Chattermill Insights

Get our latest insights right in your inbox

by Sam Frampton on 8 Aug 2018

How to perform effective NPS analysis

Chapter 3 NPS Guide
Chapter 3 NPS Guide

Now that you’ve learned about what NPS is, why it’s important for your business to collect it and best practices for conducting your first NPS survey, it’s time to analyze that data you’ve collected (or plan to collect). While your NPS is score is useful for maintaining a constant finger on the pulse of how your customers feel about your business or service, analyzing the open-ended questions included on your survey give you the “why” of your score. This brings the customer’s voice to the table and gives you a clear, actionable path toward improving your NPS. The problem is that since that feedback is unstructured, there are some challenges to analyzing it.

Theme vs. Sentiment: Organizing your analytics

The first tool you can use to organize your NPS survey information for analysis is tagging, and two ways you can tag your feedback is by tracking themes and sentiments.

Themes

A theme is simply feedback around a recurring topic; for example, shipping times, support tickets or prices. You may already have an idea of some themes just from reading through feedback you’ve already received. Or, as you begin to tag your feedback, you’ll likely start to see themes naturally emerging. One potential difficulty with tracking themes is how many different ways there are to give feedback around one theme; for example, if your theme is “shipping times,” you may see feedback like “fast shipping,” “lightning quick” and “arrived on time.” There are so many different ways for people to refer to the same theme, which can make theme tagging labor intensive.

Sentiment

Another way to organize analytics is by the sentiment of the customer. Sentiment is more closely related to NPS because, broadly, it means assigning a metric to a piece of feedback that details how positive or negative that feedback is.

Approaches to text analytics

There are a number of ways to analyze your customers’ feedback, each with pros and cons.

Rule-based approaches

Rule-based analytics rely on sets of rules to determine which keywords found in customer feedback are positive and which are negative, and then determine a cumulative score for the feedback based on how many positive and negative words were found. Since rule-based analytics can function only within the boundaries of their existing rules, they work well when the scope of what they’re analyzing is narrow.

Excel Macro approaches

It’s possible to use Excel macros to collect rule-based analytics. Keywords can have values attached that denote how positive or negative they are, and then Excel will assign an average value to each piece of feedback depending on the number of positive and negative keywords found. This method is limited because keywords must be manually entered and assigned value.

AI approaches

AI, or machine learning approaches, are data driven and use existing rules to assign sentiments to feedback, but also to predict outcomes and create new rules. As long as the AI has a sizeable input from which to draw its predictions, this approach tends to result in extremely accurate feedback analysis at scale.

Why analyze your NPS feedback?

As you can see, implementing a system to analyze you NPS feedback isn’t easy or quick. But going beyond your NPS score gives you the opportunity to make changes that address the specific themes that impact your NPS, which means a direct path toward improving your NPS overall.

What’s next?

In the next chapter of this series, you’ll learn how to optimize your NPS findings to better your CX for every kind of customer: promoter, detractor or passive.


  • img

    Sam Frampton

    Growth Marketing at Chattermill - A.I for better customer experience