SERP analysis: reading the results page and turning it into a plan

SERP analysis means reading, query by query, what Google actually displays: organic results, SERP features, ads and AI Overviews. It adds the "market" dimension to your website analysis: intent, winning formats and meaningful competition. This guide extends our overview of SEO tools to diagnose and track your visibility. It shows how to read the results page, decide whether to defend, create or leave a query, and turn each finding into a page to produce or update.

What a SERP tells you about search intent

The results page answers one question: what does the user typing this query expect? Before looking at who ranks first, look at which page types and formats Google promotes. They reveal the intent better than the keyword itself.

Name the dominant intent and its variants

Start by naming the dominant intent visible in the top 10, then list its variants, driven by People Also Ask or by pages with different angles. In B2B, the same query often combines three intents:

• Informational: explain, define, frame a topic
• Comparative: criteria, alternatives, trade-offs between solutions
• Transactional: service, demo, integration, rollout.

Then segment your queries by buying-stage maturity (discovery, comparison, selection) and map each one to an offer and a target page.

Set your observation context

Comparing two SERPs without context is like comparing two different markets. Your browsing and your location change what you see. Document every observation:

• Country, language and Google domain (google.co.uk or google.com)
• Device: modules and their order change between mobile and desktop
• Precise location (city, region), decisive as soon as a local pack appears
• Date and period, to spot the query's seasonality and volatility.

Work in a private browsing window, in a stable environment, and keep a screenshot together with the list of modules present. You can then compare before and after without noise.

SERP features: what they reveal and what you do about them

Every module that takes up the top of the page competes directly with your blue links, even if you rank well. The table below links each feature to the intent it signals and the action it calls for.

The module that weighs most is the AI Overview: >50% of Google searches display an AI Overview (Squid Impact, 2025). Its presence changes the value of a ranking, since the click-through rate of the first organic position then falls to 2.6% (Squid Impact, 2025). More benchmarks are available in our GEO statistics.

At the time of writing, Google Search Console counts AI Overviews and AI Mode (available in France since summer 2026) in the Performance report without separating them: only observing the SERP tells you whether the module is triggered.

Featured snippets: win the snippet without losing the click

Featured snippets deliver strong visibility, but they can also answer for you. To win one, check four points:

• a one- or two-sentence definition at the start of a section
• a numbered list when the query implies steps
• a table when the intent is comparative
• sourced and dated data, internal or external.

An overly complete snippet reduces the need to click. Accept it if the goal is awareness, reassurance or citation, on top-of-funnel queries. If the goal is conversion, add a logical post-click next step (template, checklist, detailed example) that the snippet cannot replace.

People Also Ask: a plan of sub-intents

People Also Ask reveals the sub-questions Google considers relevant around the intent. Rather than a long, repetitive FAQ, use them to organise the page into reading levels: a direct answer of two or three sentences, then criteria, steps, use cases and limits, and finally internal links to deeper resources.

Reading the top 10: what ranking pages have in common

Once the modules are identified, analyse the "winning shape" of the ranking pages. Screen each top-10 result against six criteria:

• Format: guide, list, comparison, service page, tool, video
• Length and depth: sub-topics that show up consistently
• Angle: the promise made in the title tag and the visible snippet
• Entities: notions, brands, standards and concepts cited from one page to the next
• Freshness: update date, how current the examples and data are
• Authority: displayed expertise (author, references, sources) and popularity.

What appears in most pages forms the expected baseline; what appears nowhere forms your possible angle. To connect these observations to your targeting, rely on your keyword research for SEO and a query-to-page mapping approach. When the top 10 mixes incompatible formats, search intent analysis helps you choose the one you target.

Assess the useful competition

What matters is the competition that matches your intent and your offer: publishers, media, service providers and directories do not call for the same response. On authority, the referring domains of top-10 pages scope the effort: a SERP dominated by heavily linked pages is expensive to win.

Finally, look for winnable angles: a clear definition, a process, a quantified example, a comparison table, a risks and limits section, or a recent update that the ranking pages lack.

AI Overview triggered or not: a prioritisation criterion

The presence of an AI Overview changes the value of a position. On a query that triggers one, the first organic position captures few clicks and being cited in the synthesis becomes the goal. On a query that does not, the organic click keeps its classic value.

Defend, create or leave a query

Cross the presence of the AI Overview with your current situation on the query to choose between three decisions.

The same logic applies to other generative engines. ChatGPT, Perplexity and Gemini cite sources, and a brand is "cited" when it is named or linked in the answer. To understand these mechanisms, see our guide to generative engine optimization (GEO).

Make the page citable

For AI answers, one question comes first: is your page easy to summarise correctly, without distortion, and with reusable sources? Three practices increase this "citability":

• provide concise answers at the start of sections (definition, recommendation, thresholds, steps)
• state criteria and thresholds rather than generalities, backed by sourced data
• structure with lists and tables, plus section summaries, to make extraction easier.

From findings to an execution plan

The final output of a SERP analysis is not a screenshot: it is an executable plan. Each finding leads to an action and to a page to produce or update.

