Design and Analyze NPS Survey Frameworks

This skill designs, implements, and analyzes Net Promoter Score survey frameworks - methodology, customer segmentation, statistical analysis, and turning NPS data into action. The standard question is precisely worded ("How likely is it that you would recommend [company/product/service] to a friend or colleague?", 0-10 scale), scored as Promoters (9-10), Passives (7-8), and Detractors (0-6), with NPS = %Promoters minus %Detractors, ranging -100 to +100. Timing strategies are named directly: Relationship NPS (quarterly/bi-annual overall brand measurement), Transactional NPS (post-purchase, post-support, post-onboarding), and Touchpoint NPS (specific interactions like a website visit or feature use).

Use it when NPS needs to be segmented, statistically validated, and tied to a concrete response playbook - not a single generic 0-10 question fired at every customer the same way.

def min_sample_size(confidence_level=0.95, margin_error=0.05, population_size=None):
    z_score = 1.96
    p = 0.5
    n = (z_score**2 * p * (1-p)) / margin_error**2
    if population_size:
        n = n / (1 + (n-1)/population_size)
    return math.ceil(n)

Survey structure layers the core NPS question with an open-ended "primary reason for your score" follow-up, an optional categorical driver question (product quality, customer service, pricing/value, ease of use), and at most 2-3 segmentation questions, with B2B, feature-specific, and competitive-differentiation question variants given directly. Distribution logic is shown as segment-trigger rules (new customers surveyed once 30+ days post-signup, active users quarterly, support interactions 24 hours after ticket closure), and response-rate guidance specifies Tuesday-Thursday 10 AM-2 PM sends, light-touch (not paid) incentives, mobile optimization since 60%+ of responses come from mobile, and a capped reminder sequence (initial plus two follow-ups at 3 and 7 days). Analysis includes a minimum-sample-size formula (targeting at least 100 responses per segment for statistical significance) and a SQL query computing promoter/detractor percentage and NPS per customer segment over a trailing 90-day window. Industry benchmarks are given by sector (SaaS B2B +30 to +40 good, +50+ excellent; e-commerce +10 to +30 good; financial services +30 to +50 good), and a detractor-response playbook assigns a 24-hour outreach owned by customer success versus a 48-hour promoter-activation sequence (thank-you, review request, referral invite) owned by the account manager.

Customer experience and product teams running NPS programs who need segmented triggers, real sample-size math, and a defined response playbook rather than a single survey blast with no statistical backing. Common pitfalls are called out explicitly: resurveying the same customer within 90 days causes fatigue, bot and duplicate responses need CAPTCHA and email/IP deduplication, NPS should never be reported below 30 responses, and seasonal cyclicality plus correlation-vs-causation limits must be accounted for before treating a score swing as meaningful.