3 Things You Can Learn from T-Mobile’s NPS Mistakes

npsToday, I got a text-based NPS® survey from T-Mobile. The survey missed three of the most powerful and important parts of the NPS system℠.

You can see the entire transcript here. Look closely before you read on, and tell me if you can spot the mistakes.

Do you see the mistakes yet?

(To be fair, T-Mobile might be trying to achieve a different result from their survey than the one we strive to at Promoter.io)

Here are the three lessons you can learn from T-Mobile’s NPS mistakes:

1. Never ask too many questions

A properly designed NPS survey has two questions only:

  • How likely are you to recommend our brand/product/service to a friend or colleague?
  • What is the most important reason for your score?

A rule of thumb: With each added question, your chances of someone opting out of the survey double.

Since T-Mobile asked five questions, I estimate that the portion of people who actually finish the survey is 50% * 50% * 50% * 50% * 50% = 3%!!! This is shockingly close to the average survey response rates of 3-5%, depending on who you choose to believe.

If you ask only two questions, you could get up to 25% response rate, equaling eight times more data—possibly even more, as our average response rate at Promoter.io is nearing 40%.

2. Never forget the open-ended question

Imagine how much actionable data T-Mobile missed out on by assuming they knew the answers.

The three of the four follow-up questions were designed to help T-Mobile figure out the reasons behind the responders’ scores to the first question. Don’t like your phone? Don’t like our coverage? Don’t like how much you pay?

Instead of asking three questions, they could have asked: What is the most important reason for your score?

Then people could have simply responded with: I hate my phone. Or: Your coverage sucks.

Does it take extra work to categorize these responses? Yes. In fact, before fully automated NPS platforms like Promoter.io came along, categorizing responses was one of the most tedious tasks in analyzing NPS surveys. But this is also where the gold in your survey is.

Which leads us to the last and biggest problem with this survey…

3. Always use the opportunity to follow-up individually

I’ve said it before, and I will say it again:

NPS is not about numbers. NPS is about building customer loyalty. And you don’t do that by asking for a number. You build customer loyalty by creating personal relationships with your customers.

Even if T-Mobile missed the mark on the other two points—asking too many questions and not having an open-ended question—there is no reason why they couldn’t do personalized follow-ups.

For example, after seeing that I was unhappy with my coverage but very happy with my phone and rate plan, they could have easily created a semi-personalized response like this one:

Thank you for your valuable feedback, Chad! I know it sucks when you don’t always have coverage, so how about I send you a free WiFi range extender? Would that make you more likely to recommend T-Mobile in the future?

The thing is, T-Mobile is already giving away free WiFi range extenders right now. I know because when I went to their store to complain, that’s what they did for me. And it strengthened my sense of loyalty to them.

But for every 10 people who have my problem, only one or two will go to a store to complain. That’s why such surveys are a perfect opportunity to capture the 80% of passive or unhappy detractor customers and turn them into loyal promoters of your brand.

Notice the other two modifications I made to T-Mobile’s final message:

  • I added the responder’s name. It’s a small but incredibly powerful change. People’s favorite word is their own name—use it often.
  • I added a hypothetical Would that make you more likely to recommend us in the future?

That question is based on neurolinguistic programming. It primes people to think of themselves as active promoters even if they weren’t before.


When running an NPS survey, you should be thinking NOT ABOUT YOUR SCORE but about how you can segment customers into follow-up strategy buckets based on the individualized feedback.

Every survey response turns into an opportunity to grow your customer loyalty, increase organic growth, and reduce churn. Don’t let the survey be the end of the discussion—make it the beginning of one. This is how you create fanatical customers. Treat people with respect, and find innovative ways to deliver value to them using Net Promoter.

P.S. T-Mobile, I’m a big fan and would love to show the amazing things that Promoter.io could do for your NPS program. Looking at you too, AT&T.

3 thoughts on “3 Things You Can Learn from T-Mobile’s NPS Mistakes”

  1. Great post, especially when it comes to the importance of follow-up and short surveys. I’m pretty sure around the third question my response would be “what can I do to make you stop texting me?”

  2. Pingback: The 7 Biggest NPS Mistakes And How to Prevent Them

  3. I helped build that program at TMO. How do you know they aren’t doing followups? I can tell you they were when the program started a few years ago. It is a common Net Promotor strategy to follow up only with detractors or “super detractors”. Also, while i agree there is a beauty in the simplicity of 1 question with the open ended followup, your point about response rates is erroneous. If you ask one question, you may get a 50% response rate per your example. If you keep adding questions, response rate may decline on those subsequent questions, but not on the initial question. My 50% response rate remains on the first question and the subsequent ones provide incremental insight. Further, i would say, don’t get legalistic on rules at the expense of strategic objectives. TMO has a complex business where multiple factors influence churn…coverage, handset, price/value, etc. Asking quantitative questions to pinpoint the category of dissatisfaction is of great value, both in 1:1 issue resolution, and in macro level business strategies.

Leave a Comment

Your email address will not be published.

Scroll to Top