E-Commerce Search Relevancy, and Mobile
The growing use of small format devices for online purchasing puts increased emphasis on the importance of search relevancy ranking
- Search is the default strategy of most online buyers for locating products and services
- Yet many eCommerce websites don't invest in relevancy tuning
- On mobile devices, where screen real-estate is at a premium, search relevancy is even more important
- Relevancy tuning for eCommerce delivers a clear return on investment
DO YOU DO RELEVANCY?
Back in the 1990s, I wasn’t exactly fresh-faced, but I was relatively new to the world of search engines.
I have fond memories of a visit to a well-known Scandinavian newspaper – it was one of my first visits to a potential enterprise search customer. They still had a Head Librarian (so by inference, they had more than one librarian). She was a charming lady, and as is the case with many professionals from that part of the world, her English was probably better than mine.
She showed me how they used their current (at the time) search engine. It didn’t do relevancy ranking.
Instead, the methodology was approximately this:
- Type in a basic search query using Boolean syntax
- Ask the search engine to tell you how many results your query found. In other words, results were not delivered at this stage, only a number
- Iterate with the search string, adding terms, more ANDs and NOTs for example, and keep checking on the number
- Do this until the number of results to be returned is manageable, and then request delivery, which is provided after a short pause
- Inspect all results
Give Me Relevancy
The replacement system provided relevancy ranking, and since that time, a lot of R&D dollars have been spent on relevancy algorithms. Frequent use of Web search engines such as Google and Bing has trained us to expect a relevancy-ordered results set to be returned by a search application. Further, web search engines do such a good job of this, that we’ve become collectively intolerant of search applications that do not list the content we are looking for within the first or second page of results.
As data sets grow, so does the difficulty of ensuring that the best documents make it to the top of search results. Web engines do this well, but their circumstances are very different to those of private data sets. Search applications that support key business processes in compliance, customer relations, e-commerce, and research, to name just a few, typically require tuning to fully meet user expectations, in terms of relevancy. Much depends on the content landscape, and that is different in every organization.
The Growing Use of Search-driven Navigation
During the past ten or so years, many search applications, from Amazon and Autotrader, to corporate intranets and highly customized search systems, have used the paradigm of search navigators (aka facetted search, or dynamic navigation) to help chop the data set down to size. This is a similar process to that employed by the aforementioned Head Librarian two decades ago. But these days, a simple click on a search navigation link circumvents the need for knowledge of Boolean operators. I’d go so far as to say that most of the great enterprise search systems and e-commerce search implementations I’ve seen develop during the past decade, rely heavily on navigators.
This does not mean that relevancy can be ignored. But well thought-out navigation options are an excellent complement to relevancy, as they help users to home in on the content they seek. In other words, good navigation options “share the burden,” or lighten the load put onto relevancy.
In a desktop-sized browser, great relevancy, complimented by appropriate navigation options, and presented in a clean user interface, is the recipe for success.
Mobile eCommerce Search
The use of small format devices for eCommerce continues to grow, and online retailers take care to present their products and services more succinctly on the smaller format, to make best use of the available screen real-estate. For search results, this usually means that there is no room on the main results screen for navigation options, which are typically hidden behind a menu. This brings us back to basics, because in the absence of support from one-click navigation options, relevancy ranking becomes the key to eCommerce search success.
At the same time, more and more signal data about the activities of the shopper can be captured, processed, and used to support relevancy.
So what are the implications for online retailers?
Getting eCommerce Relevancy Right
The first thing to be clear about – and this can be seen in how the leading e-Commerce properties organize themselves – is that relevancy is not just about technology.
Having an appropriate technology in place is foundational. This should include (this is a partial list):
- The ability to listen to multiple inputs that can influence and personalize relevancy
- Having fine-grain control, to support both general relevancy (for new potential customer’s first-time visits), and to present customized results to known customers
- Enabling merchandizers to promote and cross sell for time-limited periods via search results
Beyond the technology, processes are required to fine-tune relevancy. This is a continual process.
How can Search Technologies Help?
Setting up efficient and effective processes to optimize search relevancy requires experience, expertise with search engines, and in many cases software tools for benchmarking relevancy, and measuring improvement. We provide all of these capabilities through our services engagements. A great place to start is with our Search Assessment service, which will help any online commerce company to assess where they are, and make a plan for improvement. This does not necessarily need further investment in technology. Many organizations have great technology already in place – they just require some expert help with implementation, and fine-tuning the ongoing processes.
But where technology is a limiting factor, open source can be a good option too. See our white paper on e-Commerce search with Solr.
The ROI is Clear
The great news is that improved search capabilities directly impact sales revenues, and do so in an easily measurable way. So for the vast majority of search relevancy tuning projects, the return on investment can be clearly demonstrated
Search Technologies works with a range of leading commercial and open source software vendors to implement and improve e-Commerce search.