Why It’s Time to Say Goodbye to Your Zero Results with Never Null

We’ve all experienced it before: we’re shopping online and trying to find a specific item— be it a new vacuum cleaner, a style of shirt, a color of eye shadow—and we pop from site to site because we have something in mind that we really want or need. We try the big online retailers, or maybe we check out some new D2C brand that was recommended to us by a friend or an ad on social media…

But we keep getting that dreaded response: zero results.

It doesn’t take an ecommerce expert to recognize that a zero results page is a poor customer experience. That said, it’s impossible for every retailer to stock every item a customer could be looking for. As consumers, we want diversity in our options and there’s pleasure in discovering new products. Although we rely on our tried and true brands, it’s also delightful to discover something new to fall in love with. So how do you balance those two things as a retailer?

Don’t Do This

In case you needed an example, here’s what not to do. I went online to search one of the top online beauty care retailers for one of the Instagram haircare brands that my friend sent me. It’s called Prose and when you visit their bespoke website they help you design your perfect hair care routine via a fun quiz. Let’s say I’m a loyal big-box beauty care shopper so I go to their website to look for this new, trendy hair care brand. Here’s what happened:Not only do they not have the brand I’m looking for (which isn’t a surprise considering you can only purchase from Prose directly), they don’t even send me in the right direction for other products. Two out of the four suggestions have nothing to do with hair, and for the two suggestions that are hair-related, it’s not even shampoo!

In this day and age, we should not be showing zero results as an option anymore. Full stop. We have the advanced technology to create a better experience for our customers by mapping a data infrastructure that aligns our product catalog to popular queries and even items we want to surface for specific campaigns and stock pushes.

Solve Null Results with Semantic Search

This is where semantic search comes in. We’ll let the Lucidworks data science experts explain exactly what this is from a technical perspective, but for the layman, it’s essentially a method of connecting similar products with one another in order to surface items in a product catalog that will have a high rate of conversion when attached to certain search queries. When a customer searches for an item that might be out of stock or not carried at all, what they’ll receive in return is an array of similar items that are closely connected to their original search. For example, if I search for “pumpernickel crackers” that don’t exist, I’m served a mix of other similar products that have high conversion.

Groceries and beauty products aren’t the only use cases for semantic search. Consider retail: maybe a fitness guru is looking for a pair of yoga pants that was discontinued because a newer version was created with better fabric. Instead of showing a zero results page when our yogi searches their beloved stretchy bottoms, they’ll see the new version of the pants with all the matching tops and accessories that go with it. And maybe they’ll even see some branded content that explains why the fitness brand has moved to more sustainable fabric options with better moisture wicking. A home run for our yogi and a much better shopping experience overall.

Small (and Big) Teams Love Semantic Search

As the ecommerce business owner, you might be asking yourself, wouldn’t this be a lot of work for my merchandisers? Sounds like asking a small team to do a lot of rule curation and manual back end work, right? Actually, no. That’s the beauty of semantic search and using machine learning in general – a lot of this information can be trained up front so your site “learns” as your consumers navigate around products. Activating this type of deep learning doesn’t require a large monetary investment or a massive data science team, either. Providing some high-level signals along with product catalogs and business goals is about all you really need to get started. It’s a win-win all around.

“At KÜHL, we get a lot done with fewer employees than competing enterprises in the outdoor clothing industry,” said Timothy Boyle, Director of Ecommerce, KÜHL.“Given how sophisticated the Lucidworks platform is, their support and the capabilities allow our small ecommerce team to deliver personalized, relevant, and dynamic shopping experiences. The platform collects shopper signals to influence result ranking, has pre-configured ML models, and creates the compelling experiences our stakeholders and customers demand from our brand, driving higher conversion and higher AOV.”

Lucidworks recently deployed semantic vector search for one of our big box retailer customers and decreased their zero search results by 90% over the Black Friday weekend. That one methodology alone increased their revenue by millions of dollars during one of the most pivotal times of the year. Need we say more?

We’re excited to launch Never Null—our own version of semantic vector search—here at Lucidworks. Contact us to learn more or feel free to dive deeper with this blog post.