eCommerce SEO Case Study: The Category-Page Playbook That Cut Customer Acquisition Cost by 40%
Retail SEO case studies tend to talk about content and links. This one is about category pages, because that is where the revenue was. An online retailer grew monthly organic sessions from 22,000 to 78,000, lifted organic search’s share of revenue from 14% to 41% and cut its blended customer acquisition cost by 40% in ten months, and most of that came through category and sub-category pages once three things were fixed: faceted navigation that had duplicated every category into countless URL variants, templates that carried nothing but a grid, and a range structure that did not match how people searched. This page sets out that playbook step by step, the order it was run in, when the acquisition cost started to fall, and how a retailer in Australia on Shopify, BigCommerce or Magento applies it. Figures come from the client’s Google Analytics 4, Search Console, ad platforms and order system.
The headline results
- −40%Blended customer acquisition cost
- 14% → 41%Organic share of revenue
- 22,000 → 78,000Monthly organic sessions
- Month 3When category pages recovered
| Metric | Before | After | Change |
|---|---|---|---|
| Blended CAC (all channels) | Baseline | 40% lower | −40% |
| Organic share of revenue | 14% | 41% | +27 points |
| Monthly organic sessions | 22,000 | 78,000 | 3.5× |
| Indexable category URL variants | Very many, from facets | Only those with search demand | Consolidated |
| Category template | Product grid only, slow | Content, guidance, fast | Rebuilt |
CAC is total acquisition spend (paid media, agency fees including SEO, and tooling) divided by new customers, from the client’s ad platforms, invoices and order system, comparing the month before kickoff with month ten. Organic revenue share and sessions are from Google Analytics 4 with purchases reconciled to the order system. The client’s name is withheld under NDA; screenshots are available on a video call under the same terms.
Who the client is
| Client snapshot | |
|---|---|
| Business | Mid-sized online retailer with a broad catalogue across many categories |
| Primary conversion | Purchases, with new-customer orders separated from repeat |
| Starting position | Category pages diluted by facet URLs, slow and empty; organic 14% of revenue; paid search buying the category demand |
| Engagement | Managed eCommerce SEO retainer, ten months at the time of these figures |
| How decisions were made | Facet and category decisions from search data; paid reallocation from a monthly CAC model fed by ad platforms, GA4 and the order system |
| Outcome | Blended CAC down 40%; organic the largest revenue channel by month ten |
An online retailer with a wide range across many categories, a small marketing team and a developer on staff. Growth had come from paid search and paid social until each extra order cost more than the one before; the finance lead set the target as a cost per new customer, and the category pages were the lever we proposed to hit it.
The challenge
The non-brand demand the retailer was buying through paid search was exactly the demand its category pages should have captured organically, and could not. The faceted navigation had created a very large number of crawlable URL variants, so each category’s authority was split across dozens of near-duplicates; the category templates loaded slowly and offered nothing beyond a product grid; and the range was organised by internal merchandising logic rather than by how shoppers searched. Content and links on top of that would have been wasted.
The plan therefore started with the category pages themselves, in a fixed order, and treated paid reallocation as the output of organic recovery rather than a decision to be made on enthusiasm.
Why this was difficult
Three things made the category work harder than it sounds. Collapsing facet URLs is easy to do badly: block everything and you lose the long-tail traffic that some combinations earn; block nothing and the dilution continues. Each facet had to be judged on search data. Template changes on a live store with an in-house developer had to be sequenced around releases and tested so that nothing on the grid or checkout broke. And the restructure of the range had to respect how the business actually merchandised, or it would have been undone within a quarter.
What we did, month by month
- Month 1CAC model and category auditThe acquisition cost model was built from ad platforms, invoices and the order system. Every category and facet URL was crawled and matched against search data to decide which deserved to exist. The developer began with the crawl rules.
- Month 2Facets collapsed, templates rebuiltFacet combinations without demand canonicalised or blocked; the few with demand promoted to sub-categories. Category templates rebuilt for speed and for guidance content above and below the grid. Organic revenue share unchanged, as expected.
- Month 3Category pages recoverThe first categories returned to page one for their head terms and organic revenue share began to rise. The media buyer started a watch list of paid terms now covered organically.
