SERP Analysis: How to Read a Results Page Before You Choose a Page Type
A practical how-to for auditing live SERPs: capture cleanly, inventory the features that decide what 'position 1' means, and translate what you see into a page archetype decision.

SERP Analysis Is a Reading Skill, Not a Tool Report
Most teams treat a results page as a scoreboard: export ten blue links, note who sits at position 1, move on. That fails now, because a single page can hold a local pack, a video shelf, a People Also Ask cluster, a featured snippet, and an AI Overview before the first organic link. In that mix, "position 1" is a moving target.
SERP analysis is the discipline of reading what a results page says about intent, then translating that into a page-type decision before anyone writes a word. This guide walks a repeatable six-step process using one running scenario: a South African solar installer working through a 200-query shortlist, contrasted with a dealer-led SERP such as "used bakkies for sale", where the local pack dominates and the page decision changes entirely.
Every SERP you inspect should end in one of three decisions: build a new page type, merge into an existing page, or skip the query. If you finish with no decision, you analysed nothing.
What You Need Before You Analyse Your First SERP
Three prerequisites, none of them paid tools:
- A query shortlist with commercial context. For the solar installer, 200 queries tagged by service (residential installation, commercial rooftop, inverter repair) and by province, since load-shedding-driven demand clusters geographically in South Africa.
- Search Console access. The Performance report gives real impressions, clicks, and average position per query, beating any third-party volume estimate. Mine it before spending a rand on tooling.
- A clean browser profile with no plugins injecting competitor data, signed out of Google, with a planned device/location pair.
A reasonable tool budget for a small team is roughly ZAR 1,500-4,000 per month (estimate) for rank tracking plus volume data - but only after Search Console has been mined.
Step 1 - Capture the SERP Cleanly and Record the Conditions
Capture hygiene decides whether your analysis is usable. Record device, approximate location, date, language, and sign-in state for every capture. Google personalises heavily, so two analysts on the same query can screenshot different pages.
- Capture on mobile and desktop. For "solar installers Pretoria", mobile often surfaces the local pack higher, which changes your page decision.
- Set location explicitly for local queries. "Used bakkies for sale" in Johannesburg produces a dealer pack; with no location signal it produces classifieds and national portals.
- Note the ad block. Google retired right-hand desktop ads in 2016, so top-of-page paid units interleave with organic features - screenshot the whole viewport.
A SERP captured while signed in, with a VPN on, and a rank-tracker plugin injecting overlays is not evidence. Re-capture before you conclude anything.
Step 2 - Inventory the Features and Who Owns Them
List every non-standard element above and between the organic links, and record who owns each. Feature ownership is the signal: if a featured snippet for "solar inverter lifespan" is held by a manufacturer's support page, a thin service page will not displace it.
Record People Also Ask questions verbatim - they reveal the sub-intents Google expects. On "solar panel installation cost South Africa", PAA typically pulls in system cost, installation time, and finance availability; each is a distinct page or section candidate.
Step 3 - Read Intent from the Shape of the Page
Ignore positions one to three and ask: what kind of page dominates? For "residential solar installation Cape Town", it is likely local pack plus service pages plus a pricing guide - a mixed commercial-investigation SERP. For "used bakkies for sale", it is inventory-led: dealer category pages and classified portals, with a strong local pack.
Watch freshness too. Query-deserves-freshness (a concept Google described in 2007) explains why some SERPs churn: "solar tax incentive South Africa" moves every time policy changes, while "what is a hybrid inverter" stays stable for years. Freshness-sensitive SERPs demand different page maintenance, not different page types.
Step 4 - Decide the Page Archetype the SERP Is Asking For
Now write the decision down. The solar installer's 200 queries typically resolve into four archetypes: a location service page, a cost/pricing guide, a comparison page (hybrid vs off-grid), and an FAQ-style explainer. One query does not automatically mean one page - a cluster of near-identical queries should consolidate into a single authoritative page rather than spawning near-duplicates.
Write the archetype and the owning URL (new or existing) next to each query before moving on. That column is the only output of Step 4 that anyone downstream will use.

