When traffic drops, a common practice is to check Google Search Console to see if rankings have dropped, as well. If they have, the traffic loss is generally attributed to the rankings change.
A decline in clicks like this is often attributed to changes in rank position.
In some cases, however, the shift in average position in GSC does not actually indicate a rankings change. Instead, it represents a shift in demand.
Google collects rank position data when an impression occurs and uses that data to calculate an average position. Sometimes, the change in average position happens, not because rankings have changed, but because people’s search habits have changed.
Changes in Average Position Due to Search Demand
Here’s an example of the GSC Search Analytics data for a website section that has high demand on weekdays and lower demand over the weekend.
In this case, the average position drop does not actually represent a change in rankings.
It looks like rankings drop each weekend and rebound on Monday. In reality, rankings never change. Some of the site’s top ranked content isn’t searched for on weekends, so the dataset used to calculate average position changes. See the example below:
| Keyword | Weekday Impressions | Position | Weekend Impressions | Position |
|---|---|---|---|---|
| Apples | 20 | 1 | 0 | N/A |
| Bananas | 5 | 2 | 5 | 2 |
| Oranges | 10 | 2 | 10 | 2 |
| Pears | 5 | 3 | 5 | 3 |
| Weekday Average Position: | 1.625 | Weekend Average Position: | 2.25 |
As you can see, demand for “apples” (a popular and highly-ranked keyword) drops over the weekend, removing its rank position from the average position calculation, resulting in a lower average position. It’s not that “apples” actually lost its rank position (it remained in position 1) but that people stopped searching for it.
Alternative Solutions
This is why it’s valuable to track rankings with a third-party tool (STAT, SEMRush, BrightEdge, etc.). These third-party rank tracking tools can provide a more stable view of ranking performance, because they collect data independently from demand.
Alternatively, if you’re facing an average position drop and don’t have a third-party tool to verify the changes, you can export GSC keyword data from before and after the average position change, then use a VLOOKUP in Excel for an apples-to-apples comparison of rankings, instead of looking at the data in aggregate. This strategy is most useful when you have at least a few days of data to work with. Using only a single day of data can be misleading.
Of course, some average position drops truly are due to changes in rank position. We tend to see the type of shift described here over weekends, during periods of low seasonal demand, and after sale events.
Hi Justin! Excellent article. Question: “As you can see, demand for “apples” (a popular and highly-ranked keyword) drops over the weekend, removing its rank position from the average position calculation, resulting in a lower average position.” The “lower average position” –> does this refer to the “N/A” or the overall average position of “2.25”? If “2.25”, then “lower” refers to lower down the SERP, correct? Higher indicates higher up in ranking near the top of the page, yes? Just want to make sure I understand this correctly.
Hi, Ellen. Yes, generally we refer to a “higher” ranking as being higher on the page (so 1 is “higher” than 2). In the case you’re referring to, Apples has 0 impressions over the weekend, so its rank isn’t included in the average calculation, which leads to a lower (on the SERP) ranking.