Separate waiting from searching.
A page can load quickly while the task still takes too many clicks. A simple task can also be held up by a slow request. Measure those two problems separately: how long the screen takes to respond and how much work you do to reach the answer.
If only the order list is slow, keep the investigation there. If everything is slow, compare the public shop and other admin screens. A public page-speed score is useful for a visitor journey, but it does not diagnose every authenticated admin action.
Repeat one representative action.
Choose a normal order lookup or product save and describe the conditions: account, filter, dataset and time. Check whether the same delay happens consistently or only sometimes. Avoid repeated saves on a real order simply to generate timing samples.
Give the developer a precise task to profile. Query Monitor can show the current page’s database work, errors and server-side HTTP calls. A large number of queries alone is not a complete diagnosis; one slow external request may matter more.
A focused handover needs only the useful details.
This fictional example separates the symptom from the proposed explanation. Use measurements from your own system when repeating it.
| Field | Example |
|---|---|
| Action | Open an order from a customer search |
| Screen | Authenticated WooCommerce order view |
| Observed delay | Record the actual wait; do not estimate a promised saving |
| Recent change | New reporting integration added |
| Question for the developer | Is the delay local processing or an external request? |
Match the next step to the finding.
An external timeout belongs with that connection or provider. Database-heavy work may need a query or data-structure review. Hosting limits belong in the hosting discussion. Ask for a specific finding before buying more capacity or removing a whole set of tools.
For order storage, HPOS is a compatibility and migration question, not a universal switch for speed. Test the relevant setup and compare the same action after a change. Separately count the steps required to answer the customer; technical response time and workflow effort can both improve, but they are different results.