Search is turning into conversational interaction. Neither typing nor speaking is required; only clicks suffice. Open the search box and enjoy the journey. Brand are increasingly turning to the search box as the new cool place to start a dialogue with their customers and tell them stories in a refreshing and fluid way.
The second part of this series focused on designing typeless search experiences for findability: Using type-ahead, related tags, trending suggestions, visual filters and other motion-based micro-iterations to guide people through the process of finding what they’re looking for.
This post, third in the series, will focus on designing typeless search experiences for discoverability. The difference between findability and discoverability is in the declaration (or absence) of intent. That is, findability refers to satisfying a pre-existing need, helping people find a product/content they already know or assume it exists. On the other hand, discoverability is the ability to nurture serendipity, that is, inspire people to discover a product/content that they were not aware of previously, or at least the need/intent was never articulated nor declared.
Findability: Users can easily find content or functionality that they assume is present in a website.
Discoverability: Users encounter new content or functionality that they were not aware of previously.
[Quote from NN/g Nielsen Group: Low Findability and Discoverability: Four Testing Methods to Identify the Causes]
But search engines are not clairvoyant… or are they? How can we, and consequently the search experience, understand what shoppers don’t yet consciously understand about themselves? And how can it be done while preserving the privacy, the unknowability, of each individual shopper’s past behavior?
Predicting the next purchase
Your customers’ search journeys are truly insightful in terms of shopping intent diversity. Some of these relationships can be inferred by intelligently examining brand shoppers’ aggregate shopping practices while preserving shopper privacy. In this respect, intent-purchase relationship can be inferred in the following ways::
· What items were searched and bought together, in any order, during a shopping session
· In what order were common items purchased?
· What was most often bought next after each product?
Mapping these intricate relationships, future intent behavior can be predicted.
Sankey diagrams, for example, help the data science and engineering teams expose the popularity of recent identical shopper journeys. Elegant lines guide us through the most common intent order in which products were recently found, discovered and purchased; line thickness conveys the approximate popularity of each journey. In the following simplified example, dozens of shoppers bought the same items in a similar order. Some aggregate shopping sequences, shown in bold, were more common than others:
milk > egg > bacon > potato
milk > almond milk > cereal > vanilla yogurt
milk > egg > bread > mayo
cream cheese > salmon > bagels > orange juice
These sequential choices, alongside atomized glances at what shoppers typically search for and buy immediately after the currently selected product and past query, underlie the evolving understanding of aggregate shopper intentions. Those aggregates guide the deepening dialogue with the current shopper — so their their intention is better understood.
What we have until now called the priority of product Findability in a search, is evolving into a new and different search priority: Discoverability, defined by the aggregate diversity of shoppers’ intents. Whether shoppers go from search A to B to C or from search B to A to D, and richness of search-engine position, ultimately help shoppers engage with more products — and declare their intent.
Search acts as a shop assistant by amplifying the customers intent journeys. The example below displays how when someone searches for “milk”, finds the milk they’re looking for, adds it to the cart, and then voila!, a new carousel with fresh suggestions from other anonymous shoppers’ past search journeys is displayed below the product card.
The very same behaviour can be trigered from the search box too, in this case a journey designed to guide customers in a basket building mode.
Conveniently, next query suggestions can directly offer related products blended with organic results. These type of lateral or digressive suggestions are best offered once the user has scrolled down into the second or third page in order to avoid cognitive overload. Below two examples.
Prediction and facilitation of typeless shopper activity is much simpler in concept, and more respectful of privacy, than we might have imagined in the previous era of e-commerce.
Future use cases for typeless
These innovations, though, are just a preview of the relationship that a typeless dialogue-oriented shopping journey makes possible.
Buying milk, for example? “Don’t forget cheese,” the search engine says — “and how about some gouda or Havarti today? It’s on sale!” Not sure what to do with that enticing queso? The search engine already knows: “We have recipes!” Buying bacon, too? “How about adding sandwich items to your cart — and congratulations, here’s a related coupon and sandwich recipe to make your day both tasty and healthy!”
Without confusing the shopper, typeless dialogue makes the shopper’s journey more flexible by making confounding decisions intuitive and seamless:
· Related brands
· Tags reflecting related types of the same product
· What shoppers tend to buy next
· Useful product information
· Inspired product uses
· Related coupons & savings
All these automated assists give the shopper a sense of being understood, of having their own personal shopping assistant, and of being entertained and informed during their journey. Intuitive typeless dialogue instills satisfaction, fresh discovery, trust, and joy in a loyal relationship between customer and retailer.