Listing Optimizer¶

Capability slug: marketplace_listing_optimizer Plan Tier: Accelerate (Quick Score) · Command/Premium (Full Analysis + Competitive Benchmark) Feature flag: settings_features.listing_optimizer_v2_enabled

Overview¶

The Listing Optimizer turns a single AWS Marketplace listing into one 0–100 score plus a clear, per-field path to raise it. The score combines four components — the same four that determine whether buyers and their AI assistants find, trust, and surface your listing — and weights them according to your listing's category, because each category is discovered differently.

It scores listings that Listing Sync imports, and feeds the SEO signal behind the Discovery Dashboard's Hero score.

The four components¶

Component	What it measures	Source
Listing Quality	Rule-based copy quality — the grounding text agents read and cite (title, descriptions, highlights, categories)	`FieldSeoScoringService`
Backlink Authority	External trust / domain authority — is this an established, safe vendor	`BacklinkAuthorityService` (DataForSEO, Bedrock fallback)
AI Visibility	Prompt coverage + structured data + brand voice — how AI assistants surface and describe you	`PromptCoverageService`, `FaqStructuredDataService`, `BrandVoiceMatchService`
Rich Media	Demo video presence + agent legibility + visual assets — whether your media exists and whether an AI buyer can read it	`RichMediaScoreService`

The overall score is the weighted combination of these four, computed by UnifiedListingScoreService over whichever components are present — a component with nothing to measure renormalizes out rather than counting as zero.

The Rich Media component¶

Buyers watch your demo video; AI buying agents can't. Rich Media scores both halves of that reality:

Sub-check	What earns points
Presence	A demo video linked on the listing, 3+ screenshots, a logo
Agent legibility	The video URL resolves with title metadata (oEmbed), a transcript and chapters exist in your Video library, and `VideoObject` JSON-LD is published on your website — the parts an AI assistant can actually read and cite
Buyer-content coverage	Whether the video shows a live demo, pricing, and objection handling — unlocks once video transcription is live; shown honestly as "not scored" until then, and the component renormalizes over what's measurable

Each gap row carries its fix: add the video in the listing editor's Rich Media section, generate a transcript with Video Intelligence, or publish the transcript excerpt + VideoObject markup to your listing page.

Vellocity never makes pixels

The Rich Media component doesn't produce video — it scores the media you have, makes it agent-legible (transcript, chapters, structured data), and repurposes it into GTM assets. The free public Listing Score on vell.ai mirrors this component as a presence-only check: it scores what's visible on your public listing page and never guesses — the agent-legibility half unlocks in the dashboard.

Category-aware weight profiles¶

Each AWS Marketplace category has different discovery physics, so the four components are weighted per category (ListingScoreWeights::PROFILES). The weights are also a signal in themselves — a high AI-visibility weight in the agent category tells vendors how to win there. Each profile sums to 100.

Category	Listing Quality	Backlink Authority	AI Visibility	Rich Media
SaaS (baseline)	35	20	35	10
AI Agents & Tools	30	15	45	10
AMI	40	20	30	10
Container	40	20	30	10
Professional Services	40	25	25	10

Why they differ:

SaaS — mature, mixed human + agent discovery; the baseline.
AI Agents & Tools — agent-native, often consumed agent-to-agent; young domains mean thin backlinks, so AI visibility carries the most weight.
AMI / Container — infrastructure you run in your VPC; spec-driven copy plus security/vendor trust lead.
Professional Services — buying expertise; firm reputation dominates and humans procure, so AI lags.

Category resolution

The profile is resolved per listing from its catalog_type and category_tags. AI Agents & Tools is a sub-facet of SaaS (catalog_type reads SAAS for both), so the agent tag in category_tags is what distinguishes it. Data products and any unknown/unmapped category inherit the SaaS baseline.

Scoring modes¶

The optimizer runs in one of three modes (the mode pill):

Mode	Name	What it does	Tier
`quick` (default)	Quick Score	Rides the 7-day signal cache for a fast re-score	Included
`full`	Full Analysis	Force-refreshes every live signal past its cache (DataForSEO backlinks, Brave prompt coverage, Bedrock brand voice, FAQ schema) and re-scores	Premium
`bench`	Competitive Benchmark	Side-by-side scoring against named competitors	Premium

Quick Score is the everyday default — fast, because it reuses cached signals. Full Analysis is the deep pass that re-pulls everything fresh. Competitive Benchmark scores you against named rivals.

The Premium gate¶

Quick Score is available to anyone with the optimizer. Full Analysis and Competitive Benchmark sit behind a Premium tier:

Both carry a Premium badge in the mode UI.
Selecting either surfaces an upgrade CTA.
They do genuine extra work — a full live re-pull (Full Analysis) or competitor scoring (Benchmark) — which is why they're tier-gated.

Real work, hence Premium

Full Analysis re-pulls and re-caches every live signal rather than reusing the cache; Competitive Benchmark scores against named competitors. Credit billing for these is wired separately as billable entities are registered.

Score interpretation¶

Score	Meaning
80–100	Strong — minor improvements possible
60–79	Good — several optimization opportunities
Below 60	Significant room for improvement

The optimizer returns the overall score, the category profile in use, a per-component breakdown (each component shows its weight and its score, with zero-weight components collapsed), and per-field rows grouped into Listing core, Backlink authority, AI visibility, and Rich media — each with state and improvement guidance.

What this is not¶

A listing builder. The optimizer scores and improves an existing listing. To create a listing, use Vellocity's built-in AI listing generator (Dashboard → Marketplace Listings → Create), AWS, or a Build tool; imported listings arrive via Listing Sync.
A video producer. Rich Media scores the video and visuals you already have and makes them agent-legible — it never generates footage.
The standalone AI-visibility tracker. That's a separate capability — marketplace_listing_seo — covering ChatGPT/Claude/Perplexity mentions + backlinks. See Agent Discoverability Score.
A pre-publish proof check. For whether your website/reviews/AWS content back the listing before you publish, use the Listing Alignment Grader.

Plan availability¶

Capability	Starter	Accelerate	Command
Quick Score	—	✅	✅
Full Analysis (Premium)	—	—	✅
Competitive Benchmark (Premium)	—	—	✅