Ten years on one ecommerce platform.Long enough to teach it.

Commercebuild SEO from a small agency that has run this platform across multiple merchants since 2016.

We help Commercebuild merchants get the visibility, the leads and the conversions they were expecting. The platform runs the storefront and the integration. We work on what happens in front of it: whether buyers find you, and whether they buy.

Two jobs, and they are not the same job

Commercebuild provides the features

The storefront, the ERP integration, the infrastructure, the releases and the support. Functionality, built and maintained by a platform business.

Commercebuild is software as a service, the same model as BigCommerce or Shopify. You rent the infrastructure and the roadmap. What you do with them is yours, and that is the part nobody is renting you.

They are good at it, and it is not our job.

We help you know how and when to use them

Which settings matter for your catalog, what each one actually does, what the platform already handles on its own, and the order to work in.

The documentation tells you how. Ten years tells you when.

The platform has more capability built into it than most merchants use. Not because it is hidden, but because a feature list tells you what exists and not whether it is worth using on your catalog this quarter. That is a knowledge gap, not a software gap, and it is the gap we close.

The “when” is the part that is hard to buy. Anyone can read the documentation and learn how a setting works. Knowing that this particular setting is worth the effort on a four thousand product catalog, and that another one will change nothing until something upstream is fixed, comes from having watched it play out on other merchants for a decade.

How you can tell whether someone knows this platform

Seven things we know about your site before we look at it

These are platform behaviours, not faults, and every one of them has changed what an audit should say.

  1. Your structured data is microdata, not JSON-LD. A consultant who greps for one script type will report that you have no schema at all. You do. It is in the template, as itemtype and itemprop attributes, and the correct work is completing it rather than replacing it.
  2. Your meta descriptions truncate at 255 characters, mid-word. It is a column limit. Write to 150 and the problem disappears.
  3. Your prices and cart are injected after the page loads. Which means anyone testing your site with a command-line tool is reading a placeholder, and any claim they make about price or stock is untested.
  4. The same category can live at several paths, all returning 200, all claiming to be the original. We have seen four for one category.
  5. Your 404 page may be telling search engines to index it, with a canonical pointing at itself, which is how dead URLs stay in the index for years.
  6. Some of your headings are not where you think. Footer text marked up as a top-level heading will make a page look like it has none, or two.
  7. Pasted documents inside content blocks bring their own head, title and meta tags with them, so a single page ends up with several.

If you recognised three of those, that is the point. None of them are in the documentation as problems, because none of them are problems. They are just how the platform works, and knowing them is the difference between an audit you can act on and a generic report.

You are already paying for most of what you need

What the platform already gives you

Structured product markup in the template. Category and page-level control over titles and descriptions. Canonical handling. A daily sitemap. Redirect management. Custom URL control.

You are already paying for all of it.

What usually happens to it

The markup ships incomplete, missing the fields that make it eligible for anything. The meta fields get filled once at launch and never revisited. The sitemap includes pages nobody wants indexed.

Not because the platform is weak. Because nobody was shown how to use it.

The recurring pattern is capability that was never switched on. Markup present but incomplete. Fields completed at launch for a catalog that has since doubled. A sitemap doing exactly what it was configured to do three years ago. None of it is a platform failure, and all of it is costing visibility.

What we work on

Information architecture and taxonomy

How the catalog is divided, and whether those divisions match how buyers actually search. This is also where the same category ends up reachable at several paths, each one claiming to be the original.

Navigation

What the menu exposes, what it buries, and what that tells a crawler about which pages matter. On a large catalog the navigation is the single biggest statement of priority you make.

Category pages

Usually a heading and a product grid, which is the largest block of wasted surface area on most stores. They also carry the head terms, so they get rebuilt before product pages.

Product pages

Specifications, applications and the attributes buyers compare, collected from manufacturers rather than invented, and published in bulk with the same structure every time.

Marketing and content pages

The pages that answer questions before the product page has to. Also where pasted documents quietly bring their own head and title tags into a page that already had them.

