Keyword Classification
Core Keywords: high volume, high competition, 1-3 words. E.g., 益阳 website building, 益阳 app development. Hardest to rank but biggest traffic. Suit homepage placement.
Long-tail Keywords: lower individual volume, lower competition, 4+ words, more specific. E.g., how much does 益阳 corporate website cost, which 益阳 WeChat mini program dev company is best. Individual volume low but aggregate significant. Much higher conversion rate. Suit inner pages/articles/product details.
Related LSI Keywords: semantically related terms search engines use to judge topic coverage depth. E.g., for website building: domain, server, UI design, SEO, responsive.
Layout Positions by Importance
Title tag: MOST important position. Must contain core keyword. Under 30 chars. Format: Core Keyword - Brand | Supporting text. Description: does not affect ranking but affects CTR in search results. 80-120 chars with keyword plus compelling CTA hint. H1 tag: one per page, contains core keyword. H2/H3 tags: naturally incorporate related and long-tail keywords. First 200 words body: keyword appears 1-2 times naturally (do not stuff). URL path: /yiyang/wangzhanjianshe/ better than /page?id=123. Image alt attributes: describe image with keyword inclusion.
Keyword Density
Old approach: aim for 2-8%. Now: over-optimization (keyword stuffing) gets penalized. Best practice: natural writing, keyword in key positions (Title/H1/first 200) 1-2 times each, body 3-5 times depending on length. Should read fluently. Never sacrifice readability for density.
A client old homepage repeated 益阳 website building 28 times. Not only terrible reading but Baidu penalized for keyword stuffing. After reducing to natural density and reorganizing content, ranking returned from page 5 to page 1 within 3 months.
Long-tail Strategy
Long-tail is SME SEO opportunity. Core terms dominated by big players, but long-tail competition is low. Strategy: dedicated page (article/FAQ/case study) per long-tail term. Content answers user specific question. 100 long-tail pages x 10 clicks each = 1,000 total clicks, potentially exceeding single core term performance.