Feature preview

Feature Introduction

People using the in-site search convert higher, should the vocabulary be maintained separately?

People who use in-site search have a higher conversion rate, according to the e-commerce data for the day, the conversion rate of in-site search users is 4.63%, and that of non-search users is 2.77%, and about 43% of visitors go directly to the search box. The vocabulary list needs to be maintained separately: the in-site search vocabulary service handles and recalls, the SEO keyword library service attracts external traffic, and the two sets have different sources, actions, and responsible persons.

In-site search Keyword library

Feature Introduction

People who use site search convert higher, should the vocabulary be maintained separately? Yes, it should be. According to the e-commerce data verified on the day, the conversion rate of site search users is significantly higher than that of browsers, Hello Retail's e-commerce site search statistics write 'Site search users convert at 4.63% vs 2.77% for non-searchers (Econsultancy).' Additionally, '43% of shoppers head straight to the search bar.' This indicates that a significant proportion of visitors use the search box when entering the store.Search behavior should be an independent category of entry and should have an independent vocabulary to accommodate it, and it should not be mixed with external traffic SEO keywords into one table.

Why are search users more valuable

The search box is a form of self-expression: visitors actively input their needs, and their intent is clearer than the browsing path. The differences between the two types of visitors are reflected in three places:

  • With a clear purpose. Search users come with specific vocabulary, whether the page hits this word or not directly determines whether they will proceed to purchase or leave.
  • Feedback is measurable. Search terms and clicks, bounces on the results page are retained, and it can be directly calculated which words bring conversions and which words only return empty results.
  • The cost of failure is high. When there are no search results or irrelevant results, visitors usually will not change the path to browse but directly leave the site.

This also explains why 43% of visitors use the search box as soon as they enter the site, maintaining the vocabulary becomes a clear profit-making thing: these people's needs are already written in the search terms, and if they are not caught, they will be lost.

The division of labor between vocabulary and keyword library

There are two sets of things inside the site, with opposite service directions:

Project Data source Major Action Towards
In-site search vocabulary In-site access logs and search records Supplement synonyms, correct spelling errors, configure empty result fallback What are visitors looking for within the site
SEO keyword library External traffic attraction words and landing page performance Categorize, expand long-tail, arrange content What are visitors searching for from the outside

There is an intersection but not equivalent: external traffic attraction words are more informative, and in-site search words are more product-oriented and model-oriented; the matching methods of the same business word are different on both sides. AnQiCMS's keyword library is used to centrally manage SEO keywords, and in-site search is responsible for matching the words visitors input with the content, which are two different capabilities, and the vocabulary list should be maintained separately, and the intersection part should be marked separately.

Who will perform the maintenance action

Word table maintenance is not a one-time task, it is recommended to be fixed in three segments in a loop:

  1. Look at the empty result words and low click words every week. Empty results mean missing content or mismatched word forms, low clicks mean incorrect sorting of results.
  2. Every month, do synonym and misspelling normalization. Map colloquialisms, abbreviations, and model writing to the same concept to reduce the situation where 'it is unavailable but has stock'.
  3. Align with the keyword library once a quarter. Pick out high-frequency words within the site search that are not externally driven, as the basis for选题选题 and page addition; these words are often the highest conversion entry.

Maintaining responsibilities should be on the content side rather than just looking at people who only care about data: changing the vocabulary list means adding content, changing titles, adjusting categories, and these actions are all completed in the background.

Focus on different scenarios

In the applicable scenarios of AnQiCMS, the degree of dependence on the vocabulary list for marketing websites and cross-border e-commerce websites is different. Marketing websites attach more importance to long-tail word acquisition, and the in-site search vocabulary list should cover product aliases and common questions; cross-border websites need to consider multi-language, and the same concept should be mapped to the same result under different languages. The vocabulary list also needs to correspond to multi-language content, otherwise, the translated page will not be searchable due to different word forms.

Personal blogs and corporate websites also have differences: The site search of a blog is often used to find old articles by topic, with the vocabulary centered around tags and categories; The search on a corporate website usually points to documents and contact information, and the vocabulary should unify the naming conventions of these pages.

Common questions

Should the vocabulary and tags share a set?Not recommended. Tags are for categorizing content, and the vocabulary is for the expressions input by visitors. Mixing them would force tags to accommodate colloquial writing styles, which would actually reduce the manageability of the content.

How often should you review the search log?It is recommended to check for anomalies (empty results, explosive words) once a week, and to normalize the data once a month. It is often the case that the window for supplementing content has been missed during major promotions or new launches.

How should an empty results page be handled more securely?Provide alternative suggestions instead of a blank screen: display related categories, hot content, or similar words, and add the word to the to-be-handled list as a basis for subsequent content supplementation.

Related functions

Preview

Which one to manage first - the article's accompanying image is directly displayed on other sites

When the article's accompanying image is directly cited by external sites, traffic and server load are counted on the source site, so bandwidth takes precedence over watermark. Anti-hotlinking relies on source verification to reject invalid reference pages, watermark management deals with ownership rather than traffic, automatic image layout in titles and full-site replacement solve the publishing side identification. This article gives the division of labor and implementation methods of both.

Preview

Is there a difference in indexing between the articles collected and the articles written by oneself?

The difference between articles collected and self-written articles in inclusion comes from the page repetition rate and information increment, rather than the source itself. The crawling and indexing judgment is what new content this page provides, and if the collected text is highly similar to existing pages on the entire network, it is likely to be considered of low value. The available processing order is to go through content review and sensitive word filtering first, then decide the scope of publication, and finally enter the sitemap.

Preview

How to separate the content model for product cases and articles when they have different fields

The fields and processes for products, cases, and articles are different, and they should be separated into independent content models: products require parameters and specifications, cases require customers and achievements, and articles only need titles and text. After separation, the list page and detail page should take what is needed, and drafts, pending publication, and recycling stations should also follow the model, avoiding forcing three types of content into the same set of fields.