Ecommerce SEO: the parts that actually move a store

The honest version, before the tactics

Search is the slowest channel you have and often the cheapest one you end up with. A store that publishes good pages today is not competing for traffic next week. It is competing after Google has crawled the site several times, decided what the site is about, and watched it stay put. Treat the first few months as setup rather than as a campaign with a result.

That does not make search optional. It makes the order of work matter, because most store owners spend their SEO effort on the wrong half of the site. They rewrite the homepage and start a blog, while the pages that can actually rank for buying queries sit exactly as the platform generated them.

What follows is the short list of parts that move a store, roughly in the order they pay off.

Ecommerce SEO is three jobs

Strip away the vocabulary and there are only three:

1. Be findable. Every page you want traffic on has to be crawlable, indexable, and linked from somewhere a crawler will reach.
2. Be obviously about one thing. A page targeting "merino base layer men's" should say that in its title, its heading, its URL and its first paragraph — and should not also be trying to rank for "winter clothing".
3. Be the better result. Once two pages both qualify, the one with more unique information, clearer pricing, faster loading and more links pointing at it wins.

Everything below is one of those three applied to a specific part of a store.

Product page titles and descriptions

The title tag is the single highest-leverage string on a product page, and the default most platforms generate for it is wrong in the same way: Product name – Store name. That puts your least-searched words first and spends characters on a brand nobody is searching for yet.

A title that works names the thing the way a buyer types it, then qualifies it:

• Weak: The Harlow – Oakmount Studio
• Better: Harlow linen apron with cross-back straps

The meta description does not affect ranking. It affects whether the result gets clicked, which is the only reason to write it. Give it the one fact that decides the click — the material, the shipping time, the price bracket, the sizing range — not a restatement of the title.

Two rules save most of the damage:

• One page per query intent. If you sell the same apron in five colors, you probably want one page with a color selector, not five near-identical pages competing with each other.
• Write the description for the page, not the product feed. The manufacturer's blurb is on four hundred other sites. See the line-by-line version in how to write product descriptions that rank.

A full element-by-element pass on a single page is in the product page SEO checklist.

Category pages are where the money usually is

Here is the part most stores get wrong. Product pages win narrow, low-volume, high-intent queries: a model name, a SKU, a very specific spec. Category and collection pages win the queries with real volume: "linen aprons", "cross-back apron", "aprons for potters".

Those are the searches people make before they know which product they want, and the page that answers them is a collection page with:

• A heading that matches the query, not "Shop all".
• Two or three paragraphs of genuinely useful text — how to choose between the options, what the differences mean. Placed under the products, not above them.
• Filters that do not generate thousands of crawlable URLs. Faceted navigation is the most common way a small store wastes its own crawl budget.

If you only have the energy to fix one group of pages, fix these. They are fewer, they rank for more, and almost nobody writes them.

Site structure: flat and boring

A crawler should reach any product from the homepage in three clicks or fewer. That is the whole rule.

The practical implication is that deep nesting hurts. /collections/aprons/linen/cross-back/harlow is a worse address than /products/harlow-linen-apron, and a category tree five levels deep buries the pages you most want indexed. Keep the tree shallow, keep URLs short and readable, and stop changing them. Every URL change is a redirect you have to maintain and a reset of whatever that page had earned.

Internal links are free and underused

External links are hard to get. Internal links cost nothing and you control every one of them.

Three placements that do most of the work:

• Category to product. Already handled by your grid.
• Product to related product. "Pairs with", "also available in", "the heavier version". Useful to buyers, and it spreads authority to pages nothing else links to.
• Content to money page. If you write a guide about caring for linen, link it to the linen category with the words someone would search, not with "click here".

The test for an orphan page: if you cannot reach a product from your own navigation without using the search box, neither can a crawler following links.

Page speed, proportionally

Speed matters and it is also the thing store owners over-invest in. A store loading in two seconds does not have a speed problem, and shaving it to 1.6 seconds will not be what moves the rankings.

What actually causes slow stores, in the order it usually bites:

1. Unresized images. A 4,000-pixel photo displayed at 600 pixels is the most common single cause.
2. Too many third-party scripts. Each chat widget, popup tool, review app and analytics tag is a request.
3. Fonts. Two families and four weights is not free.

Run one measurement, fix the top cause, measure again. Then stop and go write a category page.

Schema: a small, reliable win

Structured data does not improve your ranking. It changes what your result looks like, which changes the click-through rate. For a store, the types that matter are Product, Offer, BreadcrumbList and — with caveats — FAQPage.

mAutomate's storefronts emit Product and Offer markup on product pages, including price, currency and stock availability, plus OnlineStore markup for the store itself, a canonical tag and a per-store sitemap.xml and robots.txt. What they do not emit is review or rating markup, because a store cannot honestly mark up reviews it has not received. Ecommerce schema markup covers which type earns what, and why invented review markup is one of the few SEO shortcuts that can get a site penalized.

How you tell whether any of this worked

Install Google Search Console on day one, before you change anything, so you have a before. It is free, it is the only report that shows what Google actually thinks of your pages, and it shows two numbers in the order they move.

Impressions move first. A page that starts appearing for a query is being considered. Clicks follow later, and sales later still. If impressions are climbing and clicks are not, your title and description are the problem, not your ranking.

The queries report tells you what you accidentally rank for. This is the most useful screen in SEO and almost nobody opens it. Sort by impressions, look for queries where you are on page two, and ask whether the page that is ranking is the right page. Half the time the answer is no — a blog post is ranking for a buying query that a category page should own, and the fix is one internal link and a better heading.

Two things to ignore. Average position is an average across queries and pages and tells you very little on its own. And daily rank-tracker screenshots will make you change things weekly, which is the fastest way to never learn whether any single change did anything. Make a change, write down the date, look again in a month.

What to fix in what order

Notice that blogging — the thing most "ecommerce SEO" advice starts with — is last. It is the slowest-paying item on a list where the other six are unfinished.

What SEO will not do for you

It will not get your first customers. This is the part that costs people the most, because they wait for search to work instead of doing the things that work now. A new store's first orders almost always come from somewhere a person already is: an existing audience, a marketplace, paid ads, a local market, direct outreach. The ranked traffic arrives later and compounds. The order matters — see how to get traffic to your online store and getting your first ten customers.

It will not fix a product nobody wants. Ranking a page for a query with no commercial intent produces visitors who read and leave.

It will not survive neglect. Prices go stale, products go out of stock, pages 404 after a catalog cleanup. A store that stopped being maintained slides down over a few quarters without anything dramatic happening.

And no tool, including this one, can promise you a position. Anything generated — titles, descriptions, schema, structured copy — is a starting draft that has removed the blank page. The judgement about whether that draft describes your product accurately is still yours, which is why in mAutomate nothing an AI writes goes live until you approve it.

If you do one thing after reading this: open your three best-selling products and read their title tags as a stranger would. Most stores find the problem in under a minute.