An intern once wrote to me about the shop he helped run, a small retailer selling handball kit. The site did not appear for "handball equipment". He wanted to know where to start: meta descriptions, product pages, a separate blog, or one of the SEO tools everyone recommends. That is the question almost every small retailer asks when the traffic curve flattens.
The pressure is real. In France, the market I know best, the annual review published by FEVAD, the French e-commerce federation, put online sales at 196.4 billion euros in 2025, up 7% year on year across goods and services, with the average basket falling to 62 euros. Those are French figures, not British ones, but the squeeze they describe travels: more buyers, smaller baskets. Organic search remains a traffic source whose cost does not rise with every click.
What follows is a sequence, not a checklist to run in parallel. Technical work first, because it decides whether anything else is visible. Then queries, pages, content, measurement.
Step 1: measure before you change anything
Optimising without a numbered starting point is the first mistake. Before touching the site, record twelve months of organic traffic (Google Analytics 4 or Matomo), the queries and pages that collect impressions and clicks, and your indexing coverage. The last two come from Google Search Console. If it is not installed, install it and let data accumulate; a fortnight tells you very little.
On the handball shop the picture was flat traffic, weak positions on the obvious product queries, and 474 indexed pages for a catalogue far smaller than that. The gap between those two numbers was the finding: an architecture problem, duplicated variants or filter URLs or both. If your indexed page count is several times your product count, start there. Our note on how Google indexing works explains how to read the coverage report.
Third-party tools such as Semrush or Ahrefs help you benchmark competitors, but they underestimate the long tail, which often accounts for most of a shop's visits. Where their estimates and Search Console disagree, your data wins.
Step 2: choose queries that are actually worth having
The intern was aiming at "handball equipment". The queries already bringing visits told a different story: the brand name first, then "handball trainers", "handball shirt", "handball ball", and only then his generic term. Your strategic keywords are frequently not the ones you imagine.
I start from the queries a site already receives, extend them with Google autocomplete, the questions customer service gets asked, and a keyword tool. Then I sort each phrase on three criteria: purchase intent, whether a page exists that could plausibly answer it, and the effort required. A broad query that only the home page could serve ranks below a narrower category query with a real page behind it. Our guide to scoping an SEO audit covers how to frame that exercise.
| Query | Intent | Target page | Priority |
|---|---|---|---|
| handball trainers | Purchase, comparison | "Trainers" category | High |
| junior handball trainers size 3 | Specific purchase | Sub-category or indexable filter | High |
| handball equipment | Generic, poorly qualified | Home page | Medium |
| what size handball trainers | Pre-purchase information | Buying guide inside the category | Medium |
Step 3: fix the technical foundations
Three faults came up on that shop, and they come up on most shops I audit. An SSL certificate installed but HTTPS never switched on: without 301 redirects, without updating internal links and the sitemap, the certificate does nothing. Google has called HTTPS a lightweight ranking signal since 2014, but the real stake is trust at the moment of payment. A home page with no H1. And 404 errors on deleted products still linked from the menu.
Then the problems specific to online retail: variants that generate near-duplicate pages, faceted filters that multiply URLs, pagination. Google's documentation on faceted navigation is blunt about the mechanism: crawlers work through an enormous number of filtered URLs without judging which are useful, and new pages wait.
<!-- On /trainers/?colour=black&size=42 -->
<link rel="canonical" href="https://example.com/trainers/">
<!-- Index a filter only if people search for it,
e.g. /trainers/junior/ with its own copy -->Speed belongs to the foundations too. Google publishes thresholds for the Core Web Vitals: LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1, measured on real visits rather than in the laboratory (Google Search Central documentation). On a shop, the usual levers are image compression, caching and third-party scripts. Our page on site speed goes through the trade-offs, and a technical SEO audit puts these fixes in order before you move on to content.
Step 4: optimise the pages that sell
The shop's home page title was the brand name, a pipe, the brand name again, then "online shop". No keyword, no proposition. The replacement put the priority query at the front, the reason to buy in the middle, the brand at the end — the reasoning every category and product page needs.
Category pages
Category pages are your best candidates for purchase queries such as "handball trainers". Give them an explicit H1, a short introduction saying what the range covers, a buying guide underneath the product list rather than above it, and indexable filters only where searches justify them. They are chronically under-written on small shops, and they sit closest to the money.