Schema

Completing the microdata the template already emits rather than bolting on a second system. The missing pieces are usually offers, price, availability, brand and the identifiers that make a product eligible.

Data feeds

What you send to shopping and AI surfaces, and whether it agrees with the site. A feed that disagrees with the page is a suspension risk, and we have seen what that costs.

Measurement

Analytics and Search Console configured to answer a question rather than fill a dashboard, including the checks that tell you whether any of the above actually reached the index.

Some of this is the platform’s job and some of it is yours. Most of it sits in between: capability the platform provides, filled in by people who have other work to do, against a catalog that keeps changing. That middle is where we are useful, and it is the part nobody owns by default.

What ten years produced

A written operating procedure

Ten years of working out what this platform rewards, what it silently ignores, and what order to do things in. Not a generic SEO checklist with the platform name at the top.

Taught, not just done

The hardest part of platform knowledge is explaining it to someone who has to use it on Monday. Your team learns which fields matter, what each one does, and why the order matters.

Verified against the platform, not against theory

Every recommendation is checked against what the platform actually does, because half of standard SEO advice is either impossible here or already handled. Knowing which half saves the budget.

The value of long platform experience is mostly negative knowledge. Knowing what not to try, what the platform already does, and which piece of standard advice will waste a month. That is difficult to acquire quickly and it is the part we can hand over.

When you run more than one store

Several stores on one platform, which is where this gets interesting

Three of the groups we work with run more than one store on a single Commercebuild platform. Two stores, three stores, a B2B site alongside a B2C one, a second brand acquired and brought in. Seven instances in total across those groups.

That architecture is genuinely useful. It also creates a question nobody asks until it bites: what should be identical across your stores, and what must differ?

Pricing and terms usually differ. Product facts usually should not. Category structure is a judgement call. Get it wrong in one direction and you maintain everything twice. Get it wrong in the other and the same product is described two different ways, which teaches search engines and AI assistants that your company is two companies.

Deciding where that line sits, per group, is most of the work. It is also the decision nobody owns by default, because it sits between the platform, the catalog and the commercial teams.

The same question arrives on every SaaS platform, and the answer changes with how much is shared. On fully separate installations there is nothing common to inherit, which is a different and harder problem: we describe that case on the BigCommerce SEO page, and the assembled-stack version of the same problem on the WooCommerce SEO page.

Knowing the platform well enough to prove it is the platform

The defect that was not in the merchant’s site

A shipping-charge calculation fault in the platform got one merchant’s Google Merchant Center account suspended. The suspension lasted twenty-six months.

No amount of work on that merchant’s own site would have lifted it, because the fault was not on their site. We took it to the platform, worked with them until it was corrected at source, and the suspension was removed.

The fix protected every merchant on the platform, not just ours. That is the practical difference between a supplier who files a ticket and waits, and someone who knows the platform well enough to prove where the problem actually is.

Who this is for, and who it is not

It fits merchants on this platform who suspect they are not getting what they paid for, teams who would rather learn to run it than hand it over, and businesses evaluating the platform who want an accurate account of what running it properly involves.

We will say no when:

  • You want us to run it indefinitely. We would rather train your team and step back. If that is not the outcome you want, we are the wrong choice.
  • The real problem is the catalog, not the platform. Then the platform work is a distraction and we will point you at the actual problem.
  • You want a guarantee we will overrule the platform. Some constraints are real. We will tell you which, and we will not invoice you for fighting them.

Where this sits beside the rest

Platform knowledge is how the work gets done here, not what the work is. The catalog itself is ecommerce SEO, crawling and rendering questions are technical SEO, and whether assistants can quote your product pages is answer engine optimization.

Questions we get asked

Q. Do you work for Commercebuild or independently?

Independently. We are a small agency that has worked on this platform for about ten years across multiple merchants, and Commercebuild has introduced clients to us. That relationship means we know the platform and the people, and it does not mean we will tell you the platform is right for everything. Where a limitation is real we will say so, because you will find out anyway and we would rather you heard it from us.

