M
M
e
e
n
n
u
u
M
M
e
e
n
n
u
u

October 27, 2026

October 27, 2026

What Is LLM SEO? LLM SEO vs Traditional SEO for Ecommerce

Definition: LLM SEO is the practice of optimizing how large language models ingest, interpret, and reproduce information about your brand and products so that…

Definition: LLM SEO is the practice of optimizing how large language models ingest, interpret, and reproduce information about your brand and products so that…

Definition

Definition: LLM SEO is the practice of optimizing how large language models ingest, interpret, and reproduce information about your brand and products so that AI-generated answers cite, describe, and recommend your catalogue accurately and favorably.

How LLM SEO Works for Ecommerce Brands

LLM SEO starts from a simple reality: large language models build their understanding of your products from the data they’re trained and grounded on. If that data is incomplete, inconsistent, or buried, AI answers will be too.

For ecommerce, two mechanics matter most: pre-training and fine-tuning, and retrieval-augmented generation (RAG).

1. Training and “mislearned” facts

During pre-training and fine-tuning, models ingest huge corpora of web pages, product listings, reviews, and documentation. Whatever patterns they see repeatedly become “facts” in their internal representation.

If the web is full of outdated specs, conflicting measurements, or thin product pages, the model can effectively mislearn your catalogue: wrong carat weights, incorrect SKU compatibility, or muddled sizing rules.

Because this training happens at scale, you rarely get to correct those facts directly inside the model. Instead, LLM SEO focuses on publishing clear, structured, and citable evidence that future retrieval steps will favor.

2. How LLM retrieval works in ecommerce answers

Modern AI search experiences use RAG and related techniques like query fan-out to ground answers in live web data. Before generating a shopping recommendation, an AI system will:

  1. Expand the query into multiple sub-queries (e.g., “best 2ct elongated cushion cut engagement ring under $5,000, durable for daily wear”).

  2. Retrieve candidate pages from its search index that match those intents.

  3. Score and filter pages based on relevance, freshness, authority, and structured signals (schema, specs, reviews).

  4. Generate a synthesized answer, often with citations and pros/cons pulled from the retrieved pages.

A 2026 arXiv study found nearly 30% of domains cited in AI Overviews did not appear on the co-displayed first page of classic results. That means being “ranked #1” on a traditional SERP is neither necessary nor sufficient to be the source an AI cites.

LLM SEO is therefore about being the best retrieval target and citation source across these fan-out queries: clean product data, consistent specs, rich structured markup, and evidence that supports the decision criteria the model cares about.

Why LLM SEO Matters More Than Classic Rankings

For ecommerce CMOs and growth leaders, the stakes are clear: AI answer engines are quickly becoming the front door for product discovery.

Adobe reports generative-AI traffic to U.S. retail sites grew 1,300% year-over-year during the 2024 holiday season and 4,700% YoY in July 2025. In their survey, 38% of U.S. consumers had already used generative AI for online shopping and 52% planned to do so that year, mainly for research (53%), product recommendations (40%), deals (36%), and shopping lists (30%).

Instead of scrolling SERPs, these shoppers ask an assistant what to buy. The model then surfaces a handful of brands and SKUs, sometimes with direct links or shopping carousels.

If your catalogue is missing, misrepresented, or framed negatively in those answers, you lose consideration-stage share of voice—even if your traditional SEO metrics look healthy.

LLM SEO matters because it:

  • Aligns your product data with AI decision criteria like price, availability, durability, and trust signals.

  • Prevents costly misrepresentation, such as incorrect specs or outdated pricing.

  • Shifts your measurement focus from blue-link rank to presence and persuasion within AI-generated recommendations.

Gartner’s research shows 31% of consumers say AI summaries make them spend more time researching and 31% consider more product options because of AI overviews. LLM SEO is your way to ensure your brand is consistently one of those options.

LLM SEO vs Traditional SEO, GEO, and AEO

LLM SEO is often confused with existing search disciplines. For ecommerce leaders, understanding the differences helps allocate budget and talent.

LLM SEO vs Traditional SEO

Traditional SEO optimizes for web search rankings and clicks: title tags, content relevance, backlinks, and technical health, measured via SERP positions.

LLM SEO optimizes for how models understand and reproduce your information:

  • Focus: ingestion quality (structured data, specs, reviews), interpretation (how the model clusters and compares SKUs), and reproduction (how it describes and recommends you).

