Ecommerce SEO Case Study

Ecommerce SEO case studies tend to celebrate traffic; this one tracks revenue, because the store in question (anonymised: a specialty retailer, ~800 SKUs, established but search-invisible) needed orders, not sessions. Nine months, the earn-broad-route-narrow doctrine as the strategy spine, and a finding worth the whole writeup: the money came from category pages, and almost everything that lifted category pages was done elsewhere on the site.

The intake picture

Classic specialty-retail SEO neglect: product pages on manufacturer boilerplate (the fifty-competitors-same-text problem), category pages that were bare product grids — no content, few inlinks, ranking pages 3–5 for their commercial terms; faceted navigation spawning parameter sprawl that ate the crawl; a blog of six orphaned posts; and a link profile of supplier listings plus little else. Revenue from organic search: under 10% of total — for a retailer whose category terms had five-figure monthly demand.

The programme, in the order run

1. Months 1–2, technical triage: facet parameters blocked/canonicalised per the standard stack, duplicate manufacturer text replaced on the 80 best-selling products (not all 800 — effort followed revenue), Product schema fixed so prices and stars actually rendered. Effect: crawl stats normalised, rich results appeared, rankings barely moved yet — the floor, cleared.
2. Months 2–4, category pages became real pages: buying-guide content added to the twelve money categories (above and below the grid — selection advice only a specialist could write), titles rewritten per the CTR findings, and internal routing rebuilt: breadcrumbs, guide-to-category links, bestseller cross-links. Effect: the twelve categories moved from average position ~24 to ~11 by month four — coverage plus routing, before a single new external link.

1. Months 3–7, the linkable layer: four assets built for the niche's actual linkers (two definitive how-to guides, a materials/compatibility reference, and the campaign's star — an annual "state of [niche] prices" data piece from the store's own sales data, per the retail-data playbook): pitched to hobby media, forums and the roundup circuit, plus product seeding to six genuinely relevant reviewers. Yield: ~35 referring domains, nearly all pointing at the assets — then routed: every asset links its relevant categories prominently. Effect on the standard lag: categories climbed into positions 3–6 across months 5–8.
2. Months 6–9, harvest and iterate: refreshes on early pages, second-tier categories given the same treatment, gift-guide season pitched (four placements, two dofollow), and the price report's refresh cycle scheduled — the compounding machinery installed.

The scoreboard, revenue-first

Month 9 versus intake: organic sessions +240% — and the number that mattered, organic revenue +310%, to just under a third of store revenue, because the growth concentrated exactly where the strategy aimed it: eleven of twelve money categories on page one, product long-tail riding the category lift, the assets themselves converting modestly but linking mightily. The series pattern in retail form: nobody linked the money pages, and the money pages won anyway — earned equity, routed by architecture, cashed by categories. The store's ongoing programme is now three habits: the annual data report, seasonal pitch cycles, and category maintenance — a few days a month defending a revenue channel that didn't exist a year prior.

Frequently asked questions

Why categories over product pages as the ranking targets?

Demand shape and durability, per the doctrine: commercial queries are mostly category-shaped ("best [type]", "[type] for [use]"), and categories persist while SKUs churn. Products caught the long tail automatically once their categories carried them — the top-down flow that makes per-SKU campaigns a misallocation.

Was rewriting only 80 of 800 product descriptions enough?

For the campaign's goals, yes — the 80 covered most product-page revenue, and the remaining long tail's fix was structural (canonicals, category strength) rather than editorial. Full-catalogue rewrites are a year-two nicety; the effort-follows-leverage rule applies to catalogues brutally.

What would you do differently?

Start the data report in month one — it out-earned the other three assets combined (retail data is the niche's rarest commodity), and its links arrive on the same lag as everything, so earlier is compounding. That asset-plus-placement motion is replicable for almost any store with sales history — the pitching muscle being rentable where in-house hours aren't (our placement side, here).