Reversing a 7-year organic decline across multi-brand fashion

Arcadia Group ran a portfolio of high-street fashion brands on an ageing Oracle Endeca stack that had accumulated a decade of migrations, half-finished replatforms, and legacy URL structures. Organic search had been declining for seven consecutive years. Each brand had its own SEO debt, its own redirect chains, and its own long tail of faceted URLs cannibalising one another at enormous scale.

The business needed a reversal — not incremental improvement — and buy-in had to come from brand leads, platform engineering, and the commercial teams across every brand simultaneously.

A two-year, group-wide programme that treated the portfolio as one technical platform rather than N brands in parallel. The headline workstreams:

• Group-wide technical audit — crawl, log file, rendering, and indexation analysis per brand — mapping the decline to specific migrations, template changes, and Endeca configuration drift.
• Consolidated redirect chains, fixed canonical and pagination patterns, and re-architected faceted navigation so category pages stopped competing with themselves (see below — this was the single biggest piece of work).
• Rolled out structured data (Product, Offer, BreadcrumbList, Organisation) consistently across brands so the portfolio benefited from a shared schema baseline.
• Partnered with platform engineering to land rendering and performance fixes — Core Web Vitals, TTFB, CLS — systematically rather than brand-by-brand.
• Embedded SEO into brand, merchandising, and BI teams so it was reflected in seasonal campaigns, range management, and reporting from day one of each season.

Endeca is a powerful faceted discovery engine — and out of the box, it's an SEO minefield. Every facet combination (brand × colour × size × price × style × …) generates a unique URL, which at retail-catalogue scale produces tens of millions of crawlable permutations, most of them near-duplicates competing for the same intent, burning crawl budget, and diluting authority across the portfolio.

I led the faceted navigation rework end-to-end:

• Built a facet indexation policy per brand — a rules framework that classified every (facet, value) combination as indexable, canonicalised, or blocked. Only combinations with genuine search demand survived; everything else got pointed at its canonical parent.
• Replaced default Endeca URL patterns with clean, stable, SEO-friendly URLs — deterministic ordering, no session parameters, collapsed redundant variants, strict canonicalisation.
• Curated XML sitemaps that exposed exactly the facet combinations that should be indexed — segmented by value, freshness, and brand — so crawl budget went to pages that could actually earn rankings.
• Designed an internal linking layer that funnelled equity to high-value facet pages via category hubs, breadcrumbs, and related-product modules rather than leaving it to accidental sidebar orderings.
• Built dynamic on-page templates — H1, meta descriptions, breadcrumbs, and unique copy blocks — that scaled with the facet combinations, turning machine-generated pages into pages search engines treated as first-class.
• Commissioned custom Endeca rule modules and SEO controls that exposed indexation and canonical behaviour directly to brand and merchandising teams — settings that the default product simply didn't surface, and that previously required a developer ticket to change.

The hardest part of a multi-year programme isn't the technical work — it's making sure it doesn't quietly regress after handover. So a large share of the project was training every brand on the new faceted navigation model, the custom SEO modules built on top of Endeca, and the rules that kept indexation clean.

I ran structured programmes for merchandising, eCommerce management, copy, digital marketing, and platform teams across every brand in the portfolio: written playbooks, hands-on sessions with the new tooling, shared dashboards so every team could see the indexation health of their own brand in real time, and a single group-wide do / don't ruleset everyone could point at.

By the end of year one, decisions that would previously have required an SEO ticket were being made inside brand teams without breaking anything. That transfer of capability — not just the URL counts or the rankings — is what made the gains stick.

The faceted navigation rework was the single highest-leverage change — it stopped the portfolio from competing with itself and reshaped where Google spent its crawl budget. Once the technical foundation was clean, schema, performance, and on-page work stacked on top and traffic started compounding month on month for the first time in years.

The second load-bearing lever was the training programme. Treating every brand as a graduate of a shared curriculum meant the gains survived team changes, brand leadership changes, and seasonal pressure.

+600% organic session growth, reversing a seven-year decline across the group's brands. The faceted navigation architecture and training model became the template for subsequent replatforms within the organisation.