Raising 'conversations from search' by 75% with Google Search Appliance
WallaShops running a public website and the users are customers and potential customers (B2C). Their previous solution was database-driven search using SQL tools. This solution was poor in terms of performance and ability to provide relevant results. The auto-complete was implemented using 3rd party application and was static. Web site that has around 600 thousand subscribers registered and a half million unique visitors each month needed search engine that knows how to deal with several QPS and show stability and efficiency.
Prior to implementing Google Search Appliance, WallaShops dealt with high bounce rate and low conversion rate from search. In fact, less than 2% of the searches ended up with a purchase). Moreover, WallaShops marketing team was aware that upgrading the search engine and improving the search experience will most certainly increase their sales and customer experience.
The old search solution couldn't also deal with misspellings (especially in Hebrew), morphology (male / female, singular / plural) and relevancy. Over 30% of search queries returned 0 results while users were not aware to their mistypes. Needless to say, that poor search results means with significant loss of revenues.
Another example of an important search feature that can affect directly on sales and sales increase is that every item has many metadata and some of them are more important than others. It is very important for search administrator to be able to config the search engine which are the most important metadata fields and which are less. Most of the items contain numbers and most of the users search items by the number of the item and The problematic issue is that some of the times customers are looking for “6400” and expect to get “LG-6400” while the previous solution returns items that cost 6,400 NIS. WallaShops previously search engine does not had the configuration to search first in the most important metadata and then in the others as the GSA.
“Our reports show that 12% of the queries contain typos in Hebrew and the GSA has been able to correct them with the built-in spell-checker. In addition to the spell checker the new version of the GSA (7.2) expanded language capabilities and it upgraded the relevancies of the results.
We are experiencing a high increase in sales and we know for a fact that one of the reasons is the auto-complete functionality that exposes visitors, during the typing, to content that they might not have been aware of.
The new version and it functionality has been able to let our customers locate items while they are searching substring of the item’s model and significantly reduces the percentage of queries that returned no results and improved the relevancy. When we want to push items and manage promotions we use GSA’s KeyMatch. This significantly increases the exposure the item to customers.
The result biasing improve the relevancy of the results, the reason for that is strictly because we could configured the GSA to search first in the most important meta data (Model) and than in the other meta data.
Last but not least, the Dynamic Navigation help our customers a great deal only by letting them drill down in the results by clicking on the relevant categories and find the most relevant items from all the results”
WallaShops is the leading e-commerce website and one of the largest retail companies in Israel, displaying a variety of auctions, groups and individual competitive prices. The site offers its visitors a variety of over 10,000 products from various categories such as appliances and electronic products, computers, photography gear, furniture, sporting goods, household goods and more. The company provides a service network leads committed to fair trade, purchasing on-line in highly secure environment and maintaining customer privacy. WallaShops has around 600,000 subscribers registered and a half million unique visitors each month.