  • Surfaces: AI answers, summaries, and agentic shopping flows, not just SERP snippets.

  • Metrics: share of voice in answers, citation rates, sentiment and pros/cons framing, cross-model consistency.

Google’s own guidance says AI search experiences still rely on its core ranking systems, but RAG and query fan-out can surface sources that differ from classic first-page results. LLM SEO treats AI surfaces as a parallel discovery layer, not a side-effect of traditional SEO.

LLM SEO vs GEO (Generative Engine Optimization)

GEO usually refers to optimizing for generative search engines specifically—the systems that combine retrieval and generation in one flow.

LLM SEO overlaps with GEO but is broader:

  • GEO: tuning content and signals so generative search features select and cite your pages.

  • LLM SEO: ensuring any LLM-powered assistant ingests and reproduces your brand correctly, including agents, chatbots, and summarization tools.

In practice, LLM SEO includes GEO, but also covers catalogue hygiene, multi-model evaluation, and long-term evidence management across all the surfaces models learn from.

LLM SEO vs AEO (Answer Engine Optimization)

AEO focuses on optimizing for direct answers—being the concise fact or recommendation a system chooses.

LLM SEO includes AEO-style thinking but extends it:

  • AEO: aims to be the chosen answer to a query.

  • LLM SEO: manages the full lifecycle of brand understanding: initial ingestion, ongoing grounding via RAG, and answer-time framing.

Where AEO might optimize a single FAQ to win a direct answer, LLM SEO for ecommerce is more architectural: pricing tables, spec sheets, merchant-of-record data, region-specific availability, and third-party reviews all work together to shape how the model reasons about your SKUs.

How to Measure AI Search Rankings

Because AI answers don’t show ten blue links, “ranking” becomes a question of presence and position inside generated content.

When people ask how to measure AI search rankings, LLM SEO programs typically track:

  • Share of voice in AI answers: what percentage of relevant queries mention your brand vs competitors.

  • Brand mention rate: how often your brand appears at all for a given topic or product category.

  • Citation depth and placement: whether answers link to you, quote your pages directly, or list your SKUs in top positions within shopping-style outputs.

  • Sentiment and framing: how pros and cons are described, including price/performance tradeoffs and durability claims.

  • Multi-model consistency: whether different assistants agree on your key attributes (e.g., cut quality, warranty, materials).

LLM SEO measurement relies on polling multiple models across locations and languages, capturing generated answers, and analyzing patterns over time. This is why teams turn to purpose-built LLM trackers and AI visibility platforms instead of relying solely on classic rank trackers.

Tools & Trackers for LLM SEO

LLM trackers Era & peers

LLM SEO requires instrumentation across the AI layer. LLM trackers like those in Era’s platform and adjacent tools monitor how models mention and rank brands in conversational flows.

A robust setup typically includes:

  • Multi-model polling: automated queries to leading assistants across regions.

  • Answer logging and parsing: extracting mentions, citations, pros/cons, and SKU lists from answers.

  • Trend reporting: visualizing share of voice, sentiment shifts, and competitive displacement.

Tools to track brand mentions in AI assistants

Under the hood, tools to track brand mentions in AI assistants behave like continuous "mystery shoppers" for the AI layer.

They:

  • Query assistants with decision-stage shopping questions.

  • Detect whether and how your brand appears.

  • Flag gaps, misrepresentations, or emerging competitors that displace you.

These insights feed directly into LLM SEO work: where to enrich product data, which SKUs need better evidence, and which regions are underperforming.

Marketplace listing optimization tools for generative search

For marketplace-heavy brands, marketplace listing optimization tools for generative search ensure your listings on major platforms are:

  • Rich in structured attributes (size, materials, compatibility, certifications).

  • Consistent across sellers and regions.

  • Cross-linked with authoritative documentation and reviews.

Because AI systems increasingly ingest marketplace data directly, these tools become part of your LLM SEO stack—clean marketplace data improves how models compare and recommend your SKUs.

How to measure AI search rankings (tool view)

From a tooling perspective, how to measure AI search rankings translates into dashboards that show:

  • Per-query and per-category visibility scores in AI answers.

  • Top cited domains and brand names per query cluster.

  • SKU-level presence in shopping-style responses.

