3 Examples of Leveraging Crowd Logic to Power Site Search Pages

Search is immensely important for online stores to drive product discovery to serve online shoppers. Depending on the source, you will find differing numbers but the statistics that comes up frequently is “Around 5% of shoppers use the search, and these users convert up to 2x better than their non-searching counterparts”. Most online stores already have a sophisticated in-store search engine and there are plenty of specialized vendors providing this functionality. However, it’s never too late to add some crowd-logic into the mix.

I’m a millennial, born in 1989. I came into this world at a time when the first successful liver transplant was performed, and the Berlin wall finally came crashing down. The concept of a personal computer was indeed invented long before my birth, but I still remember 1995 as if it were yesterday when my mother came home carrying multiple huge boxes and those good old floppy disks filled with different programs and games.

Where did it all start?

The trip down memory lane here was impulsive and completely unnecessary, but I actually had a relevant point to build upon: When Google first launched their public search engine in 1997, they completely re-imagined how the population of planet earth used the internet and continue to still influence it heavily to this very day. I personally can’t imagine a world without Google search, because every time a random question pops up into my head, it literally takes me 30 seconds to learn something completely new. And today I do this mostly by using the extremely sophisticated voice search, further lowering the barrier of use.

Search in general has become something of a benchmark for websites, online stores and other services, mainly because it just makes sense. If you have a bunch of information, why not provide a simple tool that you can use to find exactly what you are looking for? If you look around for statistics of search in ecommerce, you will find a lot of different case studies. However, a quote that frequently comes up is “Around 5% of shoppers use the search, and these users convert up to 2x better than their non-searching counterparts”.

There are multiple specialized search vendors out there that provide functionality such as NLP (natural language processing), and even voice search or other kinds of advanced functionality that assist shoppers when searching for the most relevant products. At Nosto, we set out to augment the search experience, and help retailers solve situations in which the shopper doesn’t find what they are looking for. Let’s look at some examples.

Contextual Site Search vs. Behavioral Cues

In most situations, context matters and it makes sense for search engines to index as much information about the products as possible so they can match it with the relevant search term that the shopper uses. However, behavioral cues work wonders in many cases.

Example 1: The Site Search Typo Turnaround

Stockholm-based streetwear store Caliroots has a pretty decent search engine provided by their platform. If I use the in-store search with “Mike” as a keyword, I’m presented with the two most contextually relevant products: Skateboards done in collaboration with Kodak Mike Carroll and Kodak Mikemo. However on a streetwear store, “Mike” could also just be a common typo.

The behavioral “Search and Viewed” recommendation element below the actual search results reveals that users who searched for “Mike” ended up viewing “Nike” sneakers. Who would have thought?

Example 2: Redirecting to a Brand Page During Site Search

Many online stores leveraging Nosto actually opt-in to use the behavioral search recommendation only in situations where there are no search results, with the aim to rely fully on the search engine, but reverting to crowd-logic whenever the search engine provides a no results page.

Award winning UK retailer Designer Childswear does just this. Armani is heavily represented on the site and upon searching for “Armani” a shopper gets redirected to the brand page with the relevant faceting options to further narrow down the selection. But the search engine does not take into consideration “Armani” + keyword type of queries here.

Examples 3-5: The “Search and Viewed” Recommendation

“Armani boyswear” provides no results at all but on this template, the Nosto “Search and Viewed” recommendation type is activated and provides relevant recommendations for what previous shoppers interacted with after using this search term.

Another useful strategy is to add Browsing History, or Best Seller Recommendation Types to the search results page if no results were found. This removes dead-ends for the shopper and increases the likelihood that they continue onto other products in the store instead of bouncing to the competitor.

New Zealand based powerhouse NZ Muscle provides a solid experience when searching for “Creatine”, and exposes the “Searched and Viewed” Recommendation again on the no-results page.

When deliberately introducing a typo and searching for “Ceratine”, there are no contextual search results. In this example, NZ Muscle leverages Nosto to both expose the “Searched and Viewed” recommendation, but also provides the shopper with familiar items through Browsing History, further lowering the risk of leaving the site because the shopper didn’t find what they were looking for.

Watch a 4 minute demo
Let us demonstrate why over 2,600 leading retailers worldwide have chosen Nosto’s AI-based personalization platform.