Case study | SEO + Answer Engine Optimisation
How we launched a public services information site and grew organic clicks from zero to over 34,000 a month in three months
A new public-services information site launched from a bare domain. Three months in, the site was pulling 34,612 organic clicks a month and had been picked up by Google's AI answer surfaces as the top independent source on national identity document queries, ranked alongside government domains.
- Monthly organic clicks 34,612 (Month 3) from 402 (Month 1) 86x
- Monthly organic impressions 1.3m (Month 3) from 19,483 (Month 1) 67x
- Google AI Mode citations 1,927 answers #1 non-government source
- Search volume where the site is cited 2.4m searches / mo Ahead of gov.za (1.8m)
The client
A new public-services information website, built from scratch on a fresh domain, competing in a query category where government portals normally dominate. The site answers everyday citizen questions about national identity documents: how to apply, how to renew, what to bring, where to go, how much it costs. Every URL is a direct answer to a specific query rather than a long guide covering many.
The problem
Entering a query category where government domains are the default result is a hard SEO problem. Ranking a brand-new site next to gov.za on high-volume national-service queries in a matter of weeks is not something we would normally promise. There was a second layer to the problem: Google's AI Mode was already answering these queries directly without a click, drawing its citations from a small handful of trusted sources. If the new site was not inside that citation pool, no classic ranking position would matter. The build had to hit both surfaces at once, and it had to hit them before the site had any age, backlinks or brand equity to lean on.
The plan: build a site that answers questions AI models can extract
Five workstreams running in parallel from Day 1 of the build, not sequentially.
-
One URL per query, not one URL per topic
Rather than five long guides, the site ships thirty-plus tight pages, each answering one specific citizen question. Query-to-URL alignment is exact. This is how you get ranked for a specific phrase in classic organic AND get extracted as a source by an AI answer surface: the model can point to a single URL that answers the exact question a user asked.
-
Answer at the top, depth below
Every page opens with a direct answer in the first fifty words, structured as a short block an AI model can lift verbatim. Steps, prices, opening hours and requirements are formatted as bullet lists or short paragraphs, never buried inside long prose. Classic SEO best practice, and non-negotiable for AI citation.
-
Fast, mobile-first WordPress theme, purpose-built for the cluster
Bespoke theme, no page builder, no bloat. Core Web Vitals green from launch. Schema markup on every page. Sitemap regenerating on every publish. The site loads on 3G in under a second, which matters when the target audience is mobile-first and often on data.
-
Topical saturation over topical depth
We shipped the full cluster of citizen-question URLs in Weeks 1 and 2, not one page a month. Search engines and LLMs both weight topical authority higher when a domain covers a whole cluster comprehensively, and both surface authority earlier when the cluster ships as a batch rather than a slow drip.
-
AEO-native content structure
Every page has explicit TL;DR blocks, step lists formatted as numbered ordered lists, price tables and short FAQs. All of it is deliberate: LLMs summarising the page have short, extractable, factual chunks to lift into their answers. It is a formatting discipline as much as a content one.
The trajectory (first three months post-launch)
Organic clicks compounded roughly three-fold month-over-month while the site's average position held steady around six. Impressions grew even faster, which tells you the site was winning ranking positions across a widening set of queries, not fighting harder for the same ones.
Source: Google Search Console
Source: Google Search Console
Three months of compounding growth with the site's average position holding steady around six. Click volume roughly tripled every month, which is a sign the site was not fighting harder for the same queries but was winning rankings across a widening set of queries every week.
In parallel, Google's AI answer surfaces began treating the site as a primary source. It appears in 1,927 AI-generated answers with a combined 2.4 million monthly AI search volume across the relevant national identity queries, more than the equivalent government portal (gov.za, 1,155 mentions, 1.8 million volume) over the same measurement window. That is the twin outcome we optimise for now on new-site builds: classic organic ranking and AI citation, produced by the same content, from the first month.
Launching a new site into a competitive category?
Peak Digital builds websites specifically to rank fast in classic organic AND get cited by AI answer surfaces from Day 1. If you are planning a launch, book a free audit of your build plan before you commit to a theme.
Book a free audit