Post-sale NPS via WhatsApp: the complete playbook

NPS over email gets a 3-8% response rate. NPS over WhatsApp: 35-55%. It's not just more responses — it's a better sample (the average customer replies, not only the very happy or very angry ones) and real-time action (a detractor is captured in hours, not weeks). This playbook is the full protocol: when to ask, which message to use, how to react to each band (0-6, 7-8, 9-10), and how to squeeze extra business out of your promoters.

Why NPS via WhatsApp wins

• Time to respond: median 8 minutes (vs 3-5 hours over email).
• Response rate: 6× higher (35-55% vs 3-8%).
• More representative sample: neutral customers reply on chat, rarely over email.
• Immediate action: a detractor is captured in hours, and your team reaches out before it turns into a public negative review.
• It's a conversation, not a survey: the customer replies with a score plus a short text, no long form required.

When to ask (the right timing)

HSM template for the NPS question

Name: nps_survey
Category: UTILITY (the customer already bought — it's a purchase follow-up)

Hi {{1}}! This is {{2}} from [Company].

On a scale of 0 to 10, how likely are you
to recommend us to a friend?

Just reply with the number 🙏

The 3 response flows

🟢 Promoters (9-10) — extract extra value

They already want to speak well of you. Make the most of that moment:

Bot: "Awesome, {{name}}! 😊 Would you do us a favor?
Leave a review here:
https://g.page/your-company/review
It takes 30 seconds and helps others trust us."

Possible extras:

• Referral program (the customer gets a discount for the friend they refer).
• Video testimonial request (customer gets a small gift).
• Invite to a WhatsApp VIP group.

🟡 Neutrals (7-8) — find out what's missing

Bot: "Thanks for the score, {{name}}! For us to get
to a 10, what could have been better?
Free response — I really want to understand."

Gold signal: whoever scores 7 or 8 knows exactly what to fix. Capture the response in your CRM, group by theme monthly, and attack the 3 most-mentioned themes.

🔴 Detractors (0-6) — put out the fire in <2h

Bot → human (immediate alert):
"Detractor {{name}}, score {{score}}, bought {{product}}
on {{date}}. Phone {{phone}}."

Human (within ~30 min):
"Hi {{name}}, I'm {{your_name}} from [Company].
I saw your score and I'd like to understand what happened.
Can you tell me?"

Rule: a human (not a bot) handles a detractor. Empathy + listening + concrete action (refund, swap, gift) within <24h. A well-treated detractor turns into a promoter.

How to calculate NPS in practice

NPS = % promoters − % detractors. Neutrals don't count.

Example: 100 respondents in January
  - 60 promoters (9-10)  → 60%
  - 25 neutrals (7-8)    → 25%
  - 15 detractors (0-6)  → 15%

NPS = 60% − 15% = +45

Benchmarks:
  Excellent: > 70
  Good:      50-70
  OK:        20-50
  Poor:      < 20

What to do with the NPS

1. Track it monthly — the trend matters more than the absolute number.
2. Segment by product/channel/agent — you'll find where the problem is.
3. Break it down by cohort — customers from month X vs Y.
4. Cross-reference with churn — poorly handled detractors cancel within 60 days.
5. Publish the NPS publicly (if it's good) — it becomes trust marketing.

Why a detractor captured on WhatsApp becomes a promoter

In an email survey, the detractor writes the verdict and disappears — they go to review sites, post on Google, tell 20 friends. On WhatsApp, they answer with a low score and they're still in the conversation. You have ~24h to rescue them before they post.

Internal research across 30 MercaBot stores: 40-55% of detractors rescued via WhatsApp (human response in <2h + solution in <24h) raise their score in a follow-up survey 30 days later to promoter level.

Mistakes that destroy NPS

1. Bot answering a detractor. An angry customer wants a person. The bot only fires the alert.
2. Taking >24h to reach a detractor. The rescue window is short. Each extra hour reduces the odds.
3. Asking everyone for a Google review. Only promoters. Asking a neutral/detractor is inviting a bad review.
4. Not consolidating feedback. 100 neutral responses saying "confusing service" and no one sees them — total waste.
5. Sending NPS too often. Per purchase/visit is fine. Every week is annoying.

Technical setup

The BSP platform integrates the "ticket closed" or "order delivered" event → a cron fires the HSM nps_survey after X days → the bot reads the numeric reply → routes by band → the CRM updates with score + comment + cohort