Tracking Image and Video Search After Console Interface Changes
When Google changes Search Console reports or search-type filters, a chart can still look familiar even though it shows a different kind…
In Google Search Console, a search type tells you which kind of Google Search results a Performance report covers, such as Web, Image, or Video. Selecting a different search type changes which search traffic the report measures. If a dashboard keeps the same chart line after the filter changes, readers can mistake two different measurements for one continuous trend.
Keep a separate time series, a record of values over time, for each search type. Record the property, date range, other filters, and aggregation method for each one. Compare types side by side when useful, but label any combined total as a custom roll-up. That keeps a filter change from silently changing what the dashboard’s numbers mean.
Search Console’s search types track different areas of Google Search. The API also distinguishes categories such as Discover, the News tab, and Google News. Because labels and availability can differ between the browser report and connected data sources, record the exact type your source returns. Google describes the available types in its Performance report documentation and Search Analytics API reference.
A search type is different from a search appearance. The type identifies the surface, while an appearance identifies a feature of a search result. Other report settings matter too: for example, grouping by search appearance changes the table’s aggregation to page level. Check the selected dimensions alongside the search-type filter before comparing chart and table values.
When you store historical data, include the search type with each data point, not just as a dashboard control. Identify each series by its Search Console property, date, search type, metric, aggregation level, and any other filters or grouping dimensions. When available, store a data-status field so the dashboard can distinguish finalized dates from recent or incomplete data.
For example, store Web clicks and Image clicks as separate daily series. Display them as separate lines or panels. If the dashboard also sums them, label the result as a combined total across those selected surfaces. Google’s bulk-export guidance includes a date and search-type field, but its tables can contain repeated keys, so aggregate rows rather than treating every row as a unique daily total. For related details on type-specific collection, see reporting image and video search.
After you change a search-type filter, check that the dashboard still describes the report you intend to show:
Search Console’s API supports requesting data by type and date, but its responses contain top rows rather than a guaranteed complete list of every row. The report table also has row limits, and rare queries can be omitted for privacy. As a result, a sum of displayed table rows or API rows does not necessarily match the chart total. See Google’s API data-retrieval guidance and Search Console data notes.
For historical backfills, use the report or API for dates before bulk export began. Bulk export supplies ongoing data and does not backfill earlier history, as Google explains in its export setup guidance. Record the source and date convention with each imported period; otherwise, switching collection methods can create an apparent search trend.
Start a new series if the search type or aggregation level changes. Start one when the property definition, date convention, or metric calculation changes, too. If a chart crosses that boundary, mark the date and explain what changed. Otherwise, show the histories separately so a step up or down does not look like a change in traffic.
Calculate a period click-through rate (CTR) from total clicks divided by total impressions:
period CTR = SUM(clicks) / SUM(impressions)
Do not take a simple average of daily CTR values. Days with different impression totals would receive equal weight. Likewise, do not add average-position values or average them without the underlying data and a calculation appropriate to the aggregation. Google calculates position differently at property and page levels, so a dashboard should not splice those methods into one continuous position line. Its aggregation explanation describes why property- and page-level results can differ.
Comparing search types is useful when you have a clear question and the dashboard shows each type separately. But do not assume that a metric has an identical interpretation across every search surface. Preserve the source type and aggregation in the chart labels or metadata, and document the definition whenever you create a combined total.
Give Vroni a GitHub issue, bug report, spec, or rough idea. It reads the repo, plans the change, writes code, runs checks, and works toward a review-ready pull request.
Take a look at vroni.com