Search in the website of the Business-to-consumer station-write to you who don't know the technology

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

Before this:

1, want to write this article for a long time, but has been entangled in the search engine specific technical principle details, see the more, do not understand the more places, the delay to do, this is not a program ape born pain.

2, everyone knows that search is a complex thing, this article tries to think from a technical point of view and build a Web site for the website search system, does not involve too much technical details. As for the specific implementation costs are not considered here (is done through a simple sql+ cache, with Lucene or sphinx, such as Full-text search engine to do two times, and even find Google Baidu to buy code to do two times to develop, listen to your program ape, you do not master).

3, this article has repeatedly mentioned in-site search, rather than in-site search engine, there is a huge difference between the two (I am not sure whether the final design will be a true sense of the site search engine).

4, this article refers to a lot of information, examples cited below, for reference learning

"Web information architecture-Design a large web site" This classic book (not recommended for beginners to buy)

Beautiful Siccio several articles http://blog.xiqiao.info/2009/06/02/343

Several articles of yeeach.com

And some papers on Full-text search

After that, we enter the text

1. The following 2 questions need to be considered before starting to think about planning a search in a Business-to-consumer station

The problem and significance of searching in the station

The following 2 common search scenarios are described below

A user Xiao Li, the site A is more familiar with, to buy a computer, at this time know website A has computer sales. Direct input Keywords: Thinkpad x series for more accurate query.

A user small white, heard of the consumer website A, the first landing, see a dazzling array of merchandise. Have just visited similar websites before, or have a more comprehensive understanding of the current category of commodities. Want to quickly locate some of the products that are already in your mind. Then enter a broader keyword for fuzzy search: such as the importation of wool jackets, cotton t-shirts and other fuzzy keywords.

(1) Site search just meet the needs of these two types of users.

(2) Through the analysis of User keyword search frequency, to understand the potential needs of users. (For this, I have always had an idea, if found a large number of search keywords for a category of goods, and the site does not happen.) Website to reduce the risk, whether it can be used in a predetermined way, first on the shelves with the target keyword a matching product x?

(3) For the site operators, through the analysis of the user's keyword search log, can modify the naming of products, edit a more consistent with the user's recognition of the naming method (here by the way to mention the next free powerful statistics user site search tool Google Analytics)

Analyze your website for site search

In fact, site search is basically standard in the Web site, but still nagging about whether the website needs to search, or just search instead of engine?

(1) Consider the properties of the website products: the standard products, common products for the content of the site search usage will be high, and more than the door of goods, such as embroidery thorn, gifts such as the consumer Web site, users to search for the key words to understand not enough, most users can not build a clearer mental model, the search usage rate will be low.

(2) Consider the website may use the site search users of the absolute number, 100,000-level uv/Day website, site search users have been more significant, need to consider their needs.

(3) Consider the number of commodities, brand number, SKU number, according to a little experience, a single brand SKU less than 500 of the clothing site search usage rate is much less than 5%.

(4) Consider the customer repurchase rate, purchase cycle (in fact, the essence is to consider the composition of the new guest and the elderly), the new customer-oriented users, basic exploratory search, the search usage rate is also low.

Summary: Merchandise deviation door, flow is not high, SKU less, new guest-oriented site, station Search is the basic furnishings, even if you want to search in the station, you can simply cope.

Here to mention a point: many customers often need to use the Product section number to search for goods, that very need to search in the station, this is not a user needs, only to explain that the backend system is not done.

Through such a few criteria to judge, you will find some Web site reference Taobao search site eye-catching placed in the most prominent place in the website is how stupid!

2, consider these two issues, you decide to search in the station, then the first simple understanding of the work of search engine mechanism

The following is the working mechanism of the search engine

Crawler crawl Content-build (including additions and deletions) index-store index-query (user query)-– query analysis-query result sort-render result

(1) to mention that E-commerce information through manual entry or other methods have been imported into the system, do not need to use the crawler program.

(2) The so-called index, refers to the search system to the information processing, to transform information into a search system can quickly understand and facilitate the process of inquiry. The direction to explore is what to index and how to index the components of content.

(2) query analysis, this is the most technical content of the part, involving search engine core algorithm, for Chinese full-text search, here also involves the so-called word segmentation technology.

is more complex and will be briefly explained below.

(3) The query results are sorted, which is also a more complicated part, and has a variety of sorting methods. The most common and reasonable sort is "based on relevance of search keywords," What is the most relevant search result? For the same keyword, each person's understanding may be different, but through the process to understand and determine, the technical content can be seen. The most notable is the Google-pioneered page rank algorithm. The following is a brief description of the ranking algorithm for the Business-to-consumer.

(4) Search results presented, for people who do not understand the technology, this is the easiest to grasp, and Jiangshan part. This will be highlighted below.

To facilitate the distinction between what is in-site search, what is the site search engine, here is an analysis form (unknown source):

Here, the next article will explain the following questions

What do you want the user to search?

What builds do you want users to search for?

How to balance recall and precision

In order to let the user quickly locate the keyword of the query, what work to do?

What content components to render to the user?

How are the query results arranged?

Other auxiliary functions.

Interface design

        Original address: http://aboutb2c.com/255