Services consulting
Ecommerce marketing built around margin, not traffic
Ecommerce marketing means acquiring customers at a cost your margin can absorb, then raising what each customer is worth. For most online retailers, traffic is not the constraint — contribution margin and repeat purchase rate are.
Why is more traffic sometimes the worst decision?
Traffic volume is a vanity metric when the bucket leaks. If conversion rate is 1.5% and your contribution margin per order is thin, doubling traffic doubles the cost of a problem you have not solved. This is the classic leaky bucket, and it is why growth-stage retailers should usually fix conversion and retention before buying more visitors.
The diagnostic is straightforward: what is your contribution margin per order, and what percentage of first-time buyers return within ninety days? If both numbers are weak, more traffic makes things worse. If both are strong, you have something worth scaling.
Why are category pages the highest-value pages?
Category pages capture the commercial-intent queries that produce purchases — often with high commercial intent and frequently badly served by competitors. They are also the pages that internal linking, product pages and navigation all point at, so authority concentrates on them.
| Element | What to optimize | Common mistake |
|---|---|---|
| Title and H1 | Category plus key differentiating attribute | Boilerplate title identical across categories |
| Intro copy | Genuinely useful category-specific text above the fold | Auto-generated or absent, leaving the page thin |
| Filters and sorting | Facets that reflect how people actually shop | Filters that create duplicate URLs with no indexation control |
| Product cards | Price, availability, reviews and clear differentiation visible without clicking | Cards relying on hover state that does not work on mobile |
| Pagination | Real crawlable pagination with crawl budget control | Infinite scroll with no crawlable links |
| Structured data | Breadcrumb and product markup that reflect the page | Markup describing products that are not on the page |
How do you handle faceted navigation without crawl traps?
Faceted navigation is where most large ecommerce sites lose crawl efficiency, and the wrong handling creates a genuine indexation problem rather than merely wasted effort.
The correct approach depends on whether the filtered combinations have genuine demand behind them.
| Situation | Correct handling |
|---|---|
| Filter combinations with real search demand | Allow indexing, self-canonical, unique title and copy, links from the base category |
| Filter combinations with no demand | Allow crawling, noindex,follow — not robots.txt blocking |
| Unbounded parameters — session IDs, sort order, arbitrary combinations | Block in robots.txt, or disallow via pattern |
| Sorting options | Canonical to the unsorted version, or noindex; never both canonical and robots-blocked |
| Pagination pages | Self-canonical, crawlable, linked from the category; consider limiting depth |
Does product schema actually help?
Product structured data helps by making your products eligible for enhanced search results — price, availability, rating and review information displayed directly in results — which raises click-through on commercial queries. It also improves machine comprehension of your catalog for AI systems.
It is not a direct ranking boost, and it cannot compensate for a weak page. What it does is make an existing page more presentable in the results, which is worth having on every product page.
- Markup describes what is actually on the page. Mismatched markup is worse than none.
- Price and availability are accurate and current — stale pricing generates errors and loss of trust.
- Aggregate rating and review data come from real reviews visible on the page.
- GTIN, SKU and product URL are consistent with your feed.
- Validate before deploying, and monitor for errors afterwards rather than assuming it stays clean.
What do product feeds actually do for you?
Product feeds serve two distinct purposes, and conflating them causes confusion. The first is product listings across Google's surfaces — free listings, Shopping ads and, increasingly, AI-generated shopping answers. The second is broader catalog distribution to other retailers and marketplaces.
The practical requirements are the same: accurate pricing, availability, identifiers, images and shipping information, kept continuously up to date. Feed quality problems manifest as suppressed listings — products simply not appearing — which is easy to miss without monitoring the Merchant Center diagnostics.
Why is retention the cheapest growth available?
Because acquiring a new customer costs whatever it costs, while retaining an existing one costs a fraction of it. In most ecommerce businesses the second purchase is dramatically more profitable than the first — lower CAC, no shipping subsidy, higher average order value, and the customer already understands how your product works.
- Set up the flows that matter first: welcome, abandoned cart, post-purchase and win-back. These need no new technology.
- Fix authentication for sending (SPF, DKIM, DMARC) before scaling any volume — see email marketing.
- Segment by purchase behavior rather than just recency and spend. Vague and repeat categories behave completely differently.
- Incentivise the second purchase specifically. A first-order discount of 15% is weaker at driving a second order than a modest second-order incentive.
- Measure ninety-day repeat rate as a primary metric, not a secondary one. It is the closest proxy for whether the business is actually working.
For most online retailers, improving ninety-day repeat rate by a few percentage points produces more incremental revenue than a substantial increase in paid traffic — and it costs nothing to buy.
Related services
Frequently asked questions
Should we increase traffic or improve conversion rate first?
Improve conversion first, almost always. Traffic costs money; conversion improvement is usually the cheapest growth available, and increasing traffic before conversion is fixed increases the cost of the same problem. The exception is if your conversion rate is already competitive for your category and your margin comfortably supports paid acquisition — then traffic is the lever.
How do we handle thousands of filter combinations on our category pages?
Decide per combination whether there is genuine search demand. Combinations with demand should be indexable with a self-canonical, a unique title and real copy, linked from the base category. Combinations without demand should be crawlable but noindex,follow. Unbounded parameters like session IDs must be blocked. The critical rule is never to combine robots.txt blocking with a canonical tag on the same URL — the crawler cannot read a canonical it is not allowed to fetch.
Does product schema markup improve rankings?
Not directly. It makes products eligible for enhanced results showing price, availability and reviews, which raises click-through on commercial queries, and it improves machine comprehension of your catalog. That is worth having on every product page, but it will not compensate for a weak page. The markup must describe what is actually on the page — mismatched markup is worse than none.
What is a good ecommerce conversion rate?
It varies enormously by category, price point, traffic source and device, so any generic benchmark will mislead you. The more useful question is how you compare to your own previous performance and to the segment of your traffic that actually buys. Comparing your blended rate against a published average mixes cold and warm traffic and tells you nothing actionable. Indicative ranges by business model are in the benchmark guide, with the caveats that come with them.
How important is email for ecommerce?
For most online retailers it is the most profitable channel available, because it reaches existing customers at effectively no marginal cost and converts far better than any cold channel. It is also the most commonly under-invested. Four flows — welcome, abandoned cart, post-purchase and win-back — account for most of the available return, and none of them requires new technology.
Should we sell on marketplaces as well as our own site?
Marketplaces give you reach and trust that a new brand has not yet earned, and they constrain your customer relationship, your margin and your data. A sensible pattern is to use marketplaces for discovery and to push customers toward your own site where the economics allow it — subject to marketplace terms, which vary on this and are frequently restrictive. Decide deliberately rather than by default.
How do we improve average order value without hurting conversion rate?
Bundling, free shipping thresholds, and post-add-to-cart cross-sells tend to work better than discounting, because discounts condition customers to wait for offers. The constraint is that every addition to the basket is potential friction, so changes should be tested with an eye on conversion rate as well as AOV. AOV rising while conversion falls is not a win.
Are returns a marketing problem?
Sometimes. Returns driven by size or expectation mismatch are a product-information problem — the description, imagery or sizing guidance was insufficient, and better content reduces them. Returns driven by buyer remorse or impulse purchasing are addressed by qualifying the customer before purchase. Returns caused by defects or quality are not marketing's to fix, and it is worth knowing which category yours fall into before changing anything.
Need ecommerce marketing for your business?
Get a free, no-obligation review of where you stand today on ecommerce marketing — and the three changes that would move the number that matters. No pitch deck.
Book a free 30-minute audit