Keyword Research and Targeting in the AI Era
What Is Keyword Research and Targeting?
Keyword Research: Understanding How Users Search
Keyword research refers to analyzing data to identify the words or phrases users commonly use in search engines. These keywords are often not just single words but search phrases with clear intent, such as:
- “Best coffee machine for small apartments”
- “High value laptops in 2025”
- “Which baby stroller brand is good”
Through keyword research, we can understand what potential users truly care about, their preferences, and how they express themselves, allowing us to create content that better fits their search habits. Good keyword research makes content more aligned with real searches and improves its chances of being recommended by search engines.
Keyword Targeting: Helping Search Engines Better Understand Your Content
Keyword targeting involves naturally embedding researched keywords into key areas of a page to improve its visibility in search results. Keyword placement shouldn’t feel forced or repetitive — it should help both search engines and users clearly understand the page’s topic.
Keywords should typically appear in the following places:
- Page Title
- Paragraph openings
- URL
- Meta Description
- Image alt attributes
- Natural positions within the content
For example, when a user searches for “portable washing machine for renters,” and you offer a relevant product, then the page title, content, and description should clearly contain that phrase — rather than vaguely saying “washing machine recommendations.”
Precise keyword targeting not only helps search engines understand your content but also provides readers with the answers they’re looking for.
Types of Keywords
Keywords play different roles in SEO and SEM strategies. Based on their length, purpose, and user intent, keywords can be categorized in several ways:
| Classification | Keyword Type | Definition | Example (with ‘coffee machine’) |
|---|---|---|---|
| By Search Intent | Informational Keyword | Users want to acquire knowledge or information | coffee machine principles, how to use a coffee machine, how to clean a coffee machine |
| Navigational Keyword | Users want to access a specific brand/site/page | DeLonghi official coffee machine site, Nespresso capsule coffee machine on Xiaohongshu | |
| Transactional Keyword | Users have clear buying or comparison intent | coffee machine recommendations 2025, best home coffee machine, top value coffee machines | |
| By Length & Structure | Core Keyword | Typically a single word or short phrase, broad and competitive | coffee machine |
| Long-Tail Keyword | More specific, longer phrases with clear intent, often including usage context, need, or attribute | best small apartment coffee machine, portable coffee machine for rentals | |
| By Brand Attribute | Branded Keyword | Includes specific brand or model names | DeLonghi coffee machine review, Nespresso capsule coffee machine |
| Competitor Keyword | Includes competitor names or comparison terms | DeLonghi vs Philips coffee machine, Nespresso vs Nestlé coffee machine | |
| By Geographic Feature | Local Keyword | Includes geographic location, reflects local search intent | coffee machine repair shop in Shanghai, coffee machine store in Beijing |
| By Question Format | Question Keyword | Phrased as questions like “how”, “which”, “is it worth buying” | how to choose a coffee machine? is a coffee machine worth buying? |
Which Platforms Require Keyword Research and Targeting?
