Why Findability is the Key to Ultimate Self-Service
In customer service, 90% of consumers who have made an online purchase said they used site search to access self-service content. However, many users find themselves frustrated when they can’t find the desired information in an immediate way. That’s because most websites today are still relying on keyword-based navigation.
Natural Language Search Solves the Keyword Dilemma
Keyword searches have a tough time distinguishing between words that are spelled the same way, but mean something different (i.e., hard cider, a hard stone, a hard exam, the hard drive on your computer). This often results in search returns that are irrelevant to the query.
Most sites offer two different types of searches—“basic” and “advanced.” The reality is that search tools are rarely used, and they tend to be useless because it still applies the same keyword pairing principles. You might get more refined results, but they’re still not accurate.
Unlike keyword search systems, Natural Language Processing (NLP) systems focus on the meaning behind the search. It’s what we use as an everyday means of communication among humans; English, Spanish and French are examples. However, one of the biggest challenges in NLP is how to deal with ambiguity—a main characteristic of a natural language.
For example, consider the phrase “run a test.” The word “run” frequently refers to the act of running; it’s only because of this particular context that it means something like “perform.” The difference is obvious to a person familiar with English, but not to a computer.
Semantic Search is what enables us to further analyze the contextual meaning of words. This understanding is what allows computers to choose the best definition of the word “run” even when the syntax and the word itself are the same.
Content Is Not the Problem
Because FAQs make up more than 80% of incoming support questions, it goes without saying that this section is one of the most important resources a website can offer customers.
Let’s imagine you work in a Telecom business and you oversee the knowledge base (FAQs). This content directly impacts how your customers access information for themselves. For this particular example, you have a FAQ describing the cost and plans for your clients to call other countries that has the title:
“What is the price for international calls?”
The probability that your customers will search this FAQ using the same keywords you have used in the content is remote. Instead, they may type queries with misspellings such as:
“how much wll me cost to call to francw”
Your users will always use search queries that describe their particular situation, and therefore their words will be different from yours, because yours will try to describe a more general scenario.
But content managers should only have to worry about the content they are creating, not if the content will be found.
Why Findability Is a Concern
Businesses take great strides getting customers to visit their websites, and to stay there. The last thing they want is for customers to leave a site page for Google or another search engine, where they will likely be shown retargeted ads. It’s a big risk when a frustrated customer is exposed to a competitor.
This is precisely why companies need to measure how easy it is to find information on their site. A findability—or “Semantic Coincidence”—score provides a percentage rating based on how easy it is to find content. More specifically, it details how close the recommended content is to the user query. The higher the score, the higher the match. Results are then sorted by relevance so that the content with the best answer is displayed first.
By creating an extensive dictionary that contains hundreds of thousands of terms and many kinds of semantic relationships called “lexical functions,” an intelligent search engine is able to recognize phrases that have the same meaning, despite using different words to express that meaning. And with this ultimate “meaning algorithm” we are able to know how close or far those sentences are from each other in terms of meaning.
Here is an example from a Change.org visitor, where the question and FAQ language varied but the most relevant answers were still displayed as the top result:
Q: How can I stop receiving emails?
FAQs: Unsubscribe or Update email, Notification Settings
Q: How do I get a list of those who have signed my petition?
FAQs: View, and print, signatures and comments on your petition
Q: How do I change my address?
FAQs: How to edit account and profile information
As a best practice, businesses using natural language search should encourage their users to ask real questions, and not resort to keyword searches. Like in the examples above, users typed in the questions the way they would speak them to a live agent. Businesses that use the prompt “What is your question?” in a search bar instead of just “Search” are likely to have more deflection success.
It seems counterintuitive but the more specific a user question is the more relevant the answer. That’s because while other keyword-based search engines are bogged down by the amount of words, NLP is able to successfully translate the search query using lexical functions.
Inbenta is a SaaS company that specializes in artificial intelligence for customer support and e-commerce through intelligent and semantic search. 408.213.8771, info@inbenta.com, www.inbenta.com