Add To Your Favorites  Digg This Page  Google Bookmarks  del.icio.us  StumbleUpon  Share On FaceBook  
 
Internal Site Search    9/2/2008

Internal Site Search

The performance of your internal site search should be as important to you as your external search engine optimization. I’m not saying you will spend as much time optimizing your internal search but you need to make sure it’s performing as well as possible or you will loose sales. Just about any Internet retailer with internal site search setup and site logging/analytics tools will tell you that the search (or search results) page is the most popular page once someone is already on the site. If it is not, the search is not very prominent or they have very few pages.
 

Benefits of Internal Site Search

  • People land on your website from various entry points. Having a prominent search area allows someone to quickly search for what they are looking for when it may not be an obvious navigation path from the page they are on.
  • Some people choose to use search instead of menus for site navigation. Having relevant search results will allow these types of customers to better navigate your website.
  • Internal search makes you money.  Relevant results will directly connect customers with products and services they are looking for.
  • Find new popular search terms you have not yet optimized for. This is often overlooked with internal search and can be a goldmine. Not only will this tell you what you should promote on your main pages but also what keywords and search phrases you should be use to further your SEO efforts.
 

How to Implement Internal Site search

Every page on your site should have a prominent search link or better yet a search box and button. This will insure that no matter what page someone lands on when first coming to your site they have the ability to quickly search for what they are looking for.
 
Your search needs to perform quickly and you don’t want searches to drag down the speed of the rest of your website. That means searching large varchar or text fields are out of the question if you have any quantity of records to search. What I have found to perform fast and provide relevant results is indexing the words in my product descriptions. 
 
I have a table called SearchTerms that has a [Term], [ProductID], and [Instances], (varchar(50), int, and int respectively) defined. I build a program that loops through each active product and gets a word count for each word in the product description. In addition to the product description I add a keyword for the product category and one for the brand/manufacturer. Looking at the results in this table I define a list of words that my indexing program would ignore. Words such as (the, a, and, it, this) are ignored. Some of my descriptions come from the vendors with HTML tags so I ignore those tags as well. When someone types in a search phrase convert what they type into an array using the space as the delimiter. With ASP.NET this is easy to do using the split command. For instance:
 
Dim mySearchArray as array = split(txtSearchBox.text,” “)
 
Not I can build my query and loop through my array to add all the search terms like this:
 
Dim X as Integer
Dim myQuery as String
 
myQuery = “Select ProductID, Sum(Instances) from SearchTerms where “
 
For X = 0 ot mySearchArray.length -1
 If x = 0 then
    myQuery = myQuery & “ Term = ‘“ & trim(mySearchArray(X) & “’”
 else
    myQuery = myQuery & “or Term = ‘“ & trim(mySearchArray(X) & “’”
 end if
Next X
 
Myquery = myquery & “ Group by ProductID Order by sum(Instances) desc”
 
This technique will return each product that uses any of the search terms the customer searched and get a sum of search term instances for each product. Ordering by the most instances will move to the top the products with the most density of the search terms the customer entered, thus presenting the most relevant results to the top. 
 
You may consider only displaying the top twenty or thirty results as your customers won’t typically read past the first ten or so results. Additionally I like to display links to specific brand and category pages that can be gathered from your search results. You should also index content from product group pages and informational pages to display with these results. Tagging them as Informational and Products will help your users better find what they are looking for.
 

Site Search Reports

  • Google Analytics can be used to track your internal site search, to do this you need to pass your search terms across the address line. That would change the above line of code for your array to something like this: Dim mySearchArray as array = split(Request.QueryString(“SearchTerms”),” “). Regardless of if you use Google or build your own reports you will want to evaluate the following data:
  • Popular search terms and phrases – This data can be used to determine what products and pages should be highlighted on your site as well as where you can improve with your external search engine optimization
  • Product or group of products commonly visited as a result of a specific search phrase or search term. – This will tell you what products or pages people think are the best results of a specific query, this again can be used to determine what should be highlighted on your site.
  • Search phrases/terms converted to sales and which products were sold. – If you have popular search that result in many clicks but few sales you may need to adjust pricing on your product or service. Alternatively search that convert well to sales further indicate your popular products that may be hard to find using your navigation or from external search engine optimization.
  • Search without click – This report will indicate what people are searching for that you are not providing and results they like. Based on they’re search terms you can further optimize your search results or try to find resources for the products your customers are looking for.
  • Search with less than ten results – This is a different report that will show the same results as the previous report
 

Your Comment
Your Signature

© Copyright 2008 All rights reserved. Edgar E. Kneel