| Platform Type | Example Platforms | Purpose of Keyword Research (Summary) |
|---|---|---|
| Search Engines | Google, Bing, Baidu, 360 Search | - SEO optimization (organic ranking) - Ad keyword bidding & targeting (SEM) - Understand user search behavior and trends |
| E-commerce | Amazon, eBay, Etsy, Tmall, Taobao, JD.com | - Optimize product titles and keywords - Boost internal search visibility - Identify buyer intent and hot categories |
| Video Platforms | YouTube, TikTok, Bilibili | - Optimize video titles, tags, and descriptions - Improve search ranking and recommendation weight - Discover trending topics and expressions |
| Social Content Platforms | Pinterest, Reddit, Quora, Twitter, Xiaohongshu, Zhihu, Weibo | - Capture real user language - Discover question/pain-point keywords - Identify trending topics and content opportunities |
| Content Publishing Platforms | Medium, WordPress, WeChat Official Accounts, Baijiahao, Toutiao | - Plan content structure and title keywords - Improve indexing and organic traffic - Align content with search intent |
| Ad Platforms | Google Ads, Microsoft Ads, Baidu Ads, Tencent Ads, Ocean Engine | - Precisely control ad targeting - Increase CTR and conversions - Reduce invalid clicks and cost |
| App Stores | Apple App Store, Google Play, Huawei App Market, Xiaomi Store | - Optimize app name, subtitle, and keyword fields - Improve search rankings and install conversions - Implement ASO strategy |
Key Metrics for Keyword Research and Targeting
When conducting keyword research and targeting, choosing the right keywords depends not only on their relevance but also on key metrics to evaluate their value. Different platforms have different evaluation standards and methods:
General Keyword Research Metrics (Applicable to Most Platforms)
| Metric Name | Description |
|---|---|
| Search Volume | Average monthly searches; measures potential traffic scale |
| Keyword Difficulty | Level of competition, usually related to existing content and domain authority |
| Keyword Length/Structure | Long-tail vs. core keywords; affects precision and competition |
| Trend | Search volume trends over time; helps judge if the keyword is seasonal or trending |
| Search Intent Type | Informational, navigational, transactional, question-type, etc.; crucial for matching content/ads |
Platform-Specific Keyword Metrics
| Platform Type | Proprietary Metric Name | Description |
|---|---|---|
| Search Engines | CPC (Cost per Click) | Common in SEM platforms; measures paid competition level |
| SERP Features | Whether the search triggers rich results like FAQ, images, videos, etc. | |
| Video Platforms | Tag CTR, Watch Time-Linked Keywords | Used by YouTube/TikTok algorithms to evaluate content relevance and recommendation weight |
| Title/Thumbnail CTR | Measures the attractiveness of video titles and thumbnails in search/recommendation | |
| E-commerce | Category Hot Terms | Hot terms within browsing paths and category tags |
| Purchase Trigger Terms | Keywords that drive conversions, e.g., “discount”, “limited-time” | |
| Social Platforms | Hashtag Volume | Frequency and popularity of specific tags, e.g., #coffeelover |
| User Questions & Comments Language | Authentic user language from platforms like Reddit or Zhihu, ideal for question-style or conversational keywords | |
| App Stores | ASO Keyword Weight | Varying importance of name, subtitle, keyword fields for ranking |
| Install Conversion-Linked Keywords | Keywords that lead to higher install rates |
While search engines, e-commerce platforms, and app stores may appear to involve similar behavior—entering keywords → system returns results—they differ fundamentally in goals, ranking logic, and user intent. Hence, keyword metrics include overlaps and distinctions:
| Dimension | Search Engines (e.g. Google) | E-commerce Platforms (e.g. Amazon) | App Stores (e.g. App Store) |
|---|---|---|---|
| User Search Intent | Primarily to “acquire information” (also includes navigation, transaction, etc.) | Strong purchase intent (“find something to buy”) | Looking for specific functions or usage needs |
| Ranking Mechanism | Relevance + Authority + Behavior signals | Relevance + Sales data + Price + Reviews + Conversions | Relevance + Downloads + Ratings + Retention + Keyword match |
| Keyword Research Goal | Capture volume, difficulty, semantic relevance, structure | Target conversion-driving phrases, competitor terms, strong buyer intent phrases | Identify high-converting functional words, competitor terms, category terms |
| Keyword Lifecycle | Mix of evergreen and trending topics | Closely tied to inventory and promo cycles | Adjusts dynamically with updates, feature changes, user feedback |
| Metric Differences | Volume, difficulty, intent, trend | Volume, competition + Conversion terms, category hot terms | Volume + ASO keyword weight, install conversion rate, feedback keywords |
How to Research Keywords?
With the rise of generative AI, many content creators and SEO professionals are rethinking a classic question: how should we do keyword research?
