Keyword Research

If the foundational layer teaches you “how search engines work,” then the Keyword Research layer answers a more practical question: what users are actually searching for, and who you should be writing content for. It sits between “laying the foundation” and “putting up the building”—crawling and indexing make a page visible, but only the right keywords make a visible page something people genuinely need. In a nutshell: technology decides whether you can get indexed, and keywords decide whether being indexed is worth anything. Direction sets your traffic ceiling: pick the wrong race, and no matter how clean your code is, you’re just doing busywork for terms with zero search volume.

Keyword Types: Know What You’re Picking First

Think of keywords as the “granularity” of a query, laid out from broad to narrow:

Seed Keyword: the broadest core term, like SEO or Python. Search volume is huge but competition is brutal and intent is vague; new sites basically can’t win these, so they mainly serve as a “starting point for expansion.”
Long-tail Keyword: a more specific query of three or more words, like how to generate a sitemap for an astro site. Individual search volume is small, but together they make up the bulk of total search traffic—intent is clear, conversion is high, and they’re easier to rank for. This is the main battlefield for beginners.
LSI / Semantically Related Keywords (Latent Semantic Indexing): words that naturally co-occur with a topic. When you write about “keyword difficulty,” terms like “search volume,” “SERP,” and “competitors” naturally appear; they help search engines confirm your content is “actually about this topic” rather than just stuffing keywords.

🧑‍💻 Developer’s view: think of it like API design—a seed keyword is like GET /search, broad and generic; a long-tail keyword is like GET /search?q=...&filter=..., where the more specific the parameters, the more precise the hit.

Understanding Metrics: Use Data to Judge Whether a Keyword Is Worth It

Picking keywords isn’t about gut feeling; it’s about reading a combination of four numbers:

Search Volume: the rough number of monthly searches. High doesn’t mean good—you have to weigh it against the three items below.
Keyword Difficulty (KD): usually a 0–100 estimate reflecting how hard it is to rank on page one. New sites should prioritize terms with low KD and medium search volume.
Click-Through Rate (CTR): the share of actual clicks a ranking brings in. Note: many queries have their clicks eaten up by “featured snippets” or ads, so high search volume doesn’t necessarily mean high real traffic.
Commercial Value (Search Intent Value): how close the user behind this term is to “a sale/your goal.” what is SEO is educational traffic; SEO outsourcing pricing is business traffic.

⚠️ Caution: don’t chase search volume alone. One term with KD 70 and 50,000 searches is far less valuable than ten long-tail terms each with KD 15 and 500 searches.

Keyword Mapping: Group by Intent, Then Map to Pages

Once you’ve collected a pile of terms, the core move is “group → map,” not building one page per keyword.

Group by Search Intent: usually four types—Informational (wants to learn something), Navigational (looking for a specific site), Commercial (comparing options), and Transactional (ready to buy).
One intent group = one page: cluster terms that share the same intent and overlapping meaning into one “Topic Cluster,” matched to a single page, to avoid multiple pages competing for the same ranking (Keyword Cannibalization).
Map into your information architecture: translate the grouping results into your URL and directory structure, for example:

/learn/what-is-seo        → 信息型（种子+长尾）
/compare/ahrefs-vs-semrush → 商业调查型
/tools/keyword-planner     → 交易型/工具型

💡 Tip: build a table—three columns are enough: 关键词 | 意图 | 目标 URL. This table is the blueprint you’ll use later to write content and set your title / H1.

Tools: Where to Mine Keywords and How to Validate Them

Tool	Main Use	Free?
Google Search Console	See terms you’re already ranking for, plus real clicks/impressions	Free
Google Keyword Planner	Official search volume ranges, ad-oriented	Free (requires an ads account)
AnswerThePublic	Generate “question/preposition” long-tails around a seed keyword	Partially free
Ahrefs	KD, search volume, competitor reverse lookup; comprehensive data	Paid
Semrush	Keyword Magic Tool, intent labeling, competitor comparison	Paid

Starting with zero budget: first use Search Console to find “low-hanging” terms you already have, then expand long-tails with AnswerThePublic + Keyword Planner.
Leveling up with a budget: Ahrefs / Semrush let you check KD in bulk and reverse-look-up which terms competitors rank for, saving a ton of manual work.

🧑‍💻 Developer’s view: most tools have an API or can export CSV, so you can absolutely write a script to pull search volume and KD and automatically filter candidate terms by a threshold (e.g., KD < 20 && volume > 100).

📌 This Layer Is Under Construction

The full tutorial is in progress—it will walk through the keyword-selection process, intent-judgment rules, and hands-on steps for the mapping template, item by item. For now, use the quick checklist below to run through the most critical actions of this layer:

Use 1–2 seed keywords and a tool to expand into at least 30 long-tail keywords
Annotate each candidate term with search volume, KD, and search intent
Cut terms with KD that’s too high or intent that doesn’t match your goal
Cluster terms into topic clusters by search intent, with each cluster mapped to one page
Build a 关键词 | 意图 | 目标 URL mapping table as your blueprint
Open Search Console and dig out “low-hanging” terms that already get impressions but rank low