Searching for Intelligent Search
The two grocery stores closest to me shut down about a year ago. It was a small chain that declared bankruptcy. Some of the stores were sold to another, larger chain, while a few remain vacant. Luckily, the ones near me have very recently reopened under a new name and new management. Having been without a nearby grocery store for about 10 months, I was extremely happy with that outcome.
The new company took its time about remodeling the stores before their grand openings. The result, for me, was that although the building looked the same, the floor plan was completely different. My ingrained patterns of search were disrupted. Walking in the front door for the first time, my instinct was to go to where items used to be shelved. But that took me to a shelf where something completely different resided. I’m looking for vegetables and I end up at dog food. Milk used to be on the left; now it’s at the back of the store. And why are the exact same brands of cheese in three different places in the new store?
In a physical grocery store, there are visual cues to help in your search for products. There’s a bit of a challenge, however, when you want to pay for your items using the self-service checkout lanes. There’s no problem with the barcoded items; it’s the fruits and vegetables that pose a search dilemma. Do I look up my zucchini squash under Z or S? The search by picture shows one possibility for avocados, but the store has two varieties. Technically, tomatoes are a fruit, but the checkout screen lists them under Vegetables. It reminds me of this shopworn joke: Realizing that tomatoes are a fruit is knowledge—not including them in a fruit salad is wisdom.
Unlike grocery stores, many companies have digital assets that aren’t readily displayable on shelves. But those assets have tremendous value to the organization and are growing at a rapid pace. It’s as if my local grocery store expanded, virtually, into selling hardware, clothing, and automotive supplies. You might be able to buy them, but it’s not at all obvious, since they are digital products.
Artificial Intelligence Tools
What’s needed is intelligent search. The components of intelligent search are rooted in artificial intelligence (AI) tools, such as machine learning (ML) and natural language processing (NLP). Sinequa’s Scott Parker thinks that the implementation of intelligent search is not only about technology but also involves a focus on people. Advances in technology enables search to be executed by line-of-business employees, who may not completely understand how to search or how to evaluate the relevance of retrieved information.
Technology to the rescue! Intelligent search can help by finding patterns and relationships among siloed data sources indentifying experts within the organization that employees can consult; presenting a holistic, unified view of relevant information from a variety of sources; and discovering new insights from existing enterprise data.
He advises against purchasing new software every time a new business requirement emerges. Instead, stick with a unified platform that can be reconfigured to meet these new requirements.
Simplify, Modernize, Automate
I’m particularly intrigued by Kelly Koelliker, from Verint Systems, Inc., suggests these three tips to make intelligent search work well: simplify, modernize, automate. Simplification in the face of the millions of potential pieces of information on an organization’s virtual shelves is an outstanding idea. A robust tagging hierarchy that enables search to go beyond the words entered in a search box means that if I type “grapes,” the results might include “raisins,” “wine,” or “trauben” (if the company does business in Germany).
She goes on to broaden the implications of intelligent search. The rise of virtual assistants means voice search and search as conversation will become increasingly relevant. Information from social content can be equally valuable as that from more traditional enterprise knowledgebases. Moving even further away from typing words into a search box, Koelliker suggests that relevance depends on context. That search for “grapes” would return different answers depending on whether the searcher is in marketing, scientific research, human resources, or farming.
At Yext, Marc Ferrentino stresses the importance of personalized search results. He suggests that information pushed to people about your company is just as important as what they pull via search. To present a positive view of the company, you need to take control of its data and concentrate on digital knowledge management. Your internal knowledgebase should include key data about products, locations, and people. Particularly for retail stores, you certainly want people to be able to find the closest store and know its operating hours.
Take a mobile-first approach. As more searches are executed on mobile devices, it’s only common sense to structure your data to be mobile-friendly. Make sure to leverage Schema.org markup is another recommendation from Ferrentino. And don’t forget to measure your performance. In that respect, intelligent search is no different from other business processes. Identify the key performance indicators (KPIs) and pay attention to them. Search is not static. As people change their search strategies and as your competitors alter how they present information, your approach to search should dynamically take these changes into account.
As search becomes as commonplace as buying groceries, the need for intelligent search so that customers and employees do not become frustrated while trying to find needed information becomes even more critical.
Happy searching and happy shopping!