DNHQ Research · May 2026
A first-look at how Google AI Overviews are reshaping search results across 18 industries, 116,918 SERPs analysed, and 1.1M+ organic results parsed - and what it means for SME owners and the agencies that serve them.
The audit
116,918 SERPs and 1.1M+ organic results analysed across 18 industries and 6 Australian cities in May 2026. Built from primary SERP capture, not US data extrapolation - the first empirical baseline for how Google AI Overviews behave in the Australian market.
Why we did it
AI search is reshaping the SME playbook faster than the industry can describe it. Existing studies came from US data with US verticals - Australian operators were left guessing whether the same rules applied here. So we ran the audit ourselves, to move the conversation from speculation to evidence.
Who it's for
If you run an Australian SME, advise on its growth, plan its marketing, or write about how search is changing - this is your ground truth. Every percentage carries a 95% confidence interval. Every finding is reproducible from the underlying dataset. Quote it. Cite it. Argue with it.
116,918
SERPs analysed
37.8%
Overall AIO trigger rate
18
Industries covered
~387K
AIO citations
19
Findings
How to read this report
The 19 findings are grouped into three thematic Parts in numerical order. Each tile shows its chart and headline by default - tap any tile to read the full narrative and action recommendations.
The Five
Five takeaways the SME community should act on. Each one carries a 95% confidence interval and is sourced from the underlying 116,918-SERP dataset.
The Full Set
Three thematic Parts. Each finding shows its chart and headline by default - tap any tile to read the full narrative and action recommendations.
The How
The audit captures 116,918 SERPs from a balanced design engineered for cross-vertical comparability: an equal stratified sample of 833 commercial keywords per vertical × 18 verticals = 14,994 unique queries, each crawled at organic depth 100 across six Australian cities in May 2026.
Verticals were sampled at equal quota - not proportional to market size - so each industry carries the same statistical weight. The 833-per-vertical figure was set to guarantee a minimum cell size of n ≥ 100 in every cross-tab cut we report (e.g. "Real Estate × local intent × mobile"), keeping confidence intervals tight even at the long tail of the distribution. Keyword universes were sourced from commercial-intent search data, deduplicated against brand and adult terms, and filtered to AU search demand.
Three metro/regional pairs span the eastern seaboard: Sydney/Newcastle (NSW), Melbourne/Geelong (VIC), and Brisbane/Toowoomba (QLD). Each keyword was issued from each city to surface genuine geographic divergence in the SERP - important because Google increasingly localises both organic results and AI-generated answers.
70% of impressions were captured on Android mobile - the dominant share of Australian search - with 30% on desktop as a control. Each impression ran through an async SERP capture pipeline with asynchronous-AIO retrieval enabled. Without this, an estimated 30-50% of AIOs that take longer than 2 seconds to render are silently dropped from the response payload, which would have materially understated the trigger rate. Parser v1.1.0 captures both side-panel and inline citations, including domain, URL, rank, and citation type.
Every percentage in this report carries a 95% Wilson confidence interval (tighter and better-behaved than naive ± formulas at the edges of the distribution). Findings whose CI overlaps zero are flagged as "within noise" in the narrative. Per-vertical breakdowns are only reported where n ≥ 100 SERPs in the relevant cell. Snapshot data was captured 11-12 May 2026.
International Award Winners
Across Australia & Industry Wide
With Over 165 Reviews
Our strategy team will work your business goals and provide you a clear roadmap on how to get there!
