How & When To Survey Your Customers for NPS Feedback

When it comes to using Net Promoter Score surveys to gain insights from your customers, you probably have questions about sampling.

How do you decide how many customers to survey? When should you first survey a customer? What about after that? These are three important questions to think about in advance of getting started.

In this post, I’ll discuss best practices for survey sampling for NPS. While these practices apply to many types of businesses, I’ll relate them to gathering customer feedback in the online world — inside web and mobile applications, and on websites.

When do I first show a customer a survey?

It’s nerve wracking to figure out the best time to talk to a customer. After all, you want to put yourself in a position to get the best possible feedback.

You have to ask yourself at what point is someone familiar enough with the product that they can give feedback. You don’t want to talk to them too soon, as they might not have anything meaningful to share.

Some companies sample their audience based on time (how long a customer has used a product) while others use an event-based approach (a survey is sent only after the customer completes a certain action or

Either can work. They key is to figure out what will work for you and your business.

• Time-based. From my experience, thirty days is usually a safe default. If someone’s been a customer for a month, they probably have something to share. Certain businesses testing out their free trial may want to tighten up that range to learn about the trial experience–maybe survey after seven or 14 days.
• Event-based: Perhaps there’s a certain milestone in your product a customer should reach before seeing a survey. Take Wootric, for example. To get up and running with us, you need to install a small snippet of code in your web application, or install our iOS or Android SDK on your mobile app. It’s possible that you could sign up and not get that code installed within 30 days. I use installation as a milestone that needs to be reached before considering you for a survey.Companies, especially mobile apps, may think about this in terms of number of engagements. How many times has a customer logged into the app on their phone? Once they’ve visited 5 times, perhaps it might be appropriate to survey them. For ecommerce sites, this might be completely transactional — surveying two weeks after a product is shipped, for example.

Re-sampling: When do I do I survey them again?

Your work isn’t done once you’ve sent out the first survey. NPS works best when you’re surveying your customers at a regular cadence, so you have to figure out the right time to re-survey your customers.

Time-based re-surveying

In a web or mobile application, for example, you’ll expect customers to keep logging in. If you’re frequently changing the product, if you are regularly responding to feedback, perhaps a quarterly check-in with customers is appropriate.

If you have an application that doesn’t change as often or you feel like it takes longer for you to process feedback, it might make sense to wait a bit longer before you re-survey a customer.

You should also consider product releases, how able you are to respond to feedback, and even the age of your company. Startups that are constantly updating their product can survey customers more often, as they are often more agile and can respond quickly. You may start with one strategy and alter it — your customers will give you feedback if they are feeling survey fatigue.

Event-based re-surveying

Many companies opt to re-survey their audiences after they achieve milestones, whether that is a transaction or other measure of engagement.

For example, think about a company like Lyft or Uber. A customer takes a ride and at the completion of the ride, they have to rate their experience in terms of stars. You can set NPS up to operate exactly the same way. After any given action, you can ask for feedback and get a score.

EdTech companies, like our customer Magoosh, might combine a time-based and event-based approach. Students can be surveyed periodically during study, and then again upon completion of coursework.

What percentage of my user base should I survey?

Deciding on how many users you sample is very dependent on what you want out of NPS. Are you after a statistically significant NPS that you can report to your industry? Or are you interested in listening to your customers?

On a practical level, you need to consider how much feedback you can handle, and on what schedule.

Reporting a Score. One approach is to sample enough to get an NPS that is statistically significant. The amount of customers you need to sample will depend on how many users you have. If you’re getting worried about small swings in your NPS from period to period based on a sample of fewer than 100 customers, you’re creating unnecessary heartburn for yourself and your organization.

Customer Listening. Customer centric organizations may opt to survey all customers and leverage the Net Promoter Cycle, gathering and responding to a broad base of feedback. Here are two common sampling strategies for listening:

• 100% of Customers at One Time.
  This is the classic campaign approach, sending a quarterly or annual survey out to all customers via email, for example. This is great for when you have a small number of users, and need to get a quick read on what your customers are thinking.
• 100% Of Customers Over Time.
  Sampling all the time results in trickle feedback. This is a modern approach that keeps a constant pulse on what customers are thinking, and avoids the surprises that might come to light when campaign survey results come in. Here’s how this works– if you normally send a quarterly survey campaign to 12,000 users at once, instead you would send to 33% of users, or 4000 over the course of every month. In that way, all 12,000 would be surveyed each quarter. With the advent of NPS survey platforms like Wootric, this has become very easy to do.

Segmentation – A Sample-size “Watch Out”

Do you want to segment your user base and report a statistically significant Net Promoter Score for each group? You need to consider whether you have enough users to do all of this dividing and conquering.

Say you have six different “filters” you’d like to use on your NPS feedback. If your user base is only a couple hundred users or if you’ve chosen to only sample a small percentage of your users, by the time you do your slicing and dicing, it will be impossible to have statistically meaningful scores to compare in each segment. The qualitative feedback you receive by segment will still be valuable…just don’t put energy into diagnosing differences in the quantitative results.

Closing the Loop with Customers

Whatever strategy you use, be sure to respond to customer feedback. If your company responds to feedback individually or by publicly sharing updates to your products and processes, customers will be more likely to leave feedback in the future, because they know you value their input. They’ll be more satisfied with their overall experience, and are likely to reward you with high response rates.

Once you’ve answered these “when and how many” questions and you’ll have a solid sampling strategy to launch with. Remember, it is just a starting point. An NPS platform can crunch the analytics for you so you can quickly see if you are getting the results you desire, and make it easy for you to modify sampling parameters on the fly.

Jessica Pfeifer, co-founder of Wootric, has on-boarded and coached hundreds of businesses on the effective use of Net Promoter Score.