Turn each finding into an action

The table below lists the most common findings and the decision each one calls for.

Prioritise: impact, effort, risk, dependencies

Not every idea is worth a sprint. Prioritise using a short grid that marketing, content and technical teams can all understand:

• Expected impact: clicks, leads, citations, or protection of existing revenue
• Effort: production time, approvals, technical or CMS dependencies
• Risk: cannibalisation, regression, semantic dilution
• Dependencies: data, subject-matter experts, assets to gather (video, cases, figures).

Formalise the deliverables

For each selected query, formalise at least four deliverables, which hand the analysis over to the teams who execute:

• an editorial brief: angle, outline, expected answers, sources and evidence
• an on-page checklist: title tag, headings, intro, sections, lists and tables
• an internal linking plan: source pages to the target page, consistent anchors
• tracking KPIs: impressions, click-through rate, average position, clicks, leads.

Measure the impact and re-run the analysis

Measure at the page and query level, not only at the keyword level. Also monitor cannibalisation: when several URLs share the impressions of the same query, consolidate. Each KPI reads through both an SEO lens and a GEO lens.

Connect the SERP to business value

In B2B, being visible is not enough: you must connect the SERP to the click, then to engagement and conversion. A SERP can deliver impressions without clicks (AI Overview, complete snippet) or clicks without pipeline (too top-of-funnel). So pair your observation with Google Search Console and Google Analytics 4, whose session sources isolate traffic from chatgpt.com or perplexity.ai and tie it to conversions.

When to re-run a SERP read

Re-run your analysis when the results page changes, not only when your rankings move. Triggers: an AI Overview or a local pack appears, the SERP shifts towards video, new players reach the top of the page, or strong instability. For critical queries, a quarterly cadence is a baseline, complemented by a monthly re-read of volatile SERPs.

Equipping your SERP analysis

Start simple: manual observation in a fixed context, plus Google Search Console to link pages, queries, impressions, clicks, click-through rate and average position. This baseline avoids confusing estimated data with real performance. Once your query portfolio grows, a tool needs to take over three functions:

• SERP capture: keep the results page by query, country, device and location, independently of your own browsing
• Feature detection: know, query by query, which modules are displayed, including the AI Overview
• Tracking over time: history of rankings and modules, to spot the changes worth re-reading.

A SERP tool excels at diagnosis, but it does not execute for you: production, approval and measurement still need organising. Tracking over time is the job of an SEO rank tracker, which records average position by keyword, device and location at a set frequency.

Depending on your situation

Your starting point depends on what you already know:

• You have not yet diagnosed your site: start with a website analysis before comparing your pages with the SERP
• A specific page is falling behind the top 10: run an SEO page analysis, with an on-page checklist
• You are torn between several intents on a query: go back to reading the features and the top 10 before producing
• You are choosing a platform: start from the functions described in our SEO tools guide rather than from a list of brands.

To spot these opportunities across your whole portfolio, the Incremys Opportunity Analysis module records, by keyword and by page, the SERP, the presence of an AI Overview and the gap with competitors.

FAQ on SERP analysis

What is a SERP analysis?

A SERP analysis is the process of studying the results page Google shows for a given query. It assesses intent, the formats promoted, the useful competition and ranking difficulty, according to the observation context (country, language, device, location). Its purpose is to decide what to create or update, and how to differentiate from pages that already rank.

How do you analyse a SERP methodically?

Always follow the same order. Set the observation context, name the dominant intent and its variants, record the features present, read the top-10 pages (format, angle, evidence, authority), then assess the useful competition. Finish with concrete deliverables: a brief, an on-page checklist, an internal linking plan and KPIs tracked in Google Search Console.

How do you win a featured snippet?

Answer the target question in one or two sentences, directly under a clear subheading. Use a list when the query implies steps, a table when it is comparative, and back the answer with sourced data. Then check that the snippet does not cost you clicks that matter for conversion.

How do you use People Also Ask without diluting the topic?

Treat these questions as a plan of sub-intents, not a list to copy. Group them by theme, answer briefly on the main page, then link to deeper pages through internal linking. This way you cover long-tail demand without turning the page into an endless FAQ.

How do you adapt SERP analysis for AI Overviews?

Add a citability lens to the ranking lens. Note whether the query triggers an AI Overview and whether your brand is named or your domain cited in it, then decide to defend, create or leave. Make the page extractable with direct answers, tables and sourced data, and measure its value beyond the click.

How often should you redo a SERP analysis?

Redo it as soon as the results page changes: an AI Overview, a local pack or videos appear, or new players reach the top of the page. Across a portfolio of business queries, a quarterly cadence is a solid baseline, complemented by a monthly re-read of the most unstable SERPs.

How do you connect SERP analysis to business results in B2B?

Link each analysed SERP to a target page, then track a simple funnel: impressions, click-through rate, clicks, engagement, leads and opportunities. Watch for cases of visibility without value, such as impressions without clicks or clicks without conversion, to make informed trade-offs on format, targeted intent and the post-click promise.