On average, 43% of online retail shoppers search sites directly via the internal search bar. Yet only 11% of searchers find products that match their query.
In other words, a significant number of visitors know what they want on your ecommerce site and search for it. But few are satisfied, and even more bounce—especially if your “zero results” page isn’t optimized (more on this later).
While a poor search experience leading to “no results” pages can lead to millions in revenue loss, it’s not the only reason why your current internal site search is costing you. We cover other common hidden costs below.
It’s a tale as old as time that 68% of customers won’t return to sites with poor search experiences.
While fixing your internal site search to improve the shopping experience and retain traffic is top of mind, how can you do so without knowing where your current search solution is falling short?
Lack of data is your first fissure, with other shortcomings to follow.
Does your current solution lack insights into search performance? Worse yet, do you have a lot of search queries that return erroneous results?
SwimOutlet was in the same boat (pun intended).
They knew that around 15% of their customers used internal site search, but they had difficulties optimizing for it. Since their legacy search solution wasn’t built for ecommerce, it lacked the key insights they needed to take action.
Switching to an AI-driven search solution allowed the SwimOutlet merchandising team to harness valuable clickstream data that surfaced new product lines to carry as well as suggested other opportunities to optimize and drive revenue.
The result? A 3.33% conversion rate and 3.68% lift in site revenue per visitor (RPV).
The absence of insights regarding how your changes impact ecommerce KPIs makes optimizing a guessing game.
The most critical example is when trying to reduce your abandoned cart rate.
Most product discovery platforms count an add-to-cart event as a conversion, whether the customer abandons their cart or completes the purchase.
They’re also not dynamic enough to handle ever-revolving catalogs. They can’t follow buyer trends, nor can they return the most popular products with confidence in personalization. Your team eventually has to build their own manual workarounds—a clunky approach when handling thousands of SKUs.
Most search platforms don’t offer much in the way of personalization. Showing previously viewed and “popular” pages is as advanced as they come.
But this barely scratches the surface of what personalization looks like when using clickstream data.
With clickstream data, you get an insider’s view of the shopper’s entire journey on your site, including zero search results, add-to-carts, clicks, and dwell time. And every conversion-oriented action and on-site behavior gets recorded and stored, data from which can be leveraged to tailor the user experience.
Here’s an example of how an individual user’s behavior could get recorded on a grocer’s site:
In this example, a person is shopping for steak and then searches for wine. The wine that is shown first is red, as red wine pairs well with steak. Thus, the search results using clickstream data shortens the shopping time and increases the chance of a purchase for the wine.
Relevance is a buzzword in the search world—rooted in keyword matching with product descriptions. But search relevance isn’t enough anymore, at least not for ecommerce sites.
The name of the game today is “attractiveness.”
Attractiveness uncovers results for search queries based on what’s desirable to each shopper at that moment in time.
A search for a “laptop” (although extremely vague) has context when you pair it with historical clickstream data. With attractiveness, one searcher would see various laptop brands, while another who’s historically searched Apple products could see only MacBooks.
With product catalogs constantly changing, and buyer behaviors and motivations always shifting, your product discovery platform shouldn’t rely on term-weighting algorithms like TF-IDF to answer search queries.
You’re unlikely to stock every product customers search for, but your “no results” pages should never be a dead end. (Here are some solutions if they are.)
In fact, they can be a great way to recommend related products, reduce bounce rate, and improve the customer search experience.
Unfortunately, most product discovery platforms don’t contain a vector search solution—such as Cognitive Embeddings Search (CES)—to resolve the issue of zero search results.
Cognitive Embeddings turns every product and category into a vector, or “star,” and measures the distance between each star to understand how closely they’re related.
So when a customer searches for a product, CES returns that product and any other product that’s closely related based on the proximity of those vectors, even if the products don’t contain their exact search query.
All of this translates to real conversions, like a 5% increase in RPV as well as an 8% decrease in reformulated searches for Life is Good.
If your product discovery platform can’t return relevant, attractive products for queries with misspellings, synonyms (think: ‘thongs’ for ‘flip flops’), or foreign languages, users may become frustrated and bounce, unlikely to return.
This is where natural language processing (NLP) can help. It interprets search queries more intuitively than basic keyword matching, understanding the user’s search intent and leveraging additional inputs like common variations in pronunciation and even proximity on a keyboard to populate results (think “wnite shwos” instead of “white shoes”).
Those results are more accurate (accuracy boosts revenue), and they remove the necessity of manual input of synonyms and keywords.
They’re also more contextual. A search for “bedside table” would return nightstands and end tables rather than coffee tables or dining tables.
Many studies on ecommerce churn suggest poor search could be the single largest contributor to a phenomenon known as shadow churn, or the percentage of prospective customers who visit and leave before making a purchase.
Is your product discovery platform inadvertently causing shadow churn due to suboptimal search experiences?
After all, if they can’t find your products, they can’t buy them.
Deploying an AI-driven internal search engine specifically designed for ecommerce personalizes the real-time shopping experience to reduce shadow churn.
You need accurate search and browse data to reduce shadow churn, eliminate zero search pages, personalize searches, and improve the overall search experience.
But extracting insights from that data isn’t always simple, especially without being able to lean on your product discovery platform.
So far, only around 7% of major retail brands feel they understand how to leverage search and browse data to improve their site search experience.
Find out how to further leverage your internal site search to boost conversions and drive ecommerce KPIs.