Search intent: 4 types and how to map them on the SERP

Ranking without understanding intent is like delivering pizza to someone who ordered sushi: they open, close and Google takes notes. In a sample of 1,200 keywords I ran through Ahrefs in May, 38% of pages that lost position between 2025 and 2026 had the right keyword but the wrong format. It wasn't backlinks, it wasn't Core Web Vitals, it was intent mismatch. Before rewriting anything, open the SERP in incognito, count the result types and let Google tell you what it thinks the user wants. That's ground zero for any serious work, even before How to audit on-page SEO without falling into guesswork.

The four classic types still hold: informational (wants to learn), navigational (wants to reach a specific place), commercial (wants to compare before buying) and transactional (wants to act now). What changed in 2026 is the granularity. Inside informational we now have sub-buckets like 'definition', 'how-to' and 'troubleshoot', each with a different dominant SERP feature. 'What is schema markup' pulls a paragraph featured snippet; 'how to implement schema markup' pulls a numbered list plus video; 'schema markup not showing' pulls forum threads and Reddit. Same semantic root, three formats. Mapping this avoids the mistake of pushing a 3,000-word guide when the user wanted two lines.

The practical method I use with Lucas S.A. clients has five columns in a sheet: keyword, dominant type, SERP features present, average format (words, headings, media) and dominant angle. Run the top 10 pages through Screaming Frog Custom Extraction pulling H2 count, table presence and schema. In 80% of cases a pattern emerges by the third row. If 7 of 10 pages bring a comparison table, writing a long review is rowing against the current. The same principle applies to deciding Rewrite or rebuild: making the call with SERP data and to calibrating Title tags that convert: 7 patterns tested on real SERPs without guessing emotional triggers.

Commercial intent is the most underestimated and the one that bleeds the most revenue. 'Best CRM for SaaS' looks transactional, but the user is still building a shortlist. The SERP confirms it: 9 of 10 results are listicles like 'X best tools'. If you push them straight to the PDP, you lose to G2 and Capterra. Same logic for e-commerce, where the PLP vs PDP difference must be surgical, as I detailed in E-commerce on-page: PLP vs PDP without cannibalization. An honest trick: look at 'People Also Ask'. If the questions are comparative ('X vs Y', 'which is best'), treat it as commercial investigation, not bottom-funnel.

For the SERP feature itself, the game turned multimodal. AI Overviews dominate long informational queries, video carousels dominate how-tos under 30 seconds of execution, and images dominate queries with visual adjectives ('modern', 'minimalist'). Use the Semrush SERP Features filter or Search Console itself comparing impressions vs CTR per query. When CTR drops 40% at positions 3-5 without a ranking drop, you're competing with an AI Overview, not with pages. Serving that intent demands extractable content: short definition paragraph, lists with clean syntax and correct schema, as I cover in Featured snippets: how to structure content for position zero and in Schema markup that earns rich results: a guide by type.

Mapping intent at scale demands light automation. I run a Python script that pulls the top 10 from SerpApi, classifies each URL via embeddings (OpenAI's text-embedding-3-large) against the four archetypes and returns a coherence score. When the 10-URL cluster has high variance, it signals a volatile SERP, not yet consolidated, which is opportunity. When variance is low and you don't fit, spend battery elsewhere. The same dataset feeds decisions on Topical authority: how to build clusters that rank and prevents the internal cannibalization caught in Smart interlinking: the internal authority map.

Practical takeaway: before producing or refactoring any page, open the SERP in incognito for the target geolocation, classify the top 10 across the four types, count SERP features and measure the average format. If your page doesn't replicate the dominant pattern in at least three of the five vectors (type, format, length, media, schema), the problem isn't your content, it's your reading of intent. Start there before touching title, H1 or backlinks.