Discussion
  • AI Search

Should AI Overview, AI Mode and Discover AI impressions be reported together or separately? Do they show very different click behaviour for you?

Published 1 reply

Discussion starter: an AI search question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.

Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.

1 reply

  • Niraj Raut
    Niraj Raut
    Original poster

    Report them separately for analysis and together only as a headline, with the breakdown underneath. AI Overviews, AI Mode and AI features in Discover reach people in different situations: a searcher scanning results, a user in a conversational session, and someone browsing a feed. Their impressions mean different things, and they respond to different content. On clicks, the honest answer is that Search Console’s “Search generative AI performance” reports show impressions only, with no clicks, click-through rate or queries, so you cannot compare click behaviour between the three from that data. AI Mode clicks are counted within the web search totals without a separate breakdown, and Discover clicks appear in the Discover report. Comparisons of click behaviour therefore have to be inferred from other data, and should be labelled as estimates.

    What each surface represents

    Surface User situation What an impression suggests
    AI Overviews Searcher sees an AI summary above traditional results Your page was used or shown as a source alongside regular listings
    AI Mode Conversational search session, often with follow-up questions Your page appeared in a conversational answer
    AI features in Discover Browsing a personalised feed Your content appeared in an AI-generated element in the feed

    A reporting structure

    1. Headline: total AI impressions for the period, with a trend, clearly labelled as impressions without clicks.
    2. Breakdown by surface: impressions for each, with trends.
    3. Top pages by surface: which pages appear most in each, since they often differ.
    4. Context metrics: web search clicks, Discover clicks and conversions for the same pages, to see whether AI exposure coincides with changes in traffic.

    Estimating click behaviour indirectly

    • For pages with rising AI Overview impressions, compare web search click-through trends with similar pages without AI impressions.
    • For Discover, compare Discover clicks for pages with and without AI feature impressions.
    • Use published research on AI summaries and clicks as context, such as Pew Research Center’s finding that users clicked traditional results less often when an AI summary appeared.
    • Label any resulting figures as estimates, with the method stated.

    Hypothetical example:

    A publisher’s AI impressions double over a quarter. Broken down, most of the growth comes from AI Overviews on its explainer articles, while AI Mode impressions come from a different set of pages: detailed guides that answer follow-up questions. Discover AI impressions are small and concentrated on news. Web search clicks to the explainers fall slightly, while clicks to the detailed guides hold. The publisher reports the three surfaces separately, invests in the detailed guides that do well in AI Mode, and treats the explainer losses as clicks absorbed by AI Overviews.

    Watch for this:

    Adding AI impressions to web impressions, or presenting AI impressions as if they were visits, overstates reach. Keep AI impressions in their own section, and never convert them into estimated traffic or revenue without a clearly stated method.

    Country and device splits matter too

    The generative AI reports can be filtered by country and device, and those filters often reveal more than the totals. AI features launched in different markets at different times, and AI Mode usage may differ between mobile and desktop. A rise in AI impressions concentrated in one country, or on one device, points to a product rollout rather than a change in your content. Break the surfaces down by country and device before drawing conclusions, and annotate known rollout dates in each market so trends are read against them.

    Keep the reporting honest about its limits

    Every AI section of a report should say plainly that the data shows impressions, not visits, and that clicks from AI Mode are included in web search totals without being separable. Stakeholders who understand those limits interpret the numbers sensibly; stakeholders who do not will read impressions as audience reached and expect traffic to follow.

    Why the surfaces deserve different content strategies

    If AI Overviews surface your broad explainers, the content question is whether those pages still earn clicks for the people who want more depth. If AI Mode surfaces your detailed guides in conversational sessions, those pages are serving follow-up questions, and depth and specificity matter. If Discover’s AI features show your articles, freshness and topic expertise matter. Separating the surfaces is what makes those distinctions visible.

    Report AI impressions by surface, with a labelled total as the headline, pair them with clicks and conversions from the same pages, estimate click effects only indirectly and label them as estimates, and use the differences between surfaces to guide which content to invest in.

    Need help with this on your own site?

    Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.

    See SEO services

Add a reply

Replies are open to approved members. Log in or apply to join.