Eighteen dimensions, five families, a quantified diagnosis.

The five families

Every cycle runs the whole store through the same grid: eighteen dimensions across five families, each instructed separately with its own data sources. A dimension is never sacrificed to hit a deadline — if it fails, it is replayed next cycle, not struck from the scope.

• Technical, 7 dimensions — crawlability and indexability, XML sitemaps and their real submission status, performance quantified (LCP, INP, CLS, TBT, FCP, TTFB) against 2026 thresholds, structured data, above-the-fold mobile rendering, drift against the previous cycle, and indexing crossed with field performance.
• Onsite, 5 dimensions — content quality and E-E-A-T, page/intent fit, e-commerce SEO judged on product pages and never on the homepage, local signals, social and OG images.
• Offsite, 3 dimensions — inbound link profile (volume, referring domains, concentration), Maps presence and NAP consistency, SERP and keyword metrics.
• Content, 2 dimensions — hub-and-spoke topical cluster architecture built on queries you already rank for, and FLOW framing (Find, Leverage, Optimize, Win) to place the store and decide the next move.
• Answer engines, 1 dimension — visibility in generative answers: accessibility to search assistant crawlers, presence of an llms.txt, passage-level citability.

One dimension, one specialist

The eighteen dimensions are not walked through by one person ticking a checklist. Each is instructed separately, with its own sources and its own deliverable, then consolidated. That's what avoids the classic e-commerce audit failure: concluding "no product markup" after looking only at the homepage.

• Each dimension declares the data it needs and does not conclude without it.
• A dimension that doesn't apply — local for a store with no physical presence — says so explicitly instead of inventing a finding.
• A dimension missing its data is flagged as such, never blamed on the site.
• Results are consolidated into a single diagnosis, ranked by impact on revenue.

An audit that ends in production

The report always separates two columns: what we apply ourselves, and what needs your call. The first column goes straight into fixes without coming back for line-by-line approval. The second reaches you with an expected impact, and you decide.

• Applied directly — tags, structured data, redirects, sitemaps, internal linking, image compression, template fixes.
• Referred to you — reworking a range, changing pricing policy, removing pages that still generate revenue, ad investment.

What we measure, and how we prove it

The flip side of a data-driven audit is that it makes us measurable. Every fix is logged with the quantified impact we expect from it. The next cycle compares that forecast against the real result. It's uncomfortable for us when we were wrong, and that's exactly why we do it: after three months, you know which of our calls actually produce revenue.

Data-driven: the data we connect

An audit based only on what's visible from the outside misses the point — it describes the site, not the business. We wire up your real sources during onboarding, and no priority gets set without a number behind it.

• Search Console — real queries, positions, click-through rates, index coverage.
• Analytics — journeys, landing pages, conversion by source.
• Field performance data — what your visitors actually experience, not a synthetic score.
• Store back office — real margin per product, so we prioritise on profit rather than on traffic volume.