Compare website performance before and after a hosting change against a clear baseline, not a single speed-test score. Performance includes how quickly a site loads, displays content, and responds to user actions. The comparison should include the real-user experience, server response, availability, and resource use. Learn more about website performance testing and analysis.
Record the same seven metrics before and after the change. Use the same pages, test conditions, and measurement periods so that the results can be compared with less ambiguity.
Table of Contents
1. Real-user loading experience
Real-user loading experience measures how quickly visitors experience a usable page in normal browsing conditions. It reflects the result that people see, rather than only the time taken by a server in an isolated test.
Capture this metric for the same important pages before and after the hosting change. Keep the page set consistent. A change in the page itself can affect the result, so record any content or configuration changes made during the comparison.
2. Core Web Vitals
Core Web Vitals are a group of user-experience measures used to assess loading, visual stability, and interaction responsiveness. Track the same Core Web Vitals data before and after the move, using the same page templates where possible.
Core Web Vitals may remain unchanged even when the new hosting environment returns data faster. This is not a contradiction. Server response is only one part of the time needed to display and use a page. Other parts of the page delivery process can still affect the visitor experience.
3. Server response time
Server response time measures how long the hosting environment takes to begin responding to a request. It is useful for checking the direct effect of the hosting change.
Compare the same URL before and after migration. If server response improves, the new environment is handling the initial request more quickly. If real-user results do not improve at the same time, review the other metrics instead of treating the server result as the final performance score.
4. Uptime
Uptime measures whether the website is available during the selected measurement period. Record both the percentage of available time and any interruptions that occur.
Availability is separate from speed. A site can respond quickly when it is online but still perform poorly as a service if it has interruptions. Include the same observation period before and after the change when comparing uptime.
5. CPU usage
CPU usage shows how much processing capacity the website uses in the hosting environment. Record normal usage and periods of higher activity when those periods are part of the baseline.
Compare usage with the same website activity on both environments. Lower CPU usage after the move can indicate more available processing capacity, but it should be read together with server response time and real-user results.
6. Memory usage
Memory usage shows how much working memory the website consumes. Track the usual level and any periods when usage rises.
A change in memory usage does not by itself prove that visitors have a faster website. Use it to identify whether the hosting environment has enough capacity for the current workload, then compare it with response and user-experience metrics.
7. Bandwidth usage
Bandwidth usage measures the amount of data transferred by the website during the selected period. Record the period, the pages or services included, and the level of traffic represented by the measurement.
Compare like with like. A different number of visitors, larger page files, or new media can change bandwidth usage independently of the hosting environment. Keep these conditions consistent when building the before-and-after comparison.
How to create a comparable baseline
- Select the pages. Choose the same key pages for both measurement periods. Include the pages that matter most to the website’s visitors.
- Record the seven metrics. Save the results for real-user loading experience, Core Web Vitals, server response time, uptime, CPU usage, memory usage, and bandwidth usage.
- Note the conditions. Record the measurement period, website version, traffic level, and any important configuration changes.
- Keep the old environment active during preparation. Migration transfers the website’s files to the new hosting provider, and the existing provider account must remain active while preparing for the process. See the mybox migration performance guidance.
- Repeat the same measurements after the change. Use the same pages and conditions instead of comparing a single before result with an unrelated after result.
- Compare patterns across the metrics. Look for changes that appear in more than one area. Do not use one improved number as the only measure of success.
How to interpret conflicting results
| Result | What it indicates |
|---|---|
| Faster server response, unchanged Core Web Vitals | The server is responding faster, but another part of loading, display, or interaction still has a similar effect on users. |
| Better real-user experience, similar server response | The hosting change may not have changed the initial response substantially, while the overall visitor result has changed. Review the full set of measurements and any page or configuration changes. |
| Lower CPU or memory usage, no visible speed change | The environment may have more available capacity, but capacity and user-facing speed are different metrics. |
| Good speed, reduced uptime | The site may be fast when available, but availability requires separate attention. |
What a useful comparison should show
A reliable comparison shows how the hosting environment affects the server, the website’s resources, and the visitor experience. The clearest result is not always an improvement in every metric. A server response improvement with stable Core Web Vitals is still a specific finding, while unchanged uptime or resource use provides separate information about the move.
Keep the baseline and post-change results together. This creates a repeatable record for future hosting or website changes and makes performance decisions based on several measurements rather than a vague impression.