SEO reporting has a familiar argument: ranking improved, but traffic did not; traffic improved, but position looked worse; a page appeared in an AI feature, but the old chart did not explain what happened. A September 10, 2026 Search Engine Roundtable post captured the problem through a comment from Google’s John Mueller: the old one-through-ten idea is hard to map in modern search results.
That should not turn into resignation. Search Console still gives useful evidence. But average position needs to be treated as a diagnostic clue, not as the executive KPI by itself. The practical answer is a diagnostic checklist that starts with the business question before it looks at the rank number.
What average position can still tell you
Google’s Search Console help defines average position as the position of the topmost result from your site, with aggregation rules that change between chart and table views. That means a site-level number can hide page-level and query-level movement. It also means the same query may look different depending on device, location, personalization and the search surface.
The number remains helpful when you compare like with like: the same query group, same country, same device and same page set over a sensible period. It becomes misleading when it is used as a single rank statement across AI Overviews, AI Mode, rich results, classic blue links and blended surfaces.
Why AI search changes the reporting conversation
Google’s generative AI performance reports in Search Console are meant to show visibility inside AI features such as AI Overviews and AI Mode. Google says this data is included in overall performance reporting while also giving site owners a dedicated view for AI-feature visibility.
The practical issue is that visibility, citation and click are no longer the same event. A page can be cited in an AI answer and receive fewer clicks. A brand can gain answer visibility without the same landing-page session volume. A page can lose a classic position but still influence consideration through an AI surface.
A diagnostic sequence for SEO teams
Start with the business question. Are you explaining a traffic drop, a visibility gain, a lead-quality shift or a content investment decision? The right chart depends on the decision.
Segment before interpreting. Break Search Console data by page type, query intent, country, device and search appearance where available. Do not let a homepage, a news post and a product page share one position story.
Separate visibility from demand. If impressions rise but clicks fall, ask whether the query demand changed, the result surface answered more of the question, or your snippet became less compelling.
Connect AI visibility to onsite behavior. When AI-feature data is available, compare the visible pages with landing-page engagement, assisted conversions, branded search and direct traffic. Treat it as one influence signal, not a complete attribution model.
Keep a control set. Maintain a group of stable queries and pages where the search layout changes less often. This helps distinguish your SEO work from SERP design changes.
How to report it to executives
Replace “we moved from position four to position two” with a cleaner narrative: visibility changed for these query groups, clicks changed for these pages, and business outcomes changed here. If position moved in the opposite direction, explain whether aggregation, AI surfaces, device mix or query mix caused the conflict.
A good report should include a small number of decisions: refresh this page, consolidate these two URLs, improve snippet clarity, add source-level expertise, protect a page that still converts, or stop chasing a query that now answers itself in the results.
The new SEO habit
The old rank number was comforting because it made search look linear. Search is less linear now. The discipline is not to abandon measurement, but to ask sharper questions of the data.
Average position still belongs in the SEO dashboard. It just should not sit alone. Pair it with impressions, CTR, AI visibility, page-level engagement, conversions and a written explanation of the search surface. That is how a rank debate becomes an operating decision.