Many AI visibility platforms are adding these capabilities alongside traditional SEO metrics, turning LLM SEO into an operational discipline rather than a one-off experiment.

How to Monitor Brand Mentions in Chatbots and AI Assistants

To monitor brand mentions in chatbots and AI assistants effectively, LLM SEO programs blend three categories of tooling:

  1. LLM trackers: automated polling of common shopping and research queries to see when, where, and how your brand appears.

  2. AI visibility platforms: multi-model, multi-region analytics that aggregate mentions, citations, sentiment, and competitive context into one view.

  3. Agent-testing frameworks: scenario-based tests that simulate real consumer journeys (e.g., “help me choose an engagement ring under $5,000”), capturing the full conversation and recommended SKUs.

Together, these tools show not just whether you’re visible, but how you’re framed—which is the real outcome LLM SEO is trying to control.

Tools & Reviews: Choosing AI Visibility Platforms Trusted by Marketers

Because LLM SEO is a young category, marketers should rely on AI search visibility platform reviews and hands-on trials rather than hype.

Look for:

  • Coverage across major models and regions.

  • SKU-level tracking and ecommerce-specific features.

  • Transparent methodology for query sets and polling frequency.

Independent roundups of AI visibility platforms trusted by marketers—covering products like SE Ranking, Semrush, Otterly, Peec AI, Profound, Scrunch, Birdeye, and others—can help benchmark capabilities before you commit your measurement stack.

Example: Fixing Mislearned Product Facts via LLM Seeding

Consider a mid-market jewelry brand that specializes in vintage-inspired stones. They notice that when consumers ask an AI assistant for a “2ct elongated cushion cut” engagement ring, the model repeatedly misdescribes their hero SKU:

  • It lists the stone as a standard cushion instead of elongated.

  • It confuses their modern cushion with an old mine cut vs cushion comparison drawn from blog posts that mix the terms.

  • It cites outdated dimensions and an incorrect depth percentage.

This is a classic LLM SEO problem: the model has mislearned product facts from noisy, historical data.

To correct it, the brand runs a LLM seeding campaign:

  1. Detect: Use LLM trackers to identify queries where the SKU appears with wrong specs or fails to appear at all.

  2. Publish structured evidence: Update product detail pages with machine-readable attributes (cut type, dimensions, carat range), add detailed comparison guides clarifying old mine cut vs cushion vs elongated cushion, and enrich markup so specs are easy for retrieval systems to parse.

  3. Align third-party sources: Coordinate with marketplaces and major review sites to ensure listings use the same terminology and dimensions, reducing conflicting signals.

  4. Promote fresh documentation: Publish an authoritative guide on choosing a 2ct elongated cushion cut engagement ring, explicitly describing their hero SKU in the model’s own decision criteria (budget ranges, durability, setting styles).

  5. Re-query and monitor: After indexing and propagation, re-run the LLM tracker queries to see if answers now cite the updated pages and describe the SKU correctly.

This example shows LLM SEO in action, by replacing mislearned facts with structured, citable evidence that AI systems can retrieve and trust.

Over time, the assistant’s recommendations shift: the brand’s elongated cushion cut SKU appears consistently, described accurately, and positioned against competitors with the right pros (sparkle, vintage aesthetic) and tradeoffs (price band).

For ecommerce teams, this pattern—detect, seed, monitor—is the practical heartbeat of LLM SEO.

Where Era Fits in LLM SEO Work

LLM SEO becomes much more manageable when visibility, tracking, and optimization live in one place. Platforms like Era provide multi-model LLM trackers, AI visibility analytics, and GEO/AEO-focused automation that help ecommerce brands and agencies systematically measure and improve how AI systems see, cite, and recommend their products across generative search and agentic commerce channels.

Chart showing generative-AI retail traffic growth in 2024–2025 based on Adobe data.

Definition

Definition: LLM SEO is the practice of optimizing how large language models ingest, interpret, and reproduce information about your brand and products so that AI-generated answers cite, describe, and recommend your catalogue accurately and favorably.

How LLM SEO Works for Ecommerce Brands

LLM SEO starts from a simple reality: large language models build their understanding of your products from the data they’re trained and grounded on. If that data is incomplete, inconsistent, or buried, AI answers will be too.

