IBM WebSphere Commerce Search Overview - Optimizing SOLR Search

50 %
50 %
Information about IBM WebSphere Commerce Search Overview - Optimizing SOLR Search
Technology

Published on February 28, 2014

Author: royalcyber

Source: slideshare.net

Description

Our extensive experience in SOLR eCommerce sites can guide you with the tweaking required for fast pace improvements. Inquire more: http://bit.ly/1d0Gj2A

IBM WebSphere Commerce Search Relevancy Improvement Powered by Apache Solr and Apache Lucene

Agenda  The Importance of Site Search  Effective Site Search  Search Relevancy improvements in WebSphere Commerce Server  Review task list  Statistics job  Solr Search WCBD script updates  Next steps ( status calls, VPN access)

The Importance of Site Search  Consumers use Google or Yahoo or some web search every day – They are used to searching.  The most important page on your website is the site search page.  Google analytics shows 65% of visitors on whitecap.com are using search features.  Out of which average 37% exit the site from first search result page.  Fast, precise and successful search increase your customer trust.  Without effective search, in short, merchants loose business.  By analyzing the sample keywords Barry provided, analyzing the Google analytics statistics on whitecap.com site, we are expecting over a 10 to 15% growth in revenue with more satisfied customer, and wining more customers each quarter.

Effective Site Search  The user may have done something simple such as misspell the word or put in a synonym, and when the site returns a web page informing the user that there are zero results, the user leaves, and the opportunity to make a sale has departed with the user.  The flip side of zero results is far too many results with little to no relevancy for what the user seeks. The right result has to be in the top three or you risk losing that customer.  Tracking user behavior and history through the buying funnel can yield a great deal of insight of search phrases within each stage of the funnel.  Solr can be tuned to boost certain results within a profitability scoring model  You can dramatically improve the return on investment of site search improvements. Using various search improvement techniques we can help you win more happy customers.

Spelling correction and ‘Did you mean’ When the customer types in some wrong spelling of a search word, Solr corrects the word and display the results. If the number of results are less, then did you mean feature gets triggered. Current White Cap Website is not having these Feature in their search

Search term associations Synonyms : Business user can provide a set of words separated by comma's in the Synonyms field. You can define two or more words that have the same meaning in your catalog and a search on any individual term will show the shopper results for all the terms. Replacement Term: Replacement terms are similar but are unidirectional. With replacement terms you have the option to include the original term in the search along with the additional terms you define. Alternatively, the original term can be dropped from the search and only the replacing terms are used. Landing Page : Landing pages work differently than the other types of associations. With a landing page, no search query is ever run. If the shopper searches on a term you have defined a landing page for, they are automatically redirected to that specific page. For example, searches for a brand name can be redirected to a page just for that brand. Current White Cap Website is not having this Solr Feature in their search

Search Relevancy Improvements in WebSphere Commerce server  Include keywords in the name and short description if possible Examples, hard hats, water cooler, collated nails  Add long description in the solar index schema Examples, outdoor lights, roofing Nails, Ridgid hand tools  Add additional attributes (color, size, and capacity) in the solar search index as required. Examples, safety vest lime medium, large red heated jacket, large carabiner, concrete sealer 5 gallon  Perform code changes to enable ‘Did You Mean’ feature to display auto spelling correction, and related search results. Examples, squincher will changed to sqwincher, and search results for ‘sqwincher’ will be shown.  Create search rules to adjust relevancy ranking for specific cases. Examples, Search result for batteries, can display batteries results for certain Manufacturers on top of the result. ‘Energizer’ and ‘Everyday’  Synonyms, word replacements, and wildcard searches Examples, Search results for ‘Mining Boot’ can include results for ‘miner boot’ as well.

Tasks Schedule Tasks List Estimates Hours Get access to the dev/test Solr environment. 4 hours Go over existing Solr configuration and index files. Identify custom changes done in solar 8 hours Tweak configuration files to identify the best approach for improving the search relevancy for the ‘bad scenarios’. 8 hours Implement the code changes in dev environment and perform testing for all the good and bad scenarios provided 16 hours Review the search results with team and collect feedback 4 hours Perform further changes in configurations as per feedback. 16 hours Setup of download statics job 16 hours Documentation the configuration changes done and review with team. 6 hours Assumption : Estimates for testing support in test environment, and production deployment support is not included in the proposal estimates.

