VectorCiteVectorCite
Walkthrough · ecommerce · 4 weeks

An ecommerce brand starts getting cited in product comparisons

DTC brand with rich product detail pages that were invisible to AI engines. Fixed schema + reviews + BLUF on PDPs.

Composite walkthrough — patterns aggregated from real implementations. No specific named brand. Numbers are typical, not promised.
Profile

DTC ecommerce brand, ~150 SKUs, strong Google Shopping presence but absent from ChatGPT / Perplexity product comparisons.

Before
58/100
After
84/100
+26 pts

Narrative

The diagnostic showed PDPs scoring 58 overall — strong on Trust + Structure (Shopify defaults), weak on Authority + E-E-A-T. The biggest single move was switching schema type from generic Article to Product + AggregateRating + Review. That alone shifted 4 signals from warn/fail to pass. The press-strip and review-in-HTML changes ate the remaining gap. Week 4 was content updates on 30 PDPs.

Signals moved

  • json-ld-relevance
    →

    Switched PDP schema from generic Article to Product + Offer + AggregateRating. Engines now identify the page as a buyable product.

  • bluf-answer
    →

    Top of every PDP: 1-sentence direct answer to 'why this product, in plain English' before the marketing copy.

  • case-study-evidence
    →

    Pulled real customer reviews into the PDP HTML (not just behind a JS widget). Marked up with Review schema.

  • press-mentions
    →

    Added a press-strip section ('Featured in...') to PDPs that had real press. Marked up logos with proper alt text.

  • freshness
    →

    Added 'last updated' dates to PDPs and bumped them on actual content changes. AI engines treat undated product pages as potentially stale.

  • definitions
    →

    Defined product-category terms in plain language at first use ('Merino wool — soft, naturally temperature-regulating fiber from...').

Takeaways

  1. 1.The single highest-leverage PDP fix is Product schema with AggregateRating + Review. Generic Article markup costs you nothing visible but kills citation eligibility for product queries.
  2. 2.AI engines read customer reviews when they're in the HTML. Reviews behind a JS widget or in an iframe might as well not exist for AEO purposes.
  3. 3.Dated product pages outscore undated ones for comparison queries. Add a visible 'last updated' line and bump it on real changes.

Run the audit on your site.

The full 47-signal audit tells you which of these moves applies to your URLs.

Run free audit

Other walkthroughs

A SaaS site goes from C to A in six weeks
A trade publication starts getting cited in AI Overviews
An Indian D2C brand goes from D to B in four weeks
An open-source dev tool gets cited in 6 of 10 LLM answers in 60 days