Internet Retailer Web Design 2008

Hello!

Thanks for attending my session at Internet Retailer Web Design 2008,
“Making Site Search Do More Than Just Return Results”.

As mentioned in the session, below is a checklist for improving your site search, and a few ideas for site search split and MVT tests.

– Larry Becker

PS What has your testing shown about the best way to make your site search results pages sell?
I’d welcome swapping ideas and sharing tips on what we see working. Give a call!

Site Search Checklist

The search box is clearly labeled and in a standard location.
The search engine never returns dead end result pages. Alternatives are provided.
Misspellings and plurals are consistently handled.

Results pages offer clear confirmation of user’s search:
- Page is immediately recognizable as search results.
- Headline echoes user’s search term.
- List begins above the fold.

Exact item / single result searches take user directly to relevant page.
Search results are accurate, relevant and useful. Net is neither too wide nor too narrow.
Order of results is logical and learnable.
Results can be viewed as either multiple or single pages.

Search refinements are relevant and useful:
- Results can be both sorted and filtered.
- Filtering options highlight differentiating product features.
- Most useful filters offered first.
Imagery conveys content, helps user choose.
How to choose content is offered.
Editorial / merchandising recommendations are offered.
Same results offered via browse or search.
Same refinements offered via browse or search.
Search results function as strong entry pages.
Catalog Quick Order is intuitively labeled and behaves as expected.

Site Search Testing Ideas

Test landing your PPC product category terms on search results pages for that term vs. category pages.
Test increasing and decreasing the space given to your “Why Shop Here?” message on search results pages.
Test different default sort orders for your site search results pages: best-selling vs. price ascending vs. price descending vs. alphabetical.
Test the presence, absence and size of thumbnail images on your site search results pages.
Test an array layout vs. a list layout.
Test varying the number of site search results returned per page: best 5 vs. best 10 vs. best 20 etc.
Test adding vs. increasing vs. decreasing vs. removing the amount of space dedicated to cross-selling on site search results. Even if your site search application lacks the ability to produce search-relevant cross-sell SKUs, test hard-coding static cross-sell blocks for your best selling product categories.
If you have a solid call center, test increasing the emphasis of your toll free ordering number on your site search results pages. (This takes careful tracking, as a successful test may decrease web conversions but increase total conversions by sending traffic into the call center.)