For ecommerce, two mechanics matter most: pre-training and fine-tuning, and retrieval-augmented generation (RAG).

1. Training and “mislearned” facts

During pre-training and fine-tuning, models ingest huge corpora of web pages, product listings, reviews, and documentation. Whatever patterns they see repeatedly become “facts” in their internal representation.

If the web is full of outdated specs, conflicting measurements, or thin product pages, the model can effectively mislearn your catalogue: wrong carat weights, incorrect SKU compatibility, or muddled sizing rules.

Because this training happens at scale, you rarely get to correct those facts directly inside the model. Instead, LLM SEO focuses on publishing clear, structured, and citable evidence that future retrieval steps will favor.

2. How LLM retrieval works in ecommerce answers

Modern AI search experiences use RAG and related techniques like query fan-out to ground answers in live web data. Before generating a shopping recommendation, an AI system will:

  1. Expand the query into multiple sub-queries (e.g., “best 2ct elongated cushion cut engagement ring under $5,000, durable for daily wear”).

  2. Retrieve candidate pages from its search index that match those intents.

  3. Score and filter pages based on relevance, freshness, authority, and structured signals (schema, specs, reviews).

  4. Generate a synthesized answer, often with citations and pros/cons pulled from the retrieved pages.

A 2026 arXiv study found nearly 30% of domains cited in AI Overviews did not appear on the co-displayed first page of classic results. That means being “ranked #1” on a traditional SERP is neither necessary nor sufficient to be the source an AI cites.

LLM SEO is therefore about being the best retrieval target and citation source across these fan-out queries: clean product data, consistent specs, rich structured markup, and evidence that supports the decision criteria the model cares about.

Why LLM SEO Matters More Than Classic Rankings

For ecommerce CMOs and growth leaders, the stakes are clear: AI answer engines are quickly becoming the front door for product discovery.

Adobe reports generative-AI traffic to U.S. retail sites grew 1,300% year-over-year during the 2024 holiday season and 4,700% YoY in July 2025. In their survey, 38% of U.S. consumers had already used generative AI for online shopping and 52% planned to do so that year, mainly for research (53%), product recommendations (40%), deals (36%), and shopping lists (30%).

Instead of scrolling SERPs, these shoppers ask an assistant what to buy. The model then surfaces a handful of brands and SKUs, sometimes with direct links or shopping carousels.

If your catalogue is missing, misrepresented, or framed negatively in those answers, you lose consideration-stage share of voice—even if your traditional SEO metrics look healthy.

LLM SEO matters because it:

  • Aligns your product data with AI decision criteria like price, availability, durability, and trust signals.

  • Prevents costly misrepresentation, such as incorrect specs or outdated pricing.

  • Shifts your measurement focus from blue-link rank to presence and persuasion within AI-generated recommendations.

Gartner’s research shows 31% of consumers say AI summaries make them spend more time researching and 31% consider more product options because of AI overviews. LLM SEO is your way to ensure your brand is consistently one of those options.

LLM SEO vs Traditional SEO, GEO, and AEO

LLM SEO is often confused with existing search disciplines. For ecommerce leaders, understanding the differences helps allocate budget and talent.

LLM SEO vs Traditional SEO

Traditional SEO optimizes for web search rankings and clicks: title tags, content relevance, backlinks, and technical health, measured via SERP positions.

LLM SEO optimizes for how models understand and reproduce your information:

  • Focus: ingestion quality (structured data, specs, reviews), interpretation (how the model clusters and compares SKUs), and reproduction (how it describes and recommends you).

  • Surfaces: AI answers, summaries, and agentic shopping flows, not just SERP snippets.

  • Metrics: share of voice in answers, citation rates, sentiment and pros/cons framing, cross-model consistency.

Google’s own guidance says AI search experiences still rely on its core ranking systems, but RAG and query fan-out can surface sources that differ from classic first-page results. LLM SEO treats AI surfaces as a parallel discovery layer, not a side-effect of traditional SEO.

LLM SEO vs GEO (Generative Engine Optimization)

GEO usually refers to optimizing for generative search engines specifically—the systems that combine retrieval and generation in one flow.

LLM SEO overlaps with GEO but is broader:

  • GEO: tuning content and signals so generative search features select and cite your pages.

  • LLM SEO: ensuring any LLM-powered assistant ingests and reproduces your brand correctly, including agents, chatbots, and summarization tools.

