Recent advances in text analytics allow for great improvements in the way we design surveys, dramatically improving response and completion rates. Surveys rely on the willingness of customers to share their thoughts on performance. Very often people invest time to make suggestions in the hope that they will be heard, only to get frustrated with endless, repetitive questions. This is not an efficient way of communicating. Find out how text analytics can make communicating with your customers as natural as a friendly conversation.
What users expect from their surveys
A little while back I analyzed 3.3k reviews about 14 top survey platforms to see where they could improve their user experience. Even though these were incentivized surveys, the insights I extracted apply to any survey. The main complaint people had was the fact that the surveys were too time-consuming. They mostly enjoyed short, easy and interesting surveys. Read the full article here.
Current survey design
Why is it difficult to design short, easy and interesting surveys? Regardless of the industry and type of business, employees designing surveys need to cover a lot of ground. It depends on the scope of the project, but often surveys have to capture experiences regarding the products, service, and brand across multiple touchpoints. In order to segment the responses, the survey also needs to gather data on the user. Getting the whole picture requires several questions, possibly with several types of scales to measure satisfaction and opinions.
Until not too long ago analyzing structured data was the only way to get measurable responses at scale. The analysis is often a basic percentage calculation, and the insight is easy to extrapolate: “51% of respondents agreed our burger is the best”. While using this method is convenient, it doesn’t always provide the best quality insights. It requires a lot of questions, frustrating your users and decreasing your survey completion rate. Let’s imagine you are running an NPS survey. You would like to know why someone who gave a detractor rating is not happy. If you try to figure the reason out you will need to ask a lot of questions and thus ending up with a frustratingly long survey.
Designing surveys for text analytics and humans
With text analytics, there is no need to ask many questions, because you are giving users the opportunity to express themselves naturally in human language through a comment. Going back to my previous example, you can have your NPS question “Would you recommend X to a friend/colleague?”, followed by an open-ended question “Please suggest how we can improve”. This means that the customer can quickly share what is top of mind for them, describe an individual situation and make suggestions directly. I’ve covered the full benefits of open-ended questions in this blog post.
The most interesting part about doing surveys in this way is that strategic managers can stay up to date with shifts in the market. A company that stays in touch with customers in this way will evolve with their customer base and not get left behind.
Not all text analytics solutions are the same
It’s important to note that while text analytics has been around for a while, it’s only recent developments in machine learning and deep learning that have made pure AI NLP (natural language processing) solutions possible. Without relying on human intervention our solution extracts relevant statements from comments, highlighting unusual patterns that indicate new trends. Older technology that builds up dictionaries for every vertical doesn’t have the capability to discover new topics and needs to be updated.
Curious to find out how Keatext can help you? Sign up for a trial and use your own data.
FRESH CONTENT DELIVERED STRAIGHT TO YOUR INBOX
- Scaling Corporate Culture Assessment with Artificial Intelligence
- Choosing the right Text Analytics partner for your VoC platform
- VoC platforms: Choosing the right Text Analytics technology
- Survey Platforms: The pitfalls of building text analytics in-house
- Design User-Friendly Surveys with Text Analytics