The week-by-week storyline of where you sit in Google’s organic results: rising, flat, or slipping.
At a glance
Position Trend plots your average organic position from Google Search Console as a time series across the selected period, so SEO and content teams can see direction at a glance rather than a single snapshot. Average position is reported on Google’s scale where 1 is the top of the results and higher numbers are further down, so a falling line is good news and a rising line is a warning. The card is a Supporting Metrics-class trend: it watches the slope, not just the latest value.
Calculation
Search Console reports average position as the mean of the highest position your site reached for each query and date combination, weighted by impressions. For a single result, the position is its rank on the results page when it was shown (counting from the top, so position 3 means the third organic listing). Search Console then averages those positions across every impression in the period. Vortex IQ reads thisposition metric from the Search Analytics API for each interval in the selected window and draws the sequence as a line. Two points worth holding in mind:
- The metric is impression-weighted, so a single high-volume head term sitting at position 4 moves the line far more than a hundred long-tail terms sitting at position 40. A “worsening” average can simply mean you started ranking (and so collecting impressions) for a new batch of low-position long-tail queries.
- Position is reported to one decimal place. Small wobbles of 0.2 to 0.5 between intervals are normal sampling noise, not movement worth acting on. The alert bands (
gsc_position) exist to separate noise from a real slide.
Worked example
A UK homeware brand running editorial content alongside its product catalogue connects Google Search Console. The dashboard is set to the last 12 weeks, web search, all devices. The Position Trend line reads:
Three observations the SEO lead drew from this:
- The slip is real, not noise. A move from 8.5 to 11.2 in one week is roughly 2.7 positions, far outside the 0.2 to 0.5 wobble band. The line crossed out of the green zone toward amber, which is exactly what the
gsc_positionalert rule is calibrated to flag. - Splitting the line by query type found the cause. Re-running the view filtered to non-branded queries showed the slide concentrated in a cluster of buying-guide articles. Branded queries held at position 1.2 throughout. The cause was a Google core update that re-ranked editorial content; the product pages were untouched. Pair with Branded vs Non-Branded to make this split routine.
- The weighted-average trap. Part of the drop was benign: the brand had just published 30 new long-tail articles that started ranking around position 25 to 35. Those new low-position impressions dragged the average down even though nothing existing had moved. Position Distribution confirmed the head terms were stable and the new mass sat in the tail.
Sibling cards SEO teams should reference together
Reconciling against the source
Where to look in Google Search Console:Open Performance on Search results, tick the Average position metric, and set the date range to match the dashboard window. The chart Search Console draws is the native equivalent of this card. The headline “Average position” figure above the chart is the period mean for the same scope.To reconcile cleanly, match four things between the two views:
You can also pull the same numbers programmatically from the Search Analytics API (
searchanalytics.query with position requested), which is the exact source Vortex IQ reads. API and UI agree when the dimensions and date range match.
Why the figures may legitimately differ:
- Data delay. Search Console data is typically 2 to 3 days behind. The most recent interval on the card may still be filling in and can shift slightly over the following days.
- Anonymised queries. Search Console drops rare queries to protect user privacy, so query-level totals never fully reconcile to the property-level average. The property-level average position you compare against is the safe baseline.
- Row caps. The Search Console UI caps tables at 1,000 rows. If you reconcile by exporting query rows and re-averaging, you will be missing the long tail and your hand-computed average will not match. Always compare against the property-level “Average position” figure, not a re-average of exported rows.
- Weighting. The average is impression-weighted by date and query. A naive average of positions you see in a table is unweighted and will not match.