Search term associations Synonyms : Business user can provide a set of words separated by comma's in the Synonyms field. You can define two or more words that have the same meaning in your catalog and a search on any individual term will show the shopper results for all the terms. Replacement Term: Replacement terms are similar but are unidirectional. With replacement terms you have the option to include the original term in the search along with the additional terms you define. Alternatively, the original term can be dropped from the search and only the replacing terms are used. Landing Page : Landing pages work differently than the other types of associations. With a landing page, no search query is ever run. If the shopper searches on a term you have defined a landing page for, they are automatically redirected to that specific page. For example, searches for a brand name can be redirected to a page just for that brand. Current White Cap Website is not having this Solr Feature in their search

Spelling correction and ‘Did you mean’ When the customer types in some wrong spelling of a search word, Solr corrects the word and display the results. If the number of results are less, then did you mean feature gets triggered. Current White Cap Website is not having these Feature in their search

Multi Select Facets  Facets group similar items and attributes together in the left sidebar to target meaningful product results.  The groups and shared attributes are displayed with the number of products that are contained in each facet.  The left sidebar can contain links that filter the search results or catalog navigation by brands, prices, and features. Current White Cap Website have Faceted Navigation but Multi Select Facets are not implemented in their left hand navigation

Download Statistics job  The DownloadStatistics job connects to the production server using the quick publish database connection to keep the search statistics up-to-date on the staging or authoring server.  Schedule a DownloadStatistics job on the staging or authoring server to connect and automatically copy search statistics from the production server at a scheduled interval for example, every ten minutes.

Download Statistics job (Continue) The Top Search Hits report provides statistical data about keywords your customers used to search your site that returned matching search results. The following is a sample of the Top Search Hits report:

Download Statistics job (Continue) The Top Search Misses report provides statistical data about keywords your customers used to search your site that did not return any matching search results. In this case, customers either saw no search results at all, or the search engine used a suggested keyword to return results. The following is a sample of the Top Search Misses report:

WebSphere Commerce Build And Deployment (WCBD)  Setup WCBD script to automate following tasks,  Checkout WC.ear and Search folder code changes from SVN repository  Build and Deploy search components to target environment  Restart the search solar server  Execute the di-preprocess and di-build jobs for all master catalogs  Execute di-build for Inventory index.

Questions

WebSphere Commerce search software stack  WebSphere Commerce search uses the following high-level Solr stack: WebSphere Commerce Feature Pack version Apache Solr version 3.5.0

Add a comment

Related presentations

Related pages

WebSphere Commerce V7.0 Feature Pack 2 Search Solution ...

... Overview and Deployment. An IBM ... 2 and set up the WebSphere Commerce search environment, experience the design and management of Solr ...
Read more

WebSphere Commerce Search Solution Performance Best Practices

... see Propagating the WebSphere Commerce search index. Optimizing the ... /search/solr ... the WebSphere Commerce IBM ...
Read more

Improving search engine optimization for your WebSphere ...

This article provides instructions on how to setup a WebSphere Commerce site for search ... ibm.commerce .sample.customer ... Optimization overview for ...
Read more

Product overview: WebSphere Commerce search

WebSphere Commerce search documentation; Concept Details and audience For more information; Introduction and training: Learn about WebSphere Commerce ...
Read more

WebSphere Commerce - Case Study Debugging Common SOLR ...

Debugging Common SOLR Support Questions ... Eric Rybczynski (ericryb1@us.ibm.com) WebSphere Commerce L2 Support ... com.ibm.commerce.search.crawler.task ...
Read more

IBM Front cover WebSphere Commerce V7.0 Feature Pack 2 ...

Feature Pack 2 Search Solution Overview and ... deployment overview IBM WebSphere Commerce V7.0 Feature Pack ... WebSphere Commerce search WAR ( solr ...
Read more

WebSphere Commerce - Wikipedia, the free encyclopedia

IBM WebSphere Commerce AKA WCS ... Application server which is supported with WebSphere Commerce is IBM WebSphere ... integrated search and enhanced ...
Read more

Solr with IBM WebSphere - SolrWebSphere - Solr Wiki

Solr with IBM WebSphere. Solr doesn't run out of the box on IBM WebSphere Application Server. The minimum version for Running Solar is WAS 6.1.x ...
Read more

Apache Solr -

Solr powers the search and navigation features of many of the world's largest internet sites. ... IBM Websphere Commerce ... Apache Solr, ...
Read more

WebSphere Commerce Search: Troubleshooting Common Issues

WebSphere Commerce Search: Troubleshooting Common ... com.ibm.commerce.search ... Or add ‘&debugQuery=true’ parameter to Solr search WebSphere ...
Read more