In practice, LLM SEO includes GEO, but also covers catalogue hygiene, multi-model evaluation, and long-term evidence management across all the surfaces models learn from.

LLM SEO vs AEO (Answer Engine Optimization)

AEO focuses on optimizing for direct answers—being the concise fact or recommendation a system chooses.

LLM SEO includes AEO-style thinking but extends it:

  • AEO: aims to be the chosen answer to a query.

  • LLM SEO: manages the full lifecycle of brand understanding: initial ingestion, ongoing grounding via RAG, and answer-time framing.

Where AEO might optimize a single FAQ to win a direct answer, LLM SEO for ecommerce is more architectural: pricing tables, spec sheets, merchant-of-record data, region-specific availability, and third-party reviews all work together to shape how the model reasons about your SKUs.

How to Measure AI Search Rankings

Because AI answers don’t show ten blue links, “ranking” becomes a question of presence and position inside generated content.

When people ask how to measure AI search rankings, LLM SEO programs typically track:

  • Share of voice in AI answers: what percentage of relevant queries mention your brand vs competitors.

  • Brand mention rate: how often your brand appears at all for a given topic or product category.

  • Citation depth and placement: whether answers link to you, quote your pages directly, or list your SKUs in top positions within shopping-style outputs.

  • Sentiment and framing: how pros and cons are described, including price/performance tradeoffs and durability claims.

  • Multi-model consistency: whether different assistants agree on your key attributes (e.g., cut quality, warranty, materials).

LLM SEO measurement relies on polling multiple models across locations and languages, capturing generated answers, and analyzing patterns over time. This is why teams turn to purpose-built LLM trackers and AI visibility platforms instead of relying solely on classic rank trackers.

Tools & Trackers for LLM SEO

LLM trackers Era & peers

LLM SEO requires instrumentation across the AI layer. LLM trackers like those in Era’s platform and adjacent tools monitor how models mention and rank brands in conversational flows.

A robust setup typically includes:

  • Multi-model polling: automated queries to leading assistants across regions.

  • Answer logging and parsing: extracting mentions, citations, pros/cons, and SKU lists from answers.

  • Trend reporting: visualizing share of voice, sentiment shifts, and competitive displacement.

Tools to track brand mentions in AI assistants

Under the hood, tools to track brand mentions in AI assistants behave like continuous "mystery shoppers" for the AI layer.

They:

  • Query assistants with decision-stage shopping questions.

  • Detect whether and how your brand appears.

  • Flag gaps, misrepresentations, or emerging competitors that displace you.

These insights feed directly into LLM SEO work: where to enrich product data, which SKUs need better evidence, and which regions are underperforming.

Marketplace listing optimization tools for generative search

For marketplace-heavy brands, marketplace listing optimization tools for generative search ensure your listings on major platforms are:

  • Rich in structured attributes (size, materials, compatibility, certifications).

  • Consistent across sellers and regions.

  • Cross-linked with authoritative documentation and reviews.

Because AI systems increasingly ingest marketplace data directly, these tools become part of your LLM SEO stack—clean marketplace data improves how models compare and recommend your SKUs.

How to measure AI search rankings (tool view)

From a tooling perspective, how to measure AI search rankings translates into dashboards that show:

  • Per-query and per-category visibility scores in AI answers.

  • Top cited domains and brand names per query cluster.

  • SKU-level presence in shopping-style responses.

Many AI visibility platforms are adding these capabilities alongside traditional SEO metrics, turning LLM SEO into an operational discipline rather than a one-off experiment.

How to Monitor Brand Mentions in Chatbots and AI Assistants

To monitor brand mentions in chatbots and AI assistants effectively, LLM SEO programs blend three categories of tooling:

  1. LLM trackers: automated polling of common shopping and research queries to see when, where, and how your brand appears.

  2. AI visibility platforms: multi-model, multi-region analytics that aggregate mentions, citations, sentiment, and competitive context into one view.

  3. Agent-testing frameworks: scenario-based tests that simulate real consumer journeys (e.g., “help me choose an engagement ring under $5,000”), capturing the full conversation and recommended SKUs.

Together, these tools show not just whether you’re visible, but how you’re framed—which is the real outcome LLM SEO is trying to control.

