In poll after poll, retailers report that improving the customer experience is one of their top priorities.
The 2018 Digital Trends in Retail survey by Adobe and Econsultancy asked 600 global retail leaders the primary way their organizations would be differentiating themselves from competitors in the coming five years. Customer experience was the top answer. Almost twice as many retailers (26 percent) planned to focus on customer experience as the second-ranked customer service (14 percent).
In another survey, Jabil’s The Future of Retail Technology, 95 percent of retailers said they were looking to technology to improve the customer experience, and 73 percent strongly agreed that technology innovation is imperative for meeting the high expectations of today’s shoppers.
Given these stated intentions, retailers seem to be missing a critical customer experience factor — ecommerce search.
From a user experience perspective, most ecommerce sites don’t do a good job with search, says Lauryn Smith, senior user experience researcher at Baymard Institute. “Almost every site within the 60 highest-grossing ecommerce sites in the United States would have to make substantial improvements to get to a performance that would be viewed as acceptable or good according to our benchmark,” she notes.
When Search Fails, Consumers Bounce
In a Baymard Institute usability study on ecommerce search, test subjects were either unable to find what they wanted or abandoned the search almost one-third (31 percent) of the time. Almost two-thirds of the time (65 percent), it took more than one attempt for subjects to find what they were looking for. And these numbers haven’t changed much over the last several years, Smith adds.
If potential customers don’t see what they want on the first search results page or they read dreaded words “your search returned no results,” most of them won’t recheck their spelling, try a synonym or use a filter to eliminate pages of irrelevant results. They will leave, even when the retailer actually carries what they want.
“In the absence of good search results, the average user will simply think you don’t have what they’re looking for and will go somewhere else,” Smith states. Not only is the consumer dissatisfied, but a potential sale and possible long-term customer is lost.
It should be no surprise to retailers that Amazon “has the best performing search from what we can tell in our benchmark,” according to Smith. Amazon has long been recognized for its user-friendly search tools and enticingly accurate recommendations. (For a detailed breakdown of the key features of Amazon’s search, see Lucidworks’ “How to Create an Amazon-Like Experience with Fusion.”)
Smith adds that Walmart is the other major exception to the generally poor search experience that plagues retail.
Retailers’ Big Data Solution Is AI
Despite the importance of search to the customer experience — not to mention sales — the technology that powers ecommerce search has been an afterthought for most retailers, observes Grant Ingersoll, chief technology officer for Lucidworks.
“Retailers invest millions in content management systems (CMS), ecommerce platforms and analytics solutions. They have become essential parts of an ecommerce site,” Ingersoll explains. “All three need to work together for an ecommerce site to make money, but for some reason, retailers continually underfund or don’t consider search,” even though search is also critical to making those systems work together.
“Search connects the CMS and the ecommerce platform,” Ingersoll continues. “It gets the user from intent — a query — in the CMS to execution — buy/convert — in the ecommerce platform. This is where the money is made.”
Ingersoll calls search a “force multiplier” when it comes to analytics. Search adds value to interaction data because a user query shows intent and has context. The clicks and sorts that follow add more context. “Making this ‘record of intent’ active works for you by helping keep users on your site, helping users buy/convert, helping users buy/convert more, and reducing the time and effort spent merchandising, ‘searchandising’ and tuning relevancy,” he says.
For years, the search and recommendation engines built into CMS and ecommerce platforms were adequate for most retailers, but recent advancements in artificial intelligence (AI) and machine learning (ML) are taking search functionality and performance to a new level. This requires more advanced search technology. (For a deeper dive, see What Is AI-Powered Search? on the Lucidworks blog.)
A vast amount of data is involved in ecommerce search, including user preferences and behavior — search terms, pages viewed, items added to and subtracted from the cart, purchases, etc. — and all the attributes of thousands of products.
All of this data makes search “one of the areas ripe for integration with AI,” said Pedro Palacios, associate director of commerce strategy at global marketing agency VML, in a story for Internet Retailer. Because most current search tools “rely heavily on manual product categorizations and keywords,” AI’s ability to make sense of unstructured data can be a game-changer. “Adding natural language processing and data management tools to your search feature will make a customer’s search query pull up more relevant results, creating an immediate positive effect on sales,” he explained.
