How do you find out why a page feels slow?
Start with one page and the moment that feels slow. Note the address, what you were trying to do, and whether the delay happens before the page appears or after you interact with it.
What should you measure first?
Choose something you can observe, such as how long the banner takes to appear or how long the list takes to update after a click. A page-load report can give you a lead about the banner. To investigate the list, record the actual click.
Where can you safely repeat the problem?
Use a standalone local sample or a preview you control. Before opening the preview, check where it saves data, which accounts can sign in and which services it calls. Confirm that those connections are disabled or point to separate test services with fictional records and test accounts. If you cannot confirm that, use the standalone sample or write down the steps without running the app.
For a hosted test, use a public address you have permission to send to the service. Keep private addresses, screenshots and recordings within the people authorized to see them.
Compare tools
Scroll the table sideways to read every column.
| Tool | Use it whenThe question you need to answer. | Where to testThe page access you need. | What you getThe evidence you can act on. |
|---|---|---|---|
| Chrome DevTools | You can reproduce a slow request or click. | Open the page in desktop Chrome, including a local sample. | A request timing or a browser timeline for that attempt. |
| Lighthouse | You want an automated page audit to find a first issue. | Use Chrome on a local, public or authorized signed-in page. | Findings from a controlled test you can repeat. |
| PageSpeed Insights | You want to check a public page against available visitor history. | Submit a public address to Google's report site. | A hosted Lighthouse run and, with enough eligible visits, visitor data from the past 28 days for the page or site address. |
What can a test leave unanswered about visitors?
A lab test uses selected settings to help you reproduce an issue. Field data summarizes what real visitors experienced across different devices and connections. If PageSpeed Insights only has data for the origin, it can include other pages on that site address. If it has no visitor data, the test can still help you investigate the page.
Core Web Vitals cover three separate parts of the experience. Largest Contentful Paint measures when the largest visible image or text block appears. Interaction to Next Paint measures how quickly the page responds visually to interactions. Cumulative Layout Shift measures how much the layout moves unexpectedly.
How do you check whether a change helped?
Keep a baseline for the same page and action, with the browser, device and test settings you used. Change one thing, repeat the test and compare the relevant finding. Check that the page still works and repeat if the results vary.
After launch, a scheduled test can help you spot a return of the same problem. Measurements from real visits can show who actually encounters it. Pick the important page or interaction before deciding which kind of monitoring you need.