Product pages
A product page earns its position through what the manufacturer does not supply: written description, structured specifications, customer reviews, descriptive image attributes ("Kempa Attack Pro handball trainers, size 8" rather than "product1.jpg"). Add Product structured data: merchant listing markup lets price, availability, reviews, delivery costs and return conditions appear in results (Google's Product documentation). Mark up only what the page actually shows.
Internal linking
On the handball shop, the cross-sell block was generic and identical everywhere, so it was ignored. Replace it with contextual links — "often bought together", "complete your kit" — between trainers, technical socks and a kit bag. Add breadcrumbs, a best-sellers section on the home page, and keep the tree to three levels: home, category, sub-category.
Step 5: add content without building a separate blog
The intern was planning a separate blog on a subdomain, to attract handball enthusiasts. It is a common and expensive mistake: links earned by the blog barely help the shop, and you start from zero on authority. If a blog makes sense at all, it lives in a directory of the same domain. For a small team, the priority is the pages you already have.
- Category FAQs: the questions customer service actually receives ("which size should I choose?", "which ball for a beginner?"), answered in a few sentences on the page concerned.
- Embedded buying guides: 300 words on choosing a size, under the list of trainers, is worth more than an isolated article nobody links to.
- Customer reviews: an email after delivery, guided questions that draw out technical detail, public replies from the merchant. Content competitors cannot copy.
This matches what Google says about helpful content: show who wrote it and what direct experience they have of the product (Google Search Central). A retailer who plays the sport and tests the trainers has an advantage over a marketplace listing, but it has to be legible on the page: a named author, a sentence about how the product was tried. Our approach to SEO content starts from the pages that already have impressions.
Step 6: plan three months, then measure
The schedule below is the one I proposed for that shop: a few hundred references, one person part-time. The order matters more than the pace.
| Period | Actions | Indicator tracked |
|---|---|---|
| Month 1 | Complete HTTPS migration, 404 errors, H1 and titles on the main pages, canonicals on variants | Indexed pages and errors in Search Console |
| Month 2 | Copy and buying guides on the ten main categories, the 20% of product pages that carry revenue, contextual cross-selling | Impressions and average position per page |
| Month 3 | Speed, structured data, a review campaign, a Google Business Profile if there is a physical shop | Organic clicks, revenue attributed to the organic channel |
Typical delays: technical corrections show up in Search Console within a few weeks, positions move over one to three months, the effect on revenue takes six months or more. None of that is a guarantee — competition and catalogue size change the picture. Track clicks and impressions per page, click-through rate, organic revenue, and acquisition cost against your paid campaigns.
A final point of proportion. A small shop does not beat the largest retailers on generic queries. It wins on precise queries, on advice about use, and on the relationship with the customer — a narrower target, and a more defensible one.
Common questions
Where should I start with SEO for an online shop?
With the technical foundations: HTTPS properly switched on with redirects, 404 errors fixed, a distinct H1 and title on every important page. These determine the effect of the content work that follows, and you can verify them in Google Search Console without buying a tool.
How long before SEO results show on an ecommerce site?
Technical corrections appear in Search Console within a few weeks. Positions on targeted queries usually move over one to three months, and the effect on revenue is better measured at six months or more. The range depends on competition and on how many pages you rework, so treat firmer figures with suspicion.
Do I need to optimise every product page?
Not at the start. Begin with the 20% of pages that generate most of the revenue and the categories that contain them, then extend gradually. The rest should at minimum have a unique title, a described image and a correct canonical on their variants.
Is a blog necessary to rank an ecommerce site?
No. Enriching category and product pages with FAQs, buying guides and reviews usually achieves more than a separate blog, particularly one on a subdomain. A blog is still worth having if someone can maintain it and it sits in a directory of the same domain.
Which tools are enough for a small shop?
Google Search Console and an analytics tool such as Google Analytics 4 or Matomo cover measurement. PageSpeed Insights covers speed. A crawler and a paid keyword tool help with auditing, but neither is essential to get started.
How do I handle filters and variants without damaging rankings?
Make indexable only the filter combinations people genuinely search for, and give those their own copy. Point the rest at the unfiltered page with a canonical. For variants of one product, a single reference URL should be indexed. Google also recommends returning a 404 when a filter produces no results.