At one point in time, the all-powerful Magic 8 Ball may have been capable of providing sufficient insight into your childhood dreams of fame and fortune. However, it’s likely that your questions have grown more complex and a list of 20 pre-written responses – such as “As I see it, yes”, “It is decidedly so”, or worse, “Better not tell you now” – won’t cut it, or simply won’t be relevant at all if you dare to pose a non yes-no question.
Thankfully, question answering has evolved quite a bit since the advent of this powerful oracle – especially with the rise of Smart Snippets.
But before we get into what exactly Smart Snippets are and how they work, let’s talk about why they’re necessary.
Evolving User Expectations for Search
Google has shaped our expectations for how search should work, and now voice platforms like Alexa, Google Assistant, and Siri are changing them again. Unlike the Magic 8 ball, they aren’t able to tell you if your childhood crush like-likes you. However, they can tell you what the weather’s like in Montreal, and they can also give you a direct answer in a featured snippet when you ask something such as “what are intelligent query suggestions”.
For years, the act of searching on websites and search engines, like Google, meant putting a few keywords into a search bar. Today, we find ourselves moving more towards conversational search queries, as devices like Amazon Echo and Google Home are adopted at a faster rate.
In fact, almost 70% of requests to the Google Assistant are expressed in natural language, not the typical keywords people type in a web search.
And web browsing without a screen is significantly impacting how people interact with your brand. Your customers expect to be able to search how they speak and find what they need with ease – sometimes without using a search box at all. It’s time to elevate search on your apps and websites, and in doing so, prioritize question answering, as that is what will enable your customers to quickly self-serve at every step of their digital journey and, as a result, enable you to compete with digital giants as they redefine the status quo.
Understand Questions, Not Just Keywords
As mentioned above, due to the rise in popularity of voice search, the way in which consumers phrase queries is shifting. Rather than simply inserting key terms:
- “best restaurants downtown Montreal”
They ask questions using full sentences:
- “What’s a good restaurant nearby?”
Enabling your customers to ask for what they want in the way they desire – and actually find it – ensures that you’re creating great experiences that feel more human, positioning your brand to be far more relatable and approachable. Therefore, it’s critical to develop the capacity to identify which queries are actual questions – those that start with who, what, where, when, why, and how.
Only then will you be able unpack the essential meaning of a query, precisely like a person would do while reading written text, and empower your customers to find what they need.
Leveraging Natural Language Understanding (NLU) is key, as it grants you the superpower to interpret and understand the intent behind a search, no matter how complex the query (or question) may be. NLU is what allows Siri to understand that you are looking for a popular and highly-ranked restaurant when you ask, “What’s a good restaurant nearby?”, and you can leverage that capability for yourself, to serve your own customers – primarily by allowing them to serve themselves.
Now that we’ve covered how customer expectations have changed and how you must adapt as a result – three cheers for enhancing question answering capabilities – we’re going to dig into how you can do just that with Smart Snippets and what the end result should look like.
Unearth Insights Buried in Your Text
Smart Snippets grant you the ability to extract a relevant answer from a list returned by the index and throw it to the top of the search results list – no manual intervention required. There is also no reformulation of content involved. The snippet that appears is simply a preview of the section most relevant to the query in question.
The goal here is to answer a customer’s question right away, without them needing to dig through content, and without you needing to pre-define any question-answer combinations. It serves as a real time-saver, enabling people to engage with content and obtain insights without touching the source of that content itself.
Smart Snippets are also utilized in conjunction with a Unified Index, so only the best-of-the-best information, across multiple systems, will rise to the top. Better yet, Smart Snippets reflect a dynamic content curation process. They are driven by machine learning elements that examine browsing paths, search history, user feedback (thumbs up or thumbs down) to identify successful and unsuccessful responses, ensuring that the best answer surfaces to the right individual at a given moment in time.
Surface Answers Right in the Search Results
The onus has historically been on the user to sift through the blue links of a search engine results page (SERP) in order to find answers to their questions. With Smart Snippets, they no longer have to do so. A snippet appears as a special block immediately above the search results and contains a portion of the document (snippet) providing answer(s) to the query (in the form of a paragraph, list, or table), as well as the date and link to where the answer originated.
Essentially, they make the experience of finding information easier by answering questions within the search results rather than forcing the user into further discovery by clicking off the page. This can also be extended beyond the traditional SERP, as Smart Snippets have the potential to enhance an existing chatbot workflow by bringing the power of search and machine learning into your conversation interfaces.
Smart Snippets are powerful. They automate question answering in a manner that is both relevant and timely for an individual user. It should be noted, however, that Smart Snippets are only as effective as the quality of your content.
Content should be created with the intent of it being found, and chances are, if you’re optimizing for the web, you’re already optimizing for question answering. Check out “How to Optimize Content for Smart Snippets” to find out if you’re doing it right, and, if so, how to improve upon what you’re already doing.
Coveo’s platform also offers a Smart Snippets feature. For more information, read our press release about the update here.