“Machine learning effectively represents the solution to the big data problem,” according to the FitForCommerce Annual Report 2018. “The beauty of a smart system, a solution that leverages machine learning, is that the more data it devours, the better it performs.” Plus, with software-as-a-service (SaaS) models designed specifically for retail applications, “AI technology is now robust and affordable for retailers of all sizes.”
Even though “the potential business impact of AI is massive, adoption rates are still low,” according to “Consumer Experience in the Retail Renaissance,” a report by Deloitte Digital and Salesforce. However, search is one of the most popular use cases among retailers and brands that have adopted AI for at least one application, according to the report, with 40 percent using it to provide relevant search results.
Choose an AI Vendor Based on Specific Needs
Businesses across most industries and across the globe are racing to deploy AI, with spending on AI systems projected to rise 44 percent this year to reach almost $36 billion worldwide, according to a report by the International Data Corporation. The retail sector is expected to be the top spender, with a global investment of $5.9 billion in 2019. Retail will also lead in the U.S., where two-thirds of AI spending is projected to take place.
Amidst all of this activity, retailers are faced with a legion of companies purporting to provide AI-powered systems, with some of those potential vendors stretching, misrepresenting or misusing the term to gain new customers.
“AI is based on algorithms. It uses computing power to solve specific problems faster and often more accurately than humans can,” explained Blake Morgan, a customer experience futurist and author, in a Forbes contributor post. “In order to truly be considered AI, the system needs to be able to learn contextually and then apply that learning to change how it does things, [but] people are left to believe [by] marketers that AI is in nearly everything.”
In selecting vendors for ecommerce search, it may not be useful for a retailer to spend a lot of time trying to understand the underlying technology. Rather, Ingersoll advises that retailers arm themselves with questions that will help them discover how search and AI/ML will drive value for their business.
He suggests starting by asking the vendor how the solution has helped, and will help, customers get to a retail website, to buy/convert and to buy/convert more. Next, find out how the solution tests and improves traffic volume, conversions and incremental purchases. Finally, how does the solution reduce the staff’s time and effort while accomplishing these goals?
Once they’ve gotten answers to those essential questions, Ingersoll says retailers should direct the conversation toward the key performance indicators (KPIs), processes and branding objectives of their organizations. Here are some ideas for those questions:
- How many ways can search results be ranked, e.g., relevancy, the user’s click history or past purchases, aggregated purchases, inventory availability? What other factors can be used to determine how search results are displayed?
- Do you offer recommendations? How many different factors can we use for recommendations? Are the factors pre-set or can we create our own?
- How much will we be able to adjust the settings for results, recommendations and experience?
- Does the platform use real-time data? If not, how often is it updated?
- Here are the KPIs we use to determine success; will this system allow us to keep using them? If not, what metrics would we be able to track?
- Will the system suggest synonyms and spelling corrections based on actual user data? Can it predict misspelling or new synonyms before they start showing up?
- What metrics can I use to determine how my A/B tests will be scored? Can I set a custom metric?
- What does “personalization” mean in your system? What factors are used to personalize search results and recommendations? How is it done for visitors who have no profile?
- How can I adapt this technology if my business model changes?
Hesitation Is Risky
FitForCommerce, a retail consulting firm, cautions that retailers waiting to deploy AI and ML are postponing incremental benefits. “Well-planned AI-driven solutions get more effective over time in a non-linear fashion,” the report states. “They are self-leveraging. So, the ROI you realize two years from today will be vastly greater if you start tomorrow than if you start six months from now.”
“If you want to be successful with AI and think there may be a threat from AI-driven competitors or new entrants, you should start learning now about how to adapt it to your business across multiple different applications and AI methods,” write the authors of “Why Companies That Wait to Adopt AI May Never Catch Up” in the Harvard Business Review. “Vendors are developing a vast variety of knowledge graphs and models that use techniques ranging from natural language processing to computer vision. If one exists for your industry or business problem … that will speed up the process of AI adoption.
“In short, you should get started now if you haven’t already, and hope that it’s not too late,” they warn.
Marie Griffin is a writer and editor with extensive experience covering retail, technology, media and other B2B topics. She has held multiple editorial leadership positions, including editor/associate publisher of Drug Store News, and has been freelancing for web/print publications and marketers since 2001.