Tools & Reviews: Choosing AI Visibility Platforms Trusted by Marketers

Because LLM SEO is a young category, marketers should rely on AI search visibility platform reviews and hands-on trials rather than hype.

Look for:

  • Coverage across major models and regions.

  • SKU-level tracking and ecommerce-specific features.

  • Transparent methodology for query sets and polling frequency.

Independent roundups of AI visibility platforms trusted by marketers—covering products like SE Ranking, Semrush, Otterly, Peec AI, Profound, Scrunch, Birdeye, and others—can help benchmark capabilities before you commit your measurement stack.

Example: Fixing Mislearned Product Facts via LLM Seeding

Consider a mid-market jewelry brand that specializes in vintage-inspired stones. They notice that when consumers ask an AI assistant for a “2ct elongated cushion cut” engagement ring, the model repeatedly misdescribes their hero SKU:

  • It lists the stone as a standard cushion instead of elongated.

  • It confuses their modern cushion with an old mine cut vs cushion comparison drawn from blog posts that mix the terms.

  • It cites outdated dimensions and an incorrect depth percentage.

This is a classic LLM SEO problem: the model has mislearned product facts from noisy, historical data.

To correct it, the brand runs a LLM seeding campaign:

  1. Detect: Use LLM trackers to identify queries where the SKU appears with wrong specs or fails to appear at all.

  2. Publish structured evidence: Update product detail pages with machine-readable attributes (cut type, dimensions, carat range), add detailed comparison guides clarifying old mine cut vs cushion vs elongated cushion, and enrich markup so specs are easy for retrieval systems to parse.

  3. Align third-party sources: Coordinate with marketplaces and major review sites to ensure listings use the same terminology and dimensions, reducing conflicting signals.

  4. Promote fresh documentation: Publish an authoritative guide on choosing a 2ct elongated cushion cut engagement ring, explicitly describing their hero SKU in the model’s own decision criteria (budget ranges, durability, setting styles).

  5. Re-query and monitor: After indexing and propagation, re-run the LLM tracker queries to see if answers now cite the updated pages and describe the SKU correctly.

This example shows LLM SEO in action, by replacing mislearned facts with structured, citable evidence that AI systems can retrieve and trust.

Over time, the assistant’s recommendations shift: the brand’s elongated cushion cut SKU appears consistently, described accurately, and positioned against competitors with the right pros (sparkle, vintage aesthetic) and tradeoffs (price band).

For ecommerce teams, this pattern—detect, seed, monitor—is the practical heartbeat of LLM SEO.

Where Era Fits in LLM SEO Work

LLM SEO becomes much more manageable when visibility, tracking, and optimization live in one place. Platforms like Era provide multi-model LLM trackers, AI visibility analytics, and GEO/AEO-focused automation that help ecommerce brands and agencies systematically measure and improve how AI systems see, cite, and recommend their products across generative search and agentic commerce channels.

Chart showing generative-AI retail traffic growth in 2024–2025 based on Adobe data.

YOUR FIRST STEP

My job is to make sure you leave the first call with a clear, actionable plan.

Valerie

Client Success Manager

YOUR FIRST STEP

My job is to make sure you leave the first call with a clear, actionable plan.

Valerie

Client Success Manager

YOUR FIRST STEP

My job is to make sure you leave the first call with a clear, actionable plan.

Valerie

Client Success Manager

08

Ready to start?

Get in touch

Whether you have questions or just want to explore options, we’re here.

By submitting, you agree to our Terms and Privacy Policy.

era®

B
B
a
a
c
c
k
k
 
 
t
t
o
o
 
 
t
t
o
o
p
p
Soft abstract gradient with white light transitioning into purple, blue, and orange hues

08

Ready to start?

Get in touch

Whether you have questions or just want to explore options, we’re here.

By submitting, you agree to our Terms and Privacy Policy.

era®

B
B
a
a
c
c
k
k
 
 
t
t
o
o
 
 
t
t
o
o
p
p
Soft abstract gradient with white light transitioning into purple, blue, and orange hues

08

Ready to start?

Get in touch

Whether you have questions or just want to explore options, we’re here.

By submitting, you agree to our Terms and Privacy Policy.

era®

B
B
a
a
c
c
k
k
 
 
t
t
o
o
 
 
t
t
o
o
p
p
Soft abstract gradient with white light transitioning into purple, blue, and orange hues