Harnessing AI-powered search to target millennials: Interview with Justin Sears
Most marketers are aware that AI personalisation can optimise customer experience, but how can AI be applied to search engines to yield better results in real time?
Marketing recently caught up with Justin Sears, the VP of product marketing at Lucidworks, about the use of AI-powered search and how it can help to better capture the millennial audience.
First of all, how is AI-powered search different from regular search engines?
Basic search is matching the text of your search term with the text in the document database – that could be in the form of a category, a specific product description or an FAQ page. No matter who is typing into the search bar, they will get the same results.
AI-powered search allows us to give each user highly personalised results, based not only on the browse and click signals sent by that user, but also on the signals sent by all prior users. We can learn from the user’s behaviour over time to deliver the next best action or recommendation, and the system can tune results in real time, based on the inevitable (but sometimes subtle) shifts in preferences.
In 50 words or less, how does AI-powered search work?
AI-powered search utilises signals that users send on their digital journeys: what pages they visit, what links they click on, what they download or put in their carts. Machine learning models trained on those signals predict intent and shape personalised digital experiences that match that intent.
Why is AI-powered search a powerful tool for targeting millennials specifically?
Retailers are being forced to differentiate from competitors on experience and service. Customers expect a consistent and personal experience regardless of channel, with 70 percent of millennials willing to let retailers track their browsing and shopping behaviours in exchange for a better shopping experience. Millennials are more likely to remember retailers who invest in AI-powered search and provide ‘you-know-me’ types of experience.
Another survey suggests that half of Millennials (56 percent) said they would swap their current brand for a brand that customises to them, while only 26 percent of Boomers were likely to switch from their current brand. As Millennials wield more buying power, retailers will have to work harder to retain their business.
This hard-won loyalty also extends into the digital workplace. Digital natives expect their work experience to mirror their digital consumer experiences. They want knowledge management tools or intranets that make it easy for them to find the information they need to do their jobs.
Whether it’s digital commerce or digital workplace, the mandate is the same – create a digital experience that predicts user intent and provides the information that individual users need, across all digital touchpoints.
How does AI-powered search relate to overall CX?
There are so many things that can create friction and prevent users from finding the information they need, including missed opportunities to make smarter recommendations and failing to learn from user behaviour to better serve them and others the next time.
AI-powered search can shape the customer experience in real time by responding to signals that the user sends, including search results they click on, navigation preferences and other interactions. The system can then automatically boost products or information that’s in line with a customer’s past behaviour and current intent. You want your users to feel like one in a million, not one of a million.
A good example to illustrate this is Lenovo, a customer of ours. Lenovo reported that since implementing signals to track user behaviours, they’ve seen a 30-point increase in customer satisfaction, with substantial growth in click-through-rate, reduced abandonment rate, reduced exit rate and improved ranking in terms of their overall metrics.
Can AI-powered search lead to more sales?
Data shows that AI-powered search leads to more sales for retailers. A report from Salesforce discovered that “6 percent of e-commerce visits that include engagement with AI-powered recommendations drive 37 percent of revenue”. Again, using Lenovo as an example, revenue contribution from search grew 95 percent year-over-year the first year, with 47 percent growth year-over-year in year two, putting them on track this year for 30 percent.
If customers can’t find it, they can’t buy it. And they’re less likely to come back in the future. Clunky search functionality means a clunky experience for shoppers who may never come back. Failing to invest in a personalised search experience costs your company sales and hurts your brand.
What can AI-powered search tell us about our customers/audience?
Research from Accenture and the Retail Industry Leaders Association (RILA) found that 63 percent of surveyed consumers are interested in personalised recommendations. Unfortunately, you can’t make personalised recommendations if you don’t know the person you’re recommending to.
An AI-powered search system relies on signals. By tracking metrics such as bounce rate, average order value, common search phrases, and add-to-cart rates, you can paint a more complete picture of how your unique customers prefer to shop. If you understand your shoppers’ behaviour, you can make informed decisions about how to optimise their digital experience.
Finally, there’s an important benefit of having AI-powered search with signals that extends beyond ecommerce merchandising. Understanding what your consumers want, and the various words they use to describe their desires, is very useful intelligence for positioning your brand, naming a new product, or planning a promotion. Why run costly focus groups or surveys, when you can analyse the preferences that customers share thousands of times per day, via the search box or their browsing behaviour?
With more brands investing in social media and other platforms, why are websites still important?
In the Digital Consumer Study, the Local Search Association found that 63 percent of consumers used the website to find a local business or interact with it. Plus, for bigger ticket items like plane tickets and high-end furniture, many people still opt for shopping through a website.
At the end of the day, customers want options. They want different channels to engage with, more ways to fulfill their order (as we’ve seen with the growing popularity the buy online pick up in store [BOPIS] service), more transparency around product availability, and a connected experience that recognises them as unique individuals, from one digital touchpoint to the next. It’s not that social media is more or less important than a website, it’s that they all need to be connected.
What common mistakes are brands making when it comes to their websites?
Here are four in particular:
- No autocomplete function in the search bar
- Zero results pages that end up frustrating customers
- No synonym or misspelling detection
- No chatbot or a chatbot that doesn’t use natural language processing
What impacts does AI-powered search have on the jobs of marketers?
Identifying with customers is one of the key tenets of effective selling, and many of us are still relying on a technique that was invented in the mid-80s – ‘persona building’. These personas represent a cohort of individuals who will purportedly watch, buy, or interact with the same things. And while this is okay for a cold start, it’s no longer sufficient given the range of sophisticated AI options that did not exist 30 years ago.
In a 2019 survey of retailers doing more than $100 million in annual sales, nearly two-thirds said they were collecting some sort of user behaviour, but only 49 percent said they were using AI for query intent detection. That puts them at a disadvantage. Machine learning allows you to understand intent by figuring out what a person is looking for – despite how they spell it or describe it. Signals provide marketers and merchandisers with an accurate and real-time understanding of their customer’s needs.
Justin Sears is the VP of product marketing at Lucidworks.