Q. Does this overlap with what Commercebuild already provides?

No, and the division is clean. Commercebuild is software as a service, the same model as BigCommerce or Shopify, and it provides the functionality: the storefront, the ERP integration, the infrastructure and the support that keeps it running. We help you understand how and when to use it. Their documentation can tell you how a feature works. It cannot tell you whether that feature is worth your time on your catalog this quarter, or which of three options to do first, because that depends on your products, your buyers and what is already broken upstream. Every merchant needs both, and they are genuinely different jobs.

Q. What can you do that a general SEO agency cannot?

Know what is actually possible before recommending it. A large share of standard SEO advice is either already handled by this platform or impossible on it, and telling those apart takes experience rather than intelligence. We have watched competent consultants recommend adding JSON-LD to a site whose markup is microdata and already working, or propose files the platform does not serve. Every one of those recommendations costs a client money and produces nothing.

Q. What exactly is in scope?

Information architecture and taxonomy, navigation, category and product pages, marketing and content pages, schema, data feeds and the measurement that tells you whether any of it worked. Some engagements are all of it in sequence, most are two or three of them, and the diagnostic exists to work out which. The one thing we will not do is start at the end: rewriting product pages on a catalog whose taxonomy is wrong means doing the work twice.

Q. Our site already has SEO settings filled in. Is there anything left to do?

Usually quite a lot. The fields tend to get completed once at launch and then never revisited, so they describe a catalog that has since changed. The more common gap is that the markup already in your template is incomplete: present and valid but missing the fields that make it eligible for rich results, which is a filling-in job rather than a rebuild and is considerably cheaper than it sounds.

Q. We run more than one store on the platform. Does that change things?

It changes the most important decision on the account, which is what should be identical across your stores and what must differ. Pricing and terms usually differ. Product facts usually should not. Category structure is a judgement call and depends on whether the same buyer shops both. Get that line wrong in one direction and you maintain everything twice; get it wrong in the other and the same product is described two ways, which teaches search engines and AI assistants that you are two companies. Three of the groups we work with run several stores on one platform, from a B2B site alongside a B2C one to a second brand brought in after an acquisition, and in every case that line was the work.

Q. Can you train our team rather than doing it for us?

That is the preferred outcome and it is what ten years on one platform is actually for. The difference between knowing something and being able to teach it is most of the value here. Your team learns which fields matter, what each does, what the platform does automatically, and the order to work in. Then they run it, and we stay available for the decisions rather than the tasks.

Q. How do you know a problem is the platform and not our site?

By testing the right way, which on this platform is not obvious. Prices, stock and cart are rendered after the page loads, so a command-line check reads placeholders and produces confident, wrong conclusions. We also know which behaviours are normal for the platform and which are not, so we do not raise a template feature as a fault. On one account we traced a problem beyond the merchant entirely: a platform-level calculation fault that had caused a twenty-six month Merchant Center suspension, which we escalated and had corrected at source.

Q. Do we need to move to a different platform to rank?

Almost never, and we will usually argue against it. Replatforming is expensive, puts existing traffic at risk, and most problems we are called in for are configuration and content problems that follow you to the next platform. If your catalog is invisible here, the same catalog will usually be invisible there.

Q. What does an engagement look like?

One bounded piece of work first: a review of what the platform is doing, what it is not doing, and what is worth changing in order. It ends with something your team could act on without us. If it makes sense to continue afterwards it usually looks like catalog work or a training rhythm rather than a monthly retainer with a report attached.

Q. We are evaluating this platform. Can you give an objective view?

Yes, with the relationship stated plainly: Commercebuild refers clients to us, so we are not neutral in the way a paid advisor with no ties would be, and you should weigh that. What we can give you is accurate: what it genuinely does well, what it does not support, and what running it properly takes. We would rather you chose it knowing the constraints than chose it on a demonstration.

Start with a conversation

Thirty minutes on what your store is currently doing well, what it is not doing at all, and which of the two is worth money. If the honest answer is that your platform setup is fine and the problem is elsewhere, we will tell you and point you at it.