A Statistical Exercise, NPS is Not

With Rogue One upon us, we’d like to offer you some sage advice for avoiding the Dark Side of NPS…

In a recent post, we showed you how Chime changed the scheduling of their surveys to better align with their customer lifecycle to drastically improve the accuracy of their results.

What Chime had discovered was that by waiting to survey new customers until their next quarterly batch survey, they were only hearing from customers who had chosen to stick around. In other words, the customers who had tried their product and for one reason or another decided not to continue, were long gone.

Obviously, feedback from lost customers is just as important to the growth of your business as active customers. In fact, you could argue that critical feedback is even more important in some ways.

Chime ultimately changed their approach by automatically triggering surveys to be sent roughly 30 days after a critical event occurred which effectively allowed them to hear from EVERY customer, not just the active ones.

This is just one example of how simple decisions can dramatically impact the accuracy of your customer engagement results.

Another example is a question we get asked quite often, “Which customers should we survey?”, or stated another way, “How many customers should we survey?”.

In short, the answer to both questions is the same … all of them.

Why EVERY customer matters

The Net Promoter Survey is unlike a traditional survey or traditional research in that it’s NOT an exercise in statistical analysis. This is a fact that often gets overlooked and can have a large negative impact on your results.

Within the Net Promoter System, ALL customers matter.

In a recent discussion with one of our customers, they had mentioned that they were only surveying a small subset (10%) of their entire customer base. When asked why, they explained that 10% represented a large quantity of customers and provided them with plenty of feedback for a “representative sample”.

Plenty of feedback is great and given a high enough percentage of responses, you could most likely draw concrete conclusions from the trending data, but you’ll still be missing out on the complete benefits of the process.

Unless you’re only concerned with your overall NPS score (which if you’re Promoter customer, you should know better by now), failing to survey ALL of your customers is simply a bad idea.

Imagine you have all 100 of your customers in a room. 30% of them are detractors and likely to churn. But, which 30 customers is it?

Surveying only a portion of your customers will likely allow you to extrapolate your data and conclude that 30% of your customers are going to churn, but wouldn’t it be better to know precisely who those 30 are?

I think the answer is obvious.

NPS is not about high-level aggregate statistics, it’s your customer’s personal invitation to tell you what you’re doing right and what you’re doing wrong in their individual view.

There isn’t a diminishing value scale to additional responses like traditional surveys, it’s actually the opposite. With NPS, the more customers you survey, the more results you get, the more the value of the survey increases.

Simply put, more feedback means more individual opportunities uncovered. And equally as beneficial is the ability to tie those opportunities to the individual customer and not a statistical subset.

With NPS there is no such thing as too much feedback, even if that’s 10x more that what you believe is needed. There has never been a company that has failed from too much meaningful customer input.

[bctt tweet=”There has never been a company that has failed from too much meaningful customer input. #NPS” username=”promoter_io”]

And, to put all of you statistical nuts at ease, the added benefit of running your NPS campaign correctly is that by doing so, statistical relevancy will happen naturally.

Now, I fully understand that there are a few logical questions and concerns about surveying 100% of your customers, so allow me to quickly address those:

Concern A: We have less-active/non-paying and active/paying customers and we don’t want to skew our results by mixing both together.

As Chime has shown, trial/free/inactive (or whatever your term may be) customers can unearth amazing opportunities in your business. And, determining why they are not active could be the key to 10x’ing your growth.

Solution A: There are a few ways to benefit from the insights of both customer segments while still maintaining a separation.

  1. Create separate campaigns that run in parallel. All of your active customers in one campaign, which would effectively represent your TRUE Net Promoter Score. The other campaign would cater exclusively to your inactive customers. (Also see When to create more than one campaign)
  2. Create filter attributes. This is as simple as adding a column within your database that identifies the customer as either “active” or “inactive”. Once these attributes are in place, you’ll be able to filter your data to reflect each of those groups individually.

Concern B: We don’t have a large enough staff to handle all of the feedback that is coming in.

If you follow our advice, you should be personally responding to each and every customer that takes the time to give you a score and/or feedback. Without a doubt, this can be a time consuming process, but it pays off in droves and should be at the top of your priority list.

Solution B: The easy answer is to create an internal cross-functional NPS team to champion the initiative, but that’s not always realistic. The immediate solution is to drip your surveys over time.

For example, if you have 10,000 customers, dripping your surveys out over 90 days would mean that instead of being hit with 3,000-5,000 responses over the course of a few days, you’d receive a much more manageable 30-40 responses each day over the course of each quarter (based on our average response rates).

If your business or service is more transactional in nature then you’ll end up surveying on a daily cadence automatically which is reflected by the engagement patterns of your customers.

Concern C: How can we possibly unearth any insights within a mountain of customer feedback?

Trust me when I say that too much customer feedback will never be an issue you can’t overcome.

Solution C: Now that you’re properly managing the amount of feedback you’re receiving each day by dripping your surveys over time, you’ll also be able to tag the feedback at the same time.

Tagging feedback trends is as simple as identifying what your customer is talking about (i.e. customer service, price, product, etc.) and assigning a sentiment (positive, neutral, negative).

NPS trend report
Tagging feedback is the easiest way to unearth new trends

Your tagging efforts will all be worthwhile when you see the trending report that comes as a result. That mountain of feedback will be refined down to the most critical insights ranked in order of importance. Best of all, this will be nearly 100% accurate unlike automated text or sentiment analysis tools (more on this in a future post).


The bottom line is that while statistical relevancy can be a small factor in NPS surveys, it should in no way limit the number of customers you survey. Doing so will only limit your results and potential opportunities to drive increased retention and growth.  

When done correctly, NPS will be your biggest offensive weapon against the Galactic Empire. Sometimes you just need to get out of its way so it can help you to win the battle. 

May the force be with you!

Leave a Comment

Your email address will not be published.

Scroll to Top