Traditional methods are mature—choose seed keywords, analyze search volume, expand long-tail terms, control keyword density… Clear processes, well-defined tools.
But today, users search differently—more conversationally, semantically, and unpredictably.
Let’s compare traditional keyword research with AI-driven approaches to highlight key differences:
| Dimension | Traditional Keyword Research | AI-Driven Keyword Research |
|---|---|---|
| Starting Point | Based on user behavior data + seed keyword expansion | Based on natural language input + semantic modeling — AI can extract keyword networks and intent from a description |
| Expansion Mechanism | Rule-based suggestions or manual selection | Context-driven — generates question-style, conversational, scenario-based keywords |
| Analysis Dimensions | Focus on: search volume, competition, CPC, core/long-tail structure | Adds: intent recognition, context relevance, tone, emotion, conversational style |
| Keyword Structure | Flat structure: core + long-tail | Semantic categories: “functional terms”, “question terms”, “negative intent terms”, “pain point phrases”, “scenario-bound terms” |
| Generation Method | Based on user behavior/platform history, mostly static | Generated by LLMs — can propose high-potential, never-before-searched semantic combinations |
| Representative Tools | Ahrefs, SEMrush, Google Keyword Planner, AlsoAsked, AnswerThePublic, etc. | Above + Frase, SEO.ai, ChatGPT (as prompt helper), WriterZen, NeuronWriter, Perplexity, etc. |
| Tool Capability | Requires keyword input to trigger suggestions | Can generate keywords based on goals or user intent, supports clustering and summarization |
| Content Integration | Keyword → Article structure (frequency, headers, paragraph distribution) | User intent → Search journey → Content phrasing; emphasizes natural expression and semantic coverage |
| Role Shift | Led by SEO specialist/content planner | Content creator + AI operator jointly drive process |
| Coverage Efficiency | Long-tail keywords require manual effort or enumeration tools | AI can automatically generate rich semantic variations and user language versions |
| Content Relevance | Focus on technical matches (frequency, density) | Focus on semantic fluency and intent fulfillment; emphasizes naturalness, tone, and consistency |
How to Target Keywords?
Finding keywords is only the first step. What truly determines the effectiveness of your content is how those keywords are used on the page, and whether they accurately respond to user search intent. In the past, we relied on experience to control the placement and frequency of keywords. Today, AI is beginning to help us understand “what users actually want to see,” and naturally integrate keywords into content.
| Dimension | Traditional Keyword Targeting Method | AI-Driven Keyword Targeting Method |
|---|---|---|
| Core Targeting Logic | Based on the match between keywords and page content, optimizing placement and frequency | Based on the alignment between user search intent and content semantics, optimizing context and experience |
| Keyword Distribution | Manually controlling keywords in title, H-tags, first paragraph, paragraph leads, URL, meta desc | AI automatically determines which sections, tones, and question styles keywords should appear in, aiming for natural flow |
| Page Matching Method | Keywords → Page Content (“writing around keywords”) | User Intent → Content → Related Keyword Clusters (“problem-solving centered”) |
| Intent Adaptation | Intent behind keywords judged by experience, which may involve subjective bias | LLM models can identify intent layers (e.g., informational / comparative / transactional) and generate matching phrasing |
| User Search Path | Understanding needs in terms of keywords | Understanding needs by search journey (“learn → compare → choose → purchase”) and dynamically adjusting keyword layout |
| Content-Type Matching | One keyword per page; multiple landing pages to handle different keywords | Multi-intent pages; support for multiple keywords; AI can recommend modular structures (e.g., FAQ + product comparison + reviews) |
| Ad Targeting | Select keywords + bidding; match ad copy and landing pages | AI can generate the most relevant keyword sets and copy based on ad goals and user intent (e.g., Performance Max + AI-generated visuals) |
| Supported Content Forms | Mainly for text-based pages | Also supports image-text, short video, voice and other formats for multi-modal integration |
v1.5.0