Ever wonder why random search terms come up in your report with no rhyme or reason? For example, you’re a clothing company, but “Paris Hilton” keeps coming in your search results. The same scenario happened to Verizon.
At ShopTalk, Martin Baumgartel, Head of Site Search at Verizon Wireless, tells the story of finding a number of users on their site repeatedly searching for “t-shirt.” If you know Verizon, you know they sell wireless products and services, not apparel. So why the sudden interest in t-shirts? After digging into the data, it turns out customers were looking for a ringtone based on the song “T-Shirt” by Thomas Rhett.
While not every random search will have a logical reason behind it, it lends itself to the insights gathered from search behavior and how that can be leveraged. “People can search for virtually anything, and then it’s up to you, or your search engine, to draw certain connections,” says Baumgartel. Verizon recently migrated to a search engine with a machine learning component to prepare for the AI future.
Knowing that the higher positioned products in the search results receive the highest click-through rate, the new search engine used historical customer data and search behavior to push the most relevant products to the top.
Search engines with a machine learning component will become a necessity as more consumers turn to voice technology for search. Voice will completely change how people traditionally search, says Baumgartel. Consumers are more conversational with their voice devices. They’re not going to say, “Alexa, find t-shirt.” They’re going to say, “Alexa, find the ringtone for ‘T-Shirt’ by Thomas Rhett.”
Anshuman Taneja, VP of Digital Product Management & User Experience at American Eagle, and Amy Vener, Vertical Strategy Lead, Retail at Pinterest follow Baumgartel to discuss their search strategies.
Taneja compares the evolution of search to spearfishing. Spearfishers know what they want, but the tools they’ve used to catch the fish have improved over time. Similarly, “site searchers enter our digital waters knowing pretty well what they want, just like the spearfisher’s decisive action” says Taneja.
Customers who use the search functionality demonstrate high purchase intent with high specificity in mind—qualifying them as MVP customers. Which means, the search experience needs to provide a frictionless path to purchase. Modern shoppers expect search results to reflect their previous purchases and customer behavior. For example, if a shopper typically buys a specific brand of products, these would show up higher in the search results.
To tie together the in-store experience with the online search functionality, American Eagle implemented image search in their mobile app. A customer takes a picture of an item in the apparel brand’s store using the app and the real-time results produce the product in the image along with similar products. Customers can also use the image search to check pricing or for inspiration on how to outfit an item.
The image search feature works on older products, as well, that are no longer in-stock. For example, a woman wants to replace an old American Eagle sundress with something similar. She takes a picture of the dress, uploads it to the American Eagle app, and receives results of comparable dresses. Taneja says that the customers who use image search are in a specific step in their journey—they want to go from the physical experience to a digital one.
In addition to the image search, the mobile app includes voice search functionality. Similar to Siri, the user clicks on the voice button in the app to initiate listening, then it spits back the message in text form. The app prompts customers to keep the search phrases short, “mens jeans” for example, but often people give much more context than needed when using voice search. The context can be useful to learn more about the customer, but too much can confuse the app and produce irrelevant product recommendations or ads. While voice search is a small percentage of total searches, this gives them time to test the voice feature and optimize it before it’s mainstream.
When it comes to deciding which technologies to roll out first, Taneja says they “assess the customer journey, figuring out what are those moments of truth…and how we can step up to the plate and service our customers in those moments.” He adds that at any given moment, their development team is working on ten to twenty possibilities, but only one or two may be commercialized by the end of the year.
Pinterest is another form of search that brands and retailers rely on to engage with customers. When thinking of search engines, Pinterest probably isn’t top of mind, but the platform is heavily “rooted in creating inspiration in people’s lives,” adds Vener. In fact, the platform has over 250 million visitors each month. When someone uses Pinterest to search for an item or idea, they are typically at the point of planning and doing, and much more inclined to purchase something.
Traditionally, businesses were focused on connecting customers to what they want as fast as possible and then figuring out the last signal of intent that drove them to the purchase. Today, it’s about the entire customer journey. “Consumers want to do more than transact,” says Vener. They’re going back and forth from one touchpoint to another, be it Instagram, the brand’s site, Pinterest, resulting in a much more dynamic way to experience the brand. Pinterest helps the customer feel more comfortable with the product, see how others use it in real life, and discover similar products using visual search.
“90% of searches done on Pinterest are unbranded,” states Vener, making the platform an ideal avenue to introduce your brand to new shoppers. Pinterest is focusing on creating tools that enable users to find inspiration and then pivot into action. For example, a customer uses the visual search to find an item, then completes the purchases directly on Pinterest through the shoppable ads. While not all pins are shoppable, any paid ad is eligible for surfacing in search results.