The Role of NLQ in Self-Service Business Intelligence

Have you ever walked into a ski shop when you only have a vague idea about the technical aspects of this popular winter sport?

“How can I help you?” the sales assistant will ask.

Now this sounds like it should be an easy question to answer -- but if you have no experience skiing, it can be a tricky question to answer. When you are a beginner, you don’t know what you don’t know. And if you and your sales associate don’t ask each other the right questions, you could easily end up purchasing items that are not well-suited to your needs or capability. Even worse -- your fear of failure and being ‘found out’ as a poser could prevent you from even walking through the shop’s front door.

Guess what? A lot of your line of business folks probably feel the same way when they have to query data to make a business decision. The problem is that data analysts, like the ski shop assistant, have their own language and know a lot of technical things that can make the rest of us feel … stupid.

Like the new customer in a ski shop, your employees don’t want to ask silly questions -- or risk revealing how little they know. Because nobody wants to feel stupid when it comes to data analytics, it’s not unusual for intimidated business users to put their trust in hunches and hope for the best.

The end-result? Your expensive business intelligence (BI) and analytics software sits unused, and your analysts wonder why no one is asking for help using it. It’s exactly this tension that has inspired and motivated the creation of natural language query (NLQ).

NLQ enables anyone, including non-technical business users and smart analysts, to ask questions of their data and get instant answers in the form of best practice reports and visualizations. There are two types of NLQ: open search and guided search. (In time, we should be able to literally ask a question -- or at least freely type a question -- but this is technically still some years away.)

Open search NLQ presents the user with an empty search bar. This approach has a lot of flexibility, but it requires the person who is querying the data to have a comprehensive understanding of what data is available, as well as some basic knowledge of syntax. If you’ve ever asked Alexa a question and gotten a poor response, you can understand why today’s search-based NLQ tends to work best when the questions are simple. If you don’t ask your question the exact right way, you can get an answer that doesn’t make much sense.

Guided NLQ, on the other hand, removes the barrier-to-entry issues found in search-based NLQ by giving the user a choice of filters to use when making a query. Filters mask the complexity of question syntax, language and structure and provide the engine with the context it requires to return actionable analytics. This low-code/no-code approach to BI allows even your most non-technical employees to experiment with different combinations of filters until they get the answer they need to solve a business problem.

Guided NLQ allows for ad hoc, true self-service BI. It allows even your least technical employees to slice and dice data in real time, on their own, without having to wait for someone from your data analytics team to show them how to query data. Guided NLQ will free your data analysts from spending time responding to ad-hoc queries and empower the business user by allowing them to:

Knowledge gaps create a huge barrier to entry for those employees who are new to data analytics and play a huge part in preventing business users from getting the insights they need from the data that’s available to them. In most organizations, the time it takes for an analytics team to respond to a query request can be days, weeks and in some cases -- months. In today's agile, fast paced world, that is just not good enough.

NLQ has the power to change the way your employees interact with their data. When you make data analysis accessible to employees through guided NLQ, it becomes even easier to foster a data-driven culture on an organizational level.