The NPS focus should not be the all and end all
Digital business is outside-in, customer-centric. Net Promoter is a management tool that can be used to gauge the loyalty of a firm's customer relationships. It serves as an alternative to traditional customer satisfaction research and claims to be correlated with revenue growth. (Wikipedia). There are relational NPS (based on bi-annual survey) or transactional NPS (based on immediate survey of service)., etc.
It's worth considering including a single diagnostic question with NPS along the lines of, 'What is your single biggest reason for giving this score?', depending on volume of responses and what analytics you have. This can often drive some killer insights and help avoid making assumptions about why people scored well and badly without a full blown equal piece of work. The responses can also really support any campaigns/ marketing communications that you run for customer engagement/retention/etc
Analysis of improvements and using transactional data to drive closed loop and strategic improvements has to be part of the overall program. An aggregated figure won't provide the detail needed to identify which parts of the customer journey need attention or have been revised successfully. It's not all about the number, most of work involves understanding the data that underlies the number, but it's important to get it right as it should provide a benchmark to measure improvement. The relational and transactional NPS scores need to keep separate.
Transactional NPS vs. Relational NPS: All relationships have good days and bad days, and so the immediate surveys for transactional NPS give you a view of what your customers think right now. The bi-annual survey for relational NPS should be used to show the overall NPS and to keep a check on how well you are doing overall. The immediate surveys should be used to check and react to things happening right now - are there process problems, product problems, communication problems. Particularly low scores should be contacted to discuss their issues to try and turn a detractor into a promoter. If you get a lot of detractors or neutrals, then you need to look at why and potentially change something or eliminate the cause.
Sometimes there’re significant differences between the transactional and relational scores. The transactional NPS is much higher, because at the conclusion of that event, you most likely solved a specific support problem for the customer, whereas the relational took into consideration a series of events/lifecycle that had more opportunity for both positive/negative experiences and various levels of expectation. Use the bi-annual as the true test of where the company is at and use the transactional surveys for where the company is headed. The transactional surveys hopefully contain a open ended verbatim that allows you to have a better understanding of what the rating means so you can do further analysis and work to improve the areas that are identified so that your bi-annual results improve over time.
Use all of the information, but for different reasons and in different ways. NPS is a piece but not the entire piece and what it gives you is a great place to start. At a data level, match these pieces of critical information against your churn, new sales, up-sells etc and then you really see them come alive. Remember, the data only drives one to a starting point, you would then want to use the data for action possibly even identifying a few customers that you could take a deeper dive with using some contextual design techniques etc. You have the NPS score, coupled with understanding the touch point that drove it and the themes that drove it. The combination of these attributes mapped out over time really helps drive root cause analysis and what needs to be fixed first and across your entire customer base where are your first "moments of truth" that you would start plugging away at.
The NPS focus should not be the all and end all. Whilst NPS is a straightforward metric that holds companies and employees accountable on how they treat customers, a blending of a range of research will give you more customer understanding and the insight to drive the right actions and accountability in the organization, which should already be driven from the top by a CX executive who provides the right level of business sponsorship. You should also consider trying to collect Customer Effort Score and Customer Satisfaction Score and blending that with NPS, as that gives you a well rounded view of your customers' opinion.
What you measure depends on a few things - the time at which you are capturing feedback in the customer lifecycle, the trigger/basis you use for feedback and your goal. While NPS is an important measure, it is better used as a lagging indicator of the overall customer relationship across time. As such you should be using randomly (not just 2x a year) distributed surveys with a statistically valid sample size to assess the overall relationship and brand perception. This is a good overall view of how customers see you in the market and closest to loyalty (likelihood to refer). Also, be cautious of using NPS as a single measure. If your purpose is to build loyalty for your company/brand, you need to dig a little further to find out what areas your customers would recommend you over the competition and where they would not.