Let's spend some time talking about NPS – specifically how to analyze your NPS data and prioritize your NPS driven repairs. This discussion is going to be a lot more then simply saying “improve quality, release good features, etc…”. I want to look at how to approach your NPS fixes so that you are getting the most bang for your buck. In short – how to intelligently improve the parts of your NPS score that hurts the most.
If you are unfamiliar with what NPS is then I suggest you start by reading our introduction guide to NPS. Understanding how NPS is calculated is critical to being able to tackle the right parts of your NPS problems. Let's refresh it here really quick.
NPS is the index that measures how likely a customer is to “promote” you to their friends and family. Users are asked if they would recommend you on a scale of 0-10 will they promote. 0-6 means they are a detractor, 7-8 means they are passive and 9-10 means they are a promoter. Your final score is calculated by subtracting the detractors from the promoters. Values can range from -100 to 100. A positive value is good, it means you have a net positive user base. A negative value means you are loosing customers and have work to do. A value over 50 is extremely good.
How to Analyze NPS Scores – Group Sizes
Now we come to critical part. We have this data… but what do we do with it. Analyzing your NPS data needs to be broken into different pieces. The first piece is understanding the sizes of the various groups. Understanding the various group sizes is important because it helps you to know and understand where you should put your efforts. Should you focus on helping the Detractors? or should you focus on moving your Passive users into Promoters. One easy approach would be to simply tackle which ever group is the biggest… however don't do this. Instead you want to look at the 5-6 range and the 7-8 range. Which of those two small subset ranges are the biggest. Tackle whichever range out of those two has the most people in it. The reason for this is that if the majority of your Detractor group is in the 0-3 range, then that means that these are seriously disgruntled users. You will most likely never get them back. It is important to understand what caused them to be placed in that range, but it can often be caused by isolated incidents, products or features being discontinued or other items that you might have less control over. You may also find that you invest and spend a lot of time moving users from the 1-2 range up to the 3-4 range… yet they are still not actively promoting your business to family and friends. Remember — we want to focus on smart effort here that is going to help us move the needle.
How to Analyze NPS Scores – Recent Changes
The second is understanding the trends. This is a stat that should be watched on a regular basis. If you experience a sudden drop in NPS scores it could be related to a recent release or change. Maybe software was released that isn't functioning properly. Maybe a feature was removed that users really enjoyed. Maybe a service outage occurred that prevented access and users are frustrated. Being able to notice a sudden change in your NPS score allows you top react quicker to improve your user base. If users moved from being Passive to being Detractors you can often quickly recover them by addressing the issue that occurred. If a service outage occurred you may want to consider a way to provide something free to you users to show your appreciation for the their patronage. These little efforts will often help users who slipped down into the lower levels quickly return to where they want to be long term.
How to Analyze NPS Scores – Bang for Buck
The last area to focus on is around the trends. Many NPS services will have an add on package where a user can post a comment explaining why they choose the rating they did. This data is vital to understanding what is causing or preventing your users from being advertising engines for you. Analyze this verbose data to identify trends. Place them into 10 different buckets that make sense for your business. These buckets could be things such as “Bugs”, “UI”, “Ease of Use”, “Performance”, etc… Having your data grouped together like this allows you to understand what area of your product can use the most investment. Word clouds are also very useful here. You can take all of the verbose text and place it into a word cloud generator to see which keywords appear the most. This will help you identify key categories such as “broken”, “slow”, “confusing”, etc…