Open the data for the same page and query in Yandex Webmaster. An improvement in the position of one phrase does not have to be reflected in the total number of clicks: other articles or services may have received fewer impressions during the period.
Note the start and end dates of the statistics and the date the position was checked. Then find the segment that accounts for the difference. To discuss visits, you need clicks over a period; a saved position describes the rank in a single check.
Distinguish rank, impressions, and clicks
A click is a user's action on a search result. A tracked position is the rank found for a selected query in a specific set of search results. A single search snapshot does not tell you how many people then saw the result and clicked it.
Yandex Webmaster also shows impressions and average metrics. Yandex Help on analytics describes what they are for. When reading the data, use the exact metric name rather than the general word “traffic” for every number.
An improvement in one position does not have to coincide with an increase in the site's total clicks. To reach that conclusion, you need to account for other queries, pages, demand, and the period, as well as the availability of the data itself.
Check the query list
You choose the query list for your subscription yourself. It does not necessarily include every phrase that brings visits to the site. The source statistics may include other topics and pages.
Check whether the selected phrase appears in the Yandex Webmaster data available to you. If there is no exact match, you cannot treat the overall graph as a direct measurement of that phrase. Keep the difference between the datasets in mind.
The keyword research guide can help you assess which needs your working list covers. An important service may be monitored narrowly, while a significant share of visits may come from informational content or branded queries.
Check the date range
The position may have been measured on one day, while clicks were collected over several weeks. A comparison needs a clearly defined range. Do not explain a monthly total using a single new data point if most of the period passed before it.
Write down the start and end dates of the statistics, the date of the search snapshot, and the import date. An import is not the same as the time of every event within the period. A new view of the dashboard may already include previously reported days.
To compare statistics, choose periods of equal length and suitable consecutive ranges. Overlapping ranges cannot be described as independent weeks: they include some of the same events.
Separate the site's different areas
Imagine an educational site with two areas. A service query improved in rank, but an informational article received fewer impressions. As a result, the site's total clicks declined.
| Example observation | What it describes | Next question to ask |
|---|---|---|
| The service phrase ranked higher | One selected query | How is this page performing? |
| The article received fewer impressions | A different segment of the source | Did demand or reach change? |
| Total clicks are lower | The combined events for the period | Which part accounts for the difference? |
| Some data is missing | A limitation of the source | What cannot yet be compared? |
Look at impressions alongside clicks
If impressions declined, fewer clicks may reflect a smaller available volume of impressions. If impressions are similar but there are fewer visits, that raises another question to investigate. But this pattern alone does not establish the exact cause either.
Check whether the numbers relate to the same page and period. Do not mix the site's overall total with the CTR of an individual phrase. If the data is partial, first note that limitation, then make arithmetic comparisons.
Do not promise that changing the title (the page name shown in the browser tab) will necessarily bring visits back. The snippet, demand, and composition of the search results need to be investigated, and an editorial change should be accepted based on clarity and how well it matches the page itself.
Check the click-through rate with an example
In an example period, there were one hundred impressions and ten clicks: the click-through rate is ten percent. In the second period, there were two hundred impressions and fifteen clicks: there were more clicks, while the rate was seven and a half percent. The two metrics changed in different directions without any arithmetic contradiction.
This example does not explain the causes of a real event. It only shows why clearly labeled denominators matter when drawing conclusions. Do not apply percentages to a query's position or call the difference in rates the number of lost customers. Customer actions after a visit require other evidence.
Read more: Website rank tracking: methods and tools.
Check results are available in the site monitoring dashboard.