TLDR: Net Promoter Score℠ surveys can hold valuable clues for how to improve your product and better manage your product roadmap. You’ll need a structured approach to turning those clues into actionable insights and changes that result in improved customer retention. Here’s a 4-step guide to getting the most out of NPS surveys.
1. Processing the NPS responses
Processing the responses should get you ready to answer the key questions that drive your analysis:
- What are the most prevalent issues?
- Which customer segments are the issues affecting?
- Are those customer segments critical enough that the issues require resolution?
Comments can be fantastic context for why promoters and detractors rated you as they did. The challenge is to tease the patterns and trends out of the words from your respondents. Be aware that this process is fraught with danger because of its subjectivity; it’s easy for others to argue with your category labels and bucketing decisions later.
Start with a finite list of categories (around 6-10) into which the comments can be categorized. That gives you enough granularity to meaningfully distinguish the differences in comments without slicing the data into too many little slices where no prevalence pattern emerges.
Avoid generic buckets that are meaningless or too broad. For example, a friend of mine at a very large tech company recently saw his company’s NPS results and a large percentage were “product quality.” That’s not actionable.
His company should have either used a more precise category, or sub-codes to support a broad label of “quality.” Examples of subcategories could be:
- Crashes computer / browser (reliability)
- Slows my computer down (performance)
- Features are broken (functionality)
- Hard to use (usability)
Segmenting and dimensionalizing the data
In order to know who your issues are affecting most, you’ll want to know more about the user. If you’re a B2C company, try looking at user demographics. For B2B companies, look at the role of the person and the type of company the person is associated with.
To do this, the survey response has to be joined up somehow with profile information about the respondent. Consider making pivot tables from a worksheet of responses where you’ve added columns with this supplemental data against the rows of responses.
2. Analyzing the results
Using your comment categorization scheme, you’re ready to conduct the prevalence analysis. Do the comments analysis separately for detractors, proponents and passives.
The best way to start is to rank the issue categories from greatest prevalence to least. All things being equal, focus on triaging the most frequently listed issues.
“All things being equal” is seldom true. Some customer segments are more important than others. By performing segmentation analysis – slicing and dicing the data by various dimensions – you’re looking for “hotspots” in the data where prevalent issues or skewed proponent / detractor values show up frequently along another dimension. You then have to decide if that hotspot is important to you.
A hypothetical example: the Admin users in your SMB customer tier (or Long Tail), find your product to be complicated to setup and configure. This shows up in the comments trends and in the NPS scores themselves. However, if your business is focused on large enterprises instead, where the Admin might be a full-time role, perhaps this issue isn’t worth fixing.
Some typical dimensions to slice and dice comments and NPS scores by:
- By user role (persona)
- Customer geolocation
- Acquisition channel
- By tier/segment/size of customer
- By product in a family, or product component
3. User follow-up
Some of the responses merit a follow-up with the responder. Reasons can include:
- Users’ comments were ambiguous, so you need to get to the real meaning of what’s been said
- You want to know more about the business requirement that’s implicated in the feedback
- People with great feedback can be recruited as beta testers
- People with detailed feedback can be used for user studies
These conversations can be truly enlightening, so don’t stop with your analysis. The stories you collect from these conversations will be great examples that illustrate your findings when you share them with colleagues.
4. Sharing the findings
The whole reason to share the findings is to drive positive change in your business. However, you need to make a compelling case for change. Here’s how:
- Make sure you analysis methodology is understood and agreed, so that the results can’t be disputed
- Share your top 3 recommendations to focus people’s efforts
- Use examples from your customer conversations to bring the point home. Analysis + stories = compelling case for change
Here’s a bonus tip: An additional way to get further benefit out of your NPS program is to track your product release dates with NPS scores. A change in NPS score that is closely correlated to a major release date might be an indicating factor for further improvements or proof that your new features had true ROI.