How to measure whether an SEO change worked

To measure whether an SEO change worked, record the numbers before you change anything, note the date, wait until Google has recrawled the page and a full comparison period has passed, then compare clicks, impressions, CTR and position for the same searches over equal periods. Search Console's Compare view and annotations do most of the work. You'll rarely get proof, but you can get a fair, honest read.

The rest of this guide shows the exact steps, and how to make the result more trustworthy than "the line went up".

Why it matters

Most small sites change things and never find out whether it helped. The changes pile up, some of them hurt, and nobody can say which. A repeatable habit of "baseline, change, wait, compare" fixes that and stops you repeating work that did nothing.

It also protects you from fooling yourself. Search traffic moves for reasons unrelated to what you did. Google's help on measuring the impact of page changes is blunt about it: it is "difficult to determine if a specific change you made to a page directly caused an improvement", because "other events, such as periodic changes in user sentiment, news events, or improvements to a competing site, can also contribute". Its advice is to check whether the improvement "maps closely in timing to your site changes, compared to similar time periods in the past".

How to measure an SEO change, step by step

Labels below match Google's Performance report, advanced filtering and comparison and annotations help pages.

Step 1: Change one thing at a time

If you rewrite the title, the headings and the content in one go, you won't know which helped. Pick one change per page. It's slower, but you learn something every time.

Step 2: Record the baseline before you edit

Open Performance, then Search results. Filter to the page you're about to change: click + Add filter, choose Pages, then URLs containing, and enter the page's path. Set the date range to the last 28 days, with Clicks, Impressions, Average CTR and Average position switched on.

Write down the four numbers. Then click the Queries tab and note the top five searches with their clicks, CTR and position. If you only record page totals, a change to one search can be hidden by movement in others.

Step 3: Annotate the date

Right-click the chart at the date of the change, type a short note (up to 120 characters), and click Add. Annotations are shared with every owner and full user of the property, up to 200 per property, and Google deletes ones older than 500 days. Don't put personal information in them.

Google notes that annotations don't show in comparison mode or the 24-hour view, so look for them on the normal chart. This takes ten seconds and saves you from guessing, three months later, when a title changed.

Step 4: Wait the right amount of time

Two delays apply. First, Search Console data arrives late: Google says collected data is usually available in two to three days, and the newest data can be marked preliminary with a dotted line. Second, Google has to recrawl the page and then rank it. Google's drop guide says some changes take effect in a few days and others take months, and that you'll "likely want to wait a few weeks" before analysing Search Console again.

A practical rule: wait until a full 28 days of data exists that starts after the change was crawled. If you want to nudge a recrawl, the URL Inspection tool lets owners and full users request indexing, but Google notes that requesting repeatedly won't make it faster.

Step 5: Compare equal periods

Open the date filter, choose the Compare tab and set the two date ranges:

• The "after" period: the 28 days starting after the change.
• The "before" period: the 28 days up to the change.

Click Apply. With the same page filter on, the table shows each search with a Difference column. Google recommends weekly or monthly granularity when comparing two date ranges, because it removes the day-of-the-week effect. Choose Weekly in the granularity menu.

Only one comparison can run at a time, so a new one replaces the old.

Step 6: Read it in the right order

1. Impressions first. If impressions fell for the page's main searches, demand or ranking changed, and a title tweak isn't the explanation.
2. Then position. A rise in position suggests ranking improved. Remember that position is the topmost position, averaged.
3. Then CTR and clicks. If impressions and position held and CTR rose, the snippet is doing better. That is the cleanest signal for a title or description change. To judge whether a CTR is high or low for a position, see what a good CTR looks like by position.

Step 7: Check the same period last year

Switch the comparison to year over year and look at the same window. If clicks rose by the same shape last year at the same time, your "win" may be a season.

Step 8: Use pages you didn't change as a control

This is the strongest thing you can do with free data. Pick two or three similar pages you left alone, filter to them with a Custom (regex) page filter, and run the same comparison. If your changed page rose and the control pages were flat, the case for your change gets stronger. If everything rose, it was probably the market. Our guide to regex filters has patterns for choosing groups of pages.

A worked example

This is an illustrative example, not real data.

A small software firm rewrites the title of "/blog/invoice-template" from "Invoice template" to "Free invoice template for freelancers (Word and Excel)" on 3 March. The baseline for the 28 days before:

They annotate the chart. Four weeks of data after the recrawl (starting 10 March) show:

Impressions and position barely moved, but CTR and clicks rose. That fits a snippet improvement. A year-over-year check shows the same window last year was flat, and two similar pages they didn't touch stayed within a few clicks of before. The evidence is decent, not conclusive. They keep the new title and note it as a likely win.

Common mistakes

• Comparing after one week. Not enough data, and Google may not have recrawled yet.
• Comparing a short window to a long one. Use equal lengths.
• Declaring victory from the site total. Total clicks move for many reasons. Measure the page and its searches.
• Skipping the baseline. Without it, you're comparing to a memory.
• Making three changes at once. You can't separate the effects.
• Ignoring seasonality. Always glance at last year.

Where this fits in your routine

Measurement is the last step of a loop: find a problem, make a change, wait, check. Our Google Search Console guide for founders explains the reports used here. Our 30-minute weekly SEO routine shows where each step fits in a week, and how to prioritise SEO quick wins explains which changes are worth measuring first. If you're testing a title on a page with weak clicks, start with impressions but no clicks.

Doing this every week? Sense Console does it automatically

When you click Mark done on an opportunity in Sense Console, it records that search's clicks, impressions, CTR and position for the latest 28-day window, and you can say which page you updated. After about 31 days, once your next data refresh has the new period, the card shows Change since you marked this done with the change in clicks and position for that search. The wording describes a change since you marked it done, not proof the action caused it, and a result can take longer than 31 days because the Free plan refreshes every 14 days. It measures searches that Sense Console flagged, so for a page-level test or a control group, use the Search Console steps above.

Start free. Free plan. No card required.