SEO testing in WordPress: measure each change, one page at a time
SEO testing means you change a page on purpose and then measure whether its search clicks moved more than they would have anyway. Without that last part, every rise looks like a win and every drop looks like bad luck.
This guide shows how a fair before and after test works, which traps to avoid, and how Pagereckon Pro runs the comparison for you on a control day.
What an SEO testing method needs
A simple before and after look at one page fools people all the time. Search traffic moves for many reasons: seasons, news, Google updates, a competitor's new post. Google's own guide to debugging traffic drops names seasonality and algorithm updates as common causes. So a fair test needs a few rules.
- Change one page, and one thing on it, so you know what caused the result.
- Write down the date of the change.
- Compare equal windows, the same number of days before and after, so weekdays and weekends balance out.
- Leave the change day out, since it is half old and half new.
- Control for the whole site, so a season or an update that moved every page is not credited to your edit.
- Wait until Google has seen the new version, or you are measuring the old page.
Time-based tests vs split tests
There are two main kinds of SEO testing. A time-based test changes one page and compares it with its own past. A split test changes a group of similar pages and compares them with a group you left alone over the same days.
Split tests can be stronger, but they need many similar pages, such as product or location pages. Most blogs and content sites do not have that. For them, a careful time-based test with a site-wide control is the practical choice.
Also, never show Googlebot a different page than readers see. Google's notes on website testing call that cloaking, even during a test. A before and after test avoids the issue, because everyone sees the same page.
How Pagereckon Pro runs an SEO testing report
When you change a page, you mark it as fine-tuned. You can do that from the dashboard, WP-CLI, the REST API, or an AI agent through the MCP server. The mark sets a control day, 14 days later by default.
On the control day, Pro builds a before and after report. It takes the same number of days on each side of the change, up to 28, and leaves the change day out. The window ends where Search Console data ends, about two days ago. Next, it compares the page's clicks with the whole site's Google clicks over the same days.
| Verdict | When the report gives it |
|---|---|
| Better | Page clicks rose 15%+ more than the site's, and by at least 3 clicks |
| Worse | Page clicks fell 15%+ more than the site's, and by at least 3 clicks |
| No clear change | Neither rule is met |
| Too little data | Under 5 clicks and under 50 impressions in both windows |
| Too early | Fewer than 7 days of Google data since the change |
The report also lists the top search terms before and after, so you see which ones moved. Without Search Console, it falls back to visits from Google in your server log.
The re-crawl warning in SEO testing
Google can only react to a change after Googlebot fetches the new version. Until then, searchers still see the old page, and any movement has nothing to do with your edit.
Pagereckon reads Googlebot's requests from your access log, so it knows when Google first opened the page after the change. If that has not happened yet, the report says so and warns you that Google still shows the old version. In that case, treat the verdict as early.
Why one change at a time
If you rewrite the title, add three sections and change the internal links on the same day, a good result tells you nothing about which part worked. Then you cannot repeat it on the next page.
So make one clear change, mark it, and move on to the next candidate. Our Search Console CTR guide shows how to pick pages with the most missed clicks, and the fine-tuning skill turns the whole loop into steps an AI agent can follow.
Finally, a fair note on other tools. SaaS products such as SEOTesting also run time-based and split tests on Search Console data, on a monthly plan per site. If you work across many sites that are not on WordPress, that may fit you better. Read Pagereckon vs SEOTesting for an honest comparison.
SEO testing questions
How long should an SEO testing window be?
Long enough for Google to re-crawl the page and for clicks to settle. Pagereckon Pro sets the control day 14 days after the change by default and compares up to 28 days on each side.
Can I run SEO testing on a small site?
Yes, but low-traffic pages often end as too little data. Start with pages that already get some clicks and impressions.
Why compare the page with the whole site?
Because a Google update or a season can lift or drop every page at once. Comparing with the site's clicks keeps that from looking like the effect of your change.
What is the best SEO testing tool for WordPress?
It depends on your setup. Pagereckon Pro runs inside WordPress and keeps your Search Console data on your server. SaaS tools like SEOTesting work across any platform.