“Search” keeps you from finding what you’re looking for. It needs to evolve.
The search engine has been an important resource for web surfers for longer than most people realize. As e-commerce has developed from a niche channel to a consumer behavior pattern, the notion of search has remained at the forefront of the e-commerce experience; there is no other way for a retailer to deliver the long tail of merchandise options in the absence of an endless physical shelf.
As consumers continue to adopt mobile as their primary tool for being online, the importance of search has grown and evolved. What is perhaps most important to focus on is the rapidly evolving relationship between “search” as a concept, and “mobile” as a consumer behavior pattern. This is because a consumer’s access to the internet has historically been through a keyboard, but the shift in form factor preference toward mobile will drive people away from typing, and toward interface solutions that are more in keeping with their form factor of choice. Indeed, for Google, this is nothing new.
Thus, there are 2 questions that we should try to answer: For how much longer will people put up with typing on that small screen? And what are brands and retailers expected to do about it?
“Search” is vital to the current and future health of pretty much every business. But “search” can mean one or more of several very different online resources: a general search engine, specific retailer-hosted search, social media-hosted search, and more. The interface for search to date has been almost entirely through typing, with a consumer entering their query into a form field on a web page. The data from this form would be compared to an index of available potential responses, and a list of results was returned.
With the current trend of internet access happening through devices without keyboards, voice search has moved into the spotlight as the capability that can drive faster, easier web / mobile interfaces. This dates back more than a few years as was highlighted in an excellent Wired magazine article from 2013 which means the trend certainly didn’t start with Siri, Cortana, Alexa, or even Assistant.
But when it comes to technology inception almost always takes a backseat to adoption. So, maybe it actually did start with Siri and Alexa.
The differences between a search engine and a search appliance are important to understand to see where consumer activity is headed. Search engines like Google and Bing crawl the content of the web and create indexes of what they find. Then they rank search results according to a closely guarded, evolving set of algorithms that return results according to their relevance relative to the search terms submitted. It that seems circular, it’s because it is: the more people are searching for the same thing, the higher in the search results that thing will appear. Being popular matters.
A Search Appliance is a dedicated search application that limits its index to its own information. A retailer’s search appliance works in theory very much the same as a search engine, but it presents search results that are limited to the retailer’s own inventory, in either a relative or absolute manner. The information presented to the consumer – the searcher – is delivered by the retailer, but the contents and information being searched are almost always provided by the brands whose items are for sale. Thus, it is up to the brands to take their best shots at anticipating how a consumer will search for a specific product. As with search engines, the more people are searching for the same thing, the higher in the search results that thing will appear. Being popular still matters.
That vastly oversimplified explanation highlights an impending dilemma: Retailers and brands currently manage text-based data to deliver a consumer’s search experience. As consumers move away from typing words into their phones to initiate a search, will brands and retailers be satisfied with the device’s native capabilities at converting a voice input to a text output? Or should they look to take greater control of the consumer experience by moving closer to the consumer’s own words?
Hear The Change, Feel The Impact
This pursuit of control matters tremendously in e-commerce because when consumers adopt new shopping behaviors, brands and retailers typically play catch-up. Retailers will need to create new interfaces for consumers to leverage voice search, and the brands that sell through these retailers will no doubt be tasked with providing the data that will feed voice search results.
Some of the obvious players are already in the game. Google has its own natural language processing (NLP) tools and, no surprise, so does Amazon in its Comprehend cloud-based service. They and countless other start-ups are adding to the general knowledge base surrounding NLP, pushing this technology forward at a rapid pace.
Brands and retailers should be wary of leveraging the paid services of Google or Amazon to get into voice search because neither is a neutral party in the world of e-commerce. As with search today, a competing brand can win an auction for your keywords and display their products ahead of yours. Who knows if a consumer said your brand name, or used the generic term for the product? Plus, if brands and retailers delegate this to a third-party capability provider they are distancing themselves from learning how consumers search for products, while perhaps inadvertently training another company in this key consumer behavior, which itself becomes an asset for the capability provider.
Brands would thus be wise to anticipate retailers and search engines starting to demand “data dictionaries” for voice (and probably visual) search from brands.
The size and difficulty of fulfilling this kind of data may not be readily apparent, but it is significant. Consider how you speak. Humans generally don’t speak the same way they write. When writing, you can see and edit your input before hitting “Submit”. When you speak, however, it’s out there, and what you say next generally adds on what was said before. No matter how skilled, for instance, a writer may be, it’s often helpful (or necessary) for even a celebrated author to create speaking points on cards to refer to during a debate.
Where advertising benefits from the services of some creative firm to inject extra life or humanity into the facts that define the product, performance on a retailer’s search engine is typically hampered by the kinds of indirect sentiments or images that are normal in an ad. Retail search appliances are designed to organize its catalog in absolute terms rather than relative ones. Thus, brands need a different approach to content and data that they supply a retailer’’s search engine versus the content they run on TV or in a Facebook ad. This difference is made more complex with the vagaries of how we each express ourselves in speech.
The Search Party Has Moved Across The Street
Among EMarketer’s 10 Key Digital Trends for 2018 is the prediction that “voice search will become too widespread to ignore or get wrong.” Most retailers and brands haven’t gotten regular text search right yet. Now there is a semi-urgent need to figure out how to supply content to support a whole new way for consumers to interface with the devices that link their intent to purchase with an actual sale.
The future of search is not one capability, but will most certainly be a collection of best-in-class approaches that deliver not just results, but richer context that allows for more meaningful cross merchandising, which leads to bigger baskets, better shopping experiences, and increased loyalty.
There exists an opportunity for some brands and retailers to leapfrog where they are with search and go big in NLP, which promises to be a game-changer for today’s mobile-driven consumer. Some brands and retailers, on the other hand, may start to fall back if they are slow to adopt voice as a significant interaction method. At stake is not only being current with a new consumer behavior, but gaining additional, richer insights about how people refer to products and categories when they are not restricted by the defined structure of text-based search. This is not a bright, shiny object; it is a very real consumer behavior.
Today’s online search puts the burden on the consumer to fit their inquiry within the confines of the tool to get a result. The future of search promises to turn this around, and have the search engine or appliance do the work of figuring out that running shoe, trainer, and jogger may sometimes mean the same thing, but that in the context of this specific consumer, she wants three very different things.
This is the English-language version of a column I wrote that appeared in the February 2018 issue of Ecommerce Brasil magazine, in Portuguese. The online version of the Portuguese article can be found here: https://www.ecommercebrasil.com.br/revista/