NPS for Agile Teams

---

Origins

Net Promoter Score was introduced by Fred Reichheld in a 2003 Harvard Business Review article titled "The One Number You Need to Grow."¹ The premise: a single question — "How likely are you to recommend us to a friend or colleague?" on a 0–10 scale — predicts company growth better than complex satisfaction batteries. Subtracting the percentage of detractors (0–6) from the percentage of promoters (9–10) yields the score.

NPS became dominant in the 2010s, then took heavy criticism in the late 2010s and 2020s. The honest picture today: it is a useful diagnostic when used with care, and a wildly overused single-number replacement for product judgment when used carelessly.

How NPS Adapts to Agile Teams

For agile teams building a product, NPS has two distinct applications:

• Product NPS — survey users about the product itself. Used to gauge overall product-market fit and direction.
• Feature or moment NPS — survey users immediately after a specific interaction (just shipped a feature, just completed a workflow). Tightly scoped, faster feedback.

Feature-level NPS is generally more useful to agile teams than product-level NPS. The former produces actionable signal tied to specific work; the latter is too aggregate to drive sprint-level decisions.

The Useful Mechanics

• The question stays consistent. Drift in wording invalidates trend comparisons.
• The follow-up question is the gold. "Why did you give us that score?" produces the qualitative data that actually tells you what to do.
• Segmentation matters. NPS averaged across all users hides everything. NPS by segment (new vs. returning, enterprise vs. SMB, geography) often reveals divergent products under one brand.
• Trend beats absolute. "Our NPS is 30" is meaningless. "Our NPS rose from 25 to 35 over Q1" is signal.

The Common Failures

• NPS as a target. The moment leadership says "we want NPS of 50," teams chase the number — by surveying happy users more, by gaming question framing, or by ignoring promoters' qualitative complaints.
• Survey fatigue. Asking the same users monthly produces declining response rates and biased samples (only the very happy or very unhappy bother to respond).
• Single-number tyranny. NPS reduces a complex customer relationship to one digit. Used as the only signal, it suppresses richer data the team should be paying attention to.
• Cross-industry comparison. "Our NPS of 32 is below SaaS average of 36" — average across what kind of users, what segment, what touchpoint? The comparison is usually false precision.
• The 0–6 "detractor" framing. NPS treats a 7 as neutral. For many products, a 7 is genuinely satisfied. The categorization can be misleading.

NPS Alternatives and Complements

Modern product practice supplements (and sometimes replaces) NPS with:

• CSAT (Customer Satisfaction): a more direct "how satisfied were you?" question, scoped to a specific interaction.
• CES (Customer Effort Score): "how easy was it to accomplish your task?" Often the most actionable for product teams.
• PMF survey (Sean Ellis's "how disappointed would you be if you could no longer use this?"): a sharper test of product-market fit than NPS provides.
• Behavioral metrics: retention, frequency, depth of use. Often a more honest signal than what people say in a survey.

The healthiest pattern is to use NPS as one input among several, not as the answer.

What NPS Is Good For in Agile Contexts

Despite the criticisms, NPS earns its keep in a few specific cases:

• Tracking direction over time. Repeated NPS measurement of the same segment over months produces a trend signal that no single survey can match.
• Identifying detractors for outreach. The follow-up question on detractor surveys is high-signal customer feedback.
• Cross-team alignment. A single number, however imperfect, gives a CEO and a developer a common reference. The downside (oversimplification) is sometimes also its upside.
• Quick feature-level sanity checks. Did that release improve or worsen the customer's sentiment? NPS post-launch is a cheap way to ask.

Coaching Tips

Always include "why."

The open-ended follow-up is where the value is. The number alone is barely worth collecting.

Segment ruthlessly.

NPS by user type often reveals two or three different products under one brand. The aggregate hides them.

Use feature-level NPS.

Post-feature, single-question, in-app pulse. Cheap, fast, tied to specific work.

Watch sample bias.

Self-selected responses skew. Only the very happy and very unhappy fill in surveys. Pair NPS with behavioral data to triangulate.

Never let leadership target the number.

"Improve NPS by 5 points" produces gaming. "Investigate why detractors gave us 4s" produces insight.

Reach out to detractors.

The detractor follow-up call is the single highest-signal product activity available. Their feedback is unfiltered.

Summary

NPS is neither the panacea its early advocates claimed nor the waste its critics now describe. It is a directional signal — most useful in trend form, segmented, paired with qualitative follow-up, and complemented with behavioral measures. Teams that treat it as one input among several extract value from it. Teams that treat it as the answer end up with a single number they're optimizing instead of a product that's actually getting better.

Back to Agile Metrics