Crunching the Numbers – Making Good Use of Site Search Data

By Mark Brixton | 15 Feb 2013

Harnessing the power of site search data can guide conversion optimisation and enhance the user experience. Mark Brixton shares how to get the most from your search box statistics.

The insights offered up by maintaining and analysing site search reports can reveal on-site consumer behaviours, affording retailers the ability to offer more accurate search results and fine-tune merchandising.

User experience (UX) expert Jakob Nielsen has analysed consumer site search experiences and found that most consumers will abandon a website if a relevant result is not returned after an initial query. This research proves the importance of offering a comprehensive on-site search experience to ensure that visitors are offered relevant results for just about any query.

Site search data and broader website analysis can be used to gain a clearer picture of user behaviour so as to determine any existing trends. In some cases, this may mean incorporating the idiosyncrasies of users’ language into webstore content, building synonyms into search, identifying and resolving queries that yield poor results, and also adding more items into specific inventory lists.

Applying Linguistic Eccentricities 

It should go without saying that not everyone speaks the same language, but many people don’t pause to think that this is true even on a micro-scale. Even within the same nation, the lingua franca is often fragmented into regional dialects or can be peppered with slang words and peculiar turns of phrase. In Australia, certain eccentricities of language can be identified even between various demographics. For instance, consider how a teenage girl speaks, in comparison with the language used by a middle-aged businessman.

jewellery retailer diva

diva BFF results page

Jewellery and accessories retailer, diva, discovered that its customers were using a variation of words for particular products, resulting in poor search results. The retailer found customers were searching for ‘BFF’ (best friends forever) necklaces, whereas the product was being called ‘Best Friend Necklaces’ in the on-site copy. In discovering this disparity, diva was able to include relevant keywords in order to adjust its search results to cater for this particular trend.

Misspelling and Grammatical Errors

Unfortunately human error is a fact of life. Luckily, site search reports will reveal what common misspellings and typos are made when using a webstore’s on-site search function. This information should be employed to encompass these errors and return quality results regardless.

In the case of vacuum cleaners, Deals Direct realised some people search using the misspelling ‘vacums’ and quickly moved to rectify the situation by offering up relevant results. Even better; in some cases words are predictably misspelt, and this means savvy retailers can cater for these errors before they even begin to happen. Remember: every customer that is dissatisfied after receiving results from site search the first time potentially means a missed opportunity to convert.

Example of misspelt word in on-site search

Deals Direct serves relevant results for ‘vacums’

Low Click-through Results

Searches that return no results have a zero click-through rate and keeping track of this information is useful. The input terms must either relate to content that people are searching for that you don’t have or, as is more often the case, are examples of your visitors using different keywords to describe your content, products or services.

In the first case, this knowledge is extremely valuable in helping you understand demand levels for products you don’t currently carry. It may also indicate popular products to consider stocking, and you can also tweak your merchandising to suggest alternate products.

CELLbikes has utilised intelligent search for maximum benefit. When a site user searches for a bicycle brand not stocked by the retailer, such as Giant, a clever banner is returned in the search results reading: “Don’t make a Giant mistake”, positioned above alternative product suggestions.

Moving Beyond Reactive Analysis

Employing these techniques to enhance on-site search, UX and the way you merchandise is only a starting point for a complete analysis strategy. Not only should retailers incorporate as much relevant data into their analysis as possible, they need to begin utilising this information in an anticipatory capacity, rather than a reactionary one. However, in order to do so, specialist resources with a comprehensive background in statistics and data analysis should be engaged, with multichannel retail experts used to help guide these efforts and transform the results into practicable outcomes.

0 Comment

Leave a Reply

Your email address will not be published. Required fields are marked *