Ask five people what an ecommerce SEO consultant does and you will get five answers, most of them vague. Keywords. Content. Backlinks. Optimisation. None of it tells a store owner what will actually change on their site or how they would know it worked.
The honest version is narrower and more mechanical than the pitch usually suggests. On a store with a real catalogue, most of the value comes from a small number of unglamorous fixes that let search engines see the right pages, and from getting the category structure to match the way people shop. Everything else is downstream of those two things.
The short answer
An ecommerce SEO consultant works out which of your pages should be earning search traffic, removes the technical reasons they are not, makes those pages worth ranking, and then proves with data whether it moved. On a catalogue of any size, the first of those steps is usually the one that has been skipped.
Crawl control, or making the catalogue legible
A brochure site has thirty pages. A store with two thousand products, filters on colour, size and price, plus sort orders and pagination, can generate hundreds of thousands of unique URLs without a single person noticing. Search engines have a finite appetite for crawling any one site. If most of that appetite is spent on filter combinations nobody searches for, your genuinely important pages get visited rarely and updated slowly.
Faceted navigation and parameter URLs
This is the most common thing found and the most valuable thing fixed. The consultant decides which filter combinations deserve a real indexable page, because people genuinely search for them, and which should be blocked, made non indexable or handled with a canonical back to the parent. Blue running shoes might deserve a page. Blue running shoes sorted by price ascending, page four, does not.
Getting this wrong in either direction hurts. Block too much and you lose pages that were quietly earning. Block too little and the crawler drowns.
Duplicate and thin product pages
Stores generate duplication almost by accident. The same product reachable through several category paths. Manufacturer descriptions copied word for word across every retailer selling that item. Size and colour variants split into separate pages with near identical text. Each of these has a standard fix, and the fix differs depending on whether the platform lets you control canonicals and URL structure directly.
Products that go out of stock or get discontinued
Every store has a policy for this whether it knows it or not, and the default policy is usually the worst one, which is to delete the page. A product page that has been building relevance for two years should not become a 404 because the last unit sold. The alternatives, keeping the page live with stock status and alternatives shown, or redirecting to the closest replacement or the parent category, are decisions someone has to make deliberately and then apply consistently.
Category architecture and keyword mapping
The second block of work is deciding which page owns which search. Most stores have this backwards, with category pages named after internal merchandising language rather than what customers type, and with three different pages all half targeting the same phrase.
The output should be a plain map: one primary phrase per page, the page that owns it, and the pages that support it. That map is what stops two of your own pages competing with each other, and it is what tells you when a new category page is genuinely needed rather than just another near duplicate.
This is also where the difference between a category page and a product page gets settled. Broad, comparative searches usually belong to a category page, because the searcher is still choosing. Specific model or part number searches belong to the product page. When those are mixed up, the store ranks with the wrong page and the conversion rate suffers even where the position looks fine.
Making the pages worth ranking
Once the right pages can be reached and each one owns a clear job, the content work has a point. On a category page that usually means a short piece of genuinely useful buying guidance, the sort of thing a good salesperson would say, placed where it helps rather than dumped in a grey block under the product grid.
On product pages it means original description text rather than the manufacturer feed, the specifications people actually filter on, and the questions the customer service inbox answers every week. Reviews, delivery terms and returns policy carry weight here too, because they are part of what makes the page trustworthy to a buyer, and a page that converts badly rarely holds a position for long.
The general case for this kind of investment, and what it tends to return for a smaller store, is covered in e-commerce SEO benefits for small businesses.
Internal linking and merchandising
Large catalogues almost always have pages sitting three or four clicks from anywhere useful. A consultant maps how authority moves through the site and then fixes the routes, usually with related product blocks that are chosen rather than random, category to category links that reflect how people actually shop, and breadcrumb trails that mean something.
This overlaps with merchandising, which is why it works best when the person doing it talks to whoever runs the store rather than working from a crawl in isolation. The pages you most want to push are a commercial decision, not a technical one.
Structured data, speed and the boring hygiene
Product schema with price and availability, review markup where reviews are real, breadcrumb markup, a sitemap that lists indexable pages and excludes the rest, and a robots file that is doing what someone thinks it is doing. Page speed on mobile category pages, which are usually the heaviest thing a store serves and the most likely to be opened on a phone on mobile data.
None of this is exciting and all of it is checkable, which is the point. It is the part of the job where a competent consultant should be able to show you a before and after rather than ask you to take it on faith.
Measurement, and what should be reported
The reporting is where good and bad diverge most visibly. A useful report ties search performance to pages and to revenue: which page templates gained or lost impressions and clicks, what changed on them, and what the store earned from non brand search rather than from people typing the brand name into Google.
Be wary of a report that leads with rankings for a handful of hand picked phrases, or with traffic totals that quietly include brand searches and paid visits. Those numbers can rise while the store makes less money. Ask for non brand organic revenue, segmented by page type, and ask what specifically was changed in the period being reported.
What the job is not
It is not writing a blog post a week and hoping. It is not buying links. It is not a monthly report that lists tasks completed with no reference to whether anything moved. And it is not a single audit document handed over with two hundred issues sorted by tool severity rather than by what would make a difference to this store this quarter.
Consultant, agency or in-house
A consultant works well where the store already has developers who can implement changes and needs direction rather than hands. Where there is nobody to do the work, direction alone tends to sit in a document for six months, and the store is better served by a team that can implement, which is the model an ecommerce SEO agency is built for. In-house makes sense once the catalogue and the trading volume are large enough that the work is continuous rather than project shaped.
None of the three is inherently better. The question is only whether the person deciding what to change and the person able to change it are both in place, because a store with one and not the other stalls regardless of who was hired.
How to tell the work is happening
You should be able to point at things. Specific URLs that changed. A crawl budget problem that was measured, fixed and re-measured. A keyword to page map you can read. A set of category pages that now have distinct titles and descriptions where before they had none. If three months in nobody can show you what changed on the site, that is your answer, and it does not require any technical knowledge to reach it.
