosis is specific. When the signals are ambiguous or the failure didn’t produce enough evidence during the capture window, the result may be a Null Result or a lower-confidence classification. See The Null Result: What It Means for those cases.