- Months 4–5Products, guides and the first reallocationProduct templates fixed for structured data and unique descriptions on priority items; buying guides linked into categories. With organic coverage stable on the recovered categories, the first tranche of non-brand paid spend was moved; total orders held and blended CAC fell measurably for the first time.
- Month 6Sessions pass 50,000Organic sessions reached 56,000. Links earned to category pages from product-review and industry publications.
- Months 7–10Second reallocation and the figures on this pageA further tranche of paid search reallocated after two months of stable coverage; 78,000 organic sessions, organic 41% of revenue, blended CAC 40% below the month before kickoff with total new customers higher than at the start.
Monthly organic sessions as the category pages recovered
Monthly organic sessions from Google Analytics 4 at the points reported to the client. The rise from month three follows the category-page recovery; the acquisition cost reduction lags it, beginning in month five after the first paid reallocation.
What made the difference
- ✓Judging every facet on search dataFacets with demand kept as sub-categories, the rest collapsed; this single decision let the main categories rank in month three.
- ✓Templates that give a page something to sayGuidance content above and below the grid, fast templates and product structured data turned category pages from lists into landing pages.
- ✓A range structured the way shoppers searchCategories and sub-categories renamed and regrouped around search behaviour, within the business’s merchandising logic.
- ✓Categories first, products second, content lastRevenue in retail SEO sits on category pages; the program went where the revenue was and treated content as support.
- ✓Paid reallocation as an output, not a decisionSpend was moved only after organic coverage had been stable for two months, so CAC fell without total orders dipping.
How the result was verified
Three records sit behind this page. The acquisition cost model took total acquisition spend each month (paid media from the ad platforms, agency fees including our own, and tooling from the client’s invoices) and divided it by new customers from the order system, with repeat buyers removed by customer ID. Organic revenue share and sessions came from Google Analytics 4, reconciled against the order system monthly. The facet and category decisions are preserved in the crawl and search-data worksheets from month one, so each collapsed or promoted URL can be traced to the demand figure that decided it. Baseline is the calendar month before kickoff; the endpoint is month ten. The client’s finance lead reviewed the model and this page before publication.
Who this applies to
This case applies to online retailers in Australia whose category pages are diluted by faceted navigation, slow or empty, or organised by merchandising logic rather than search behaviour, which describes most stores on Shopify, BigCommerce, Magento and WooCommerce. It applies most directly where the store is buying non-brand category demand through paid search that organic could capture. It applies less well to single-product brands, where there is no category structure to fix, and to marketplace-only sellers without a store of their own.
What to expect if you start now
Starting today, expect the first two months to be category work: facet decisions, template rebuilds and range structure, with organic revenue share barely moving. Expect categories to recover from month three, the first paid reallocation around month five once coverage is stable, and the blended acquisition cost to fall from there, with the full effect taking eight to twelve months depending on catalogue size and competition. Expect to involve your developer early and your media buyer throughout. For scoping and current pricing, see managed SEO in Australia.
“Ren Hao SEO turned organic search into our biggest pipeline source. We finally have a channel that compounds.”
“The transparency is unlike any agency we’ve worked with. We always know what’s happening and why.”
Frequently asked questions
Why did the work concentrate on category pages rather than products or content?
What was wrong with the category pages at the start?
How do you fix faceted navigation without losing long-tail traffic?
How does this apply to a retailer in Australia on Shopify or BigCommerce?
When did the acquisition cost start to fall?
How much paid spend was moved?
Can you share the client’s name or screenshots?
More case studies
About this page
Written by the Ren Hao SEO team and reviewed by Ren Hao, Founder & Lead SEO Strategist (LinkedIn), who led the engagement. Ren Hao SEO is a data-driven international SEO agency headquartered in Johor Bahru, Malaysia, serving eight country markets (company LinkedIn). Figures come from the client’s Google Analytics, Search Console and CRM and were reviewed with the client before publication; the client’s name is withheld under NDA and screenshots are available on request under the same terms. Last substantive update: October 2026. Drafting is assisted by AI tools and every page is edited and fact-checked by our team before publication; see how we create our content. Nothing on this page is a forecast for any other store.