Measurement
Seasonality
Seasonality is the regular rise and fall of search demand that repeats at the same time each year, such as holiday shopping or tax deadlines. In SEO reporting it must be separated from the effect of the work, usually by comparing with the same period a year earlier or by seasonally adjusting the series.
- Industry term
By QueryRev ResearchPublished Updated6 min read
Why it matters.
Google lists it among the reasons traffic moves. Its guide to debugging drops in Search traffic says that changes in user behavior will change the demand for certain queries, either from a new trend or from seasonality throughout the year, so traffic may drop as a result of external influences. Its example is food: people search for diets in January, turkey in November and champagne in December, and different industries have different levels of seasonality[1].
The same guide shows how to tell a season from a problem. It says to set the Performance report’s date range to the last 16 months, to make sure a drop is not one that happens every year due to a festivity or a trend, and to compare the period with the same period a year earlier. It adds a check on Google Trends, to see whether a drop was only for your website or throughout the web[1].
Google Trends shows relative interest, not a count. It divides each data point by the total searches of the geography and time range it covers, and scales the results from 0 to 100 on a topic’s proportion to all searches on all topics. Terms with low volume can appear as 0, because Trends only shows data for popular terms[2]. It is the right tool to see the shape of a season and the wrong one to size it.
Some tools carry seasonality and some do not. In Google Ads’ Keyword Planner, forecasts take seasonality into account while historical metrics, such as average monthly searches, do not[3]. A figure for one month is a point on a curve whose shape the tool does not draw.
Statistics has a standard answer, seasonal adjustment. The U.S. Census Bureau produces and maintains X-13ARIMA-SEATS, software that generates ARIMA model-based seasonal adjustment as well as nonparametric adjustments from the X-11 procedure, with diagnostics of the quality and stability of the adjustments[4]. SEO teams rarely need that program, but they need its idea: estimate the regular part of a series and take it out before reading a change.
A fixed base makes a change easier to trust. QueryRev’s study measures the same panel, the first 10,000 places of one list, so that a change between editions is a change on the web and not in the list. It has one edition, run on 15 October 2026, and makes no claim about seasons[5].
How it is measured.
The simplest measure is the same period a year earlier. For a month, divide this year’s clicks by last year’s for the same month, and report the change. It removes the season if the season repeats, and fails if the site, the market or the weeks of the calendar changed.
A seasonal index is more careful. Divide each month’s clicks by the average month of that year, then average the ratios for the same month over several years. A value of 1.2 for April means April is typically one fifth above an average month. Divide a month’s figure by its index and the season is taken out.
Check the shape against Google Trends for the topic. If interest in the topic rose and fell in the same weeks, the movement belongs to the season. If your clicks moved and interest did not, something else did[1]. Mark events that look like seasons but are not: a campaign, a holiday that moves, an update.
Measured by
- Same period a year earlier
- Seasonal index by month
- Change after adjustment
A worked example.
Illustrative: an invented site, to show the method.
A garden store relaunches its category pages in March. Its clicks in April are 40% higher than in March, which looks like a win. A year earlier, with no work done, April was 45% higher than March.
Adjusted for the season, the change is about 3% lower, a modeled range of 8% lower to 2% higher: it divides this year’s March-to-April rise by last year’s, and one earlier year is a thin base for a season. That range crosses zero. The rise was the season. The data shows no sign of an effect from the work, and a range that crosses zero cannot say it has none. The store waits for more weeks and compares the changed pages with similar pages it did not touch.
What to do, step by step.
Pull at least 16 months of clicks for the page group, and two or more years where you have them.
Mark the peaks and troughs, and check the topic in Google Trends to see whether they belong to the season.
Compare each period with the same period a year earlier, or compute a seasonal index and adjust the series.
Write an adjusted change as a range, labelled modeled, with the method and the years it rests on.
To judge a change you made, compare the changed pages with similar unchanged pages over the same weeks, as a counterfactual baseline.
Note the dates of campaigns, holidays that move and updates, so that they are not read as seasons.
Mistakes to avoid.
Comparing with the previous month. Last month had a different season. Compare with the same period a year earlier, or adjust, before you call a change an effect.
Reading a Trends chart as a count. Google Trends scales each topic from 0 to 100 against all searches[2]. It shows the shape of demand, not how many searches there were.
Planning on a yearly average. Google says historical metrics in Keyword Planner don’t take seasonality into account[3]. A flat monthly figure hides the peak and the trough.
Trusting one year as the pattern. One earlier year may have had its own event. Use more years where you have them, and say how many a figure rests on.
Adjusting without checking the adjustment. Adjustment software provides diagnostics of the quality and stability of its result[4]. Whatever method you use, test whether it removes the season before you rely on it.
How QueryRev treats it.
When QueryRev verifies a change, it compares the changed pages with a baseline that carries the season: the same period a year earlier, similar pages that were not changed, or both. The lift is written as a range and labelled incremental, and a movement the baseline explains is recorded as seasonal, not as an outcome. See pre-post analysis, the guide to SEO forecasting and the SEO ROI Calculator.
Questions people ask.
How do I know whether a drop in traffic is seasonal?
Set Search Console’s date range to the last 16 months and compare with the same period a year earlier. If the drop happened as it did the year before, it is the season. Then check the topic in Google Trends. Google’s guide to traffic drops suggests exactly these checks.
Why not compare with the previous month?
Two neighbouring months can sit at different points of the year’s demand. A month-on-month change then mixes the season with the effect of the work. Compare with the same month a year earlier, or adjust the series first.
What is seasonal adjustment?
A statistical method that estimates the regular seasonal part of a series and removes it, so that other changes show. Agencies use programs such as X-13ARIMA-SEATS, which also give diagnostics of the quality and stability of the adjustment.
Sources and citation.
Primary sources only: the search engines, the AI vendors, standards bodies and QueryRev’s own research. Each is cited in the text by its number.
Debugging drops in Google Search trafficGoogle Search Central ·
FAQ about Google Trends dataGoogle Trends Help ·
About Keyword Planner forecastsGoogle Ads Help ·
X-13ARIMA-SEATS Seasonal Adjustment ProgramU.S. Census Bureau ·
State of AI crawler access, October 2026QueryRev Research ·
Cite this definition
A suggested citation. Its link goes to the definition, whose anchor does not change.
QueryRev Research. “Seasonality.” QueryRev AI glossary, 24 October 2026. https://queryrev.com/glossary/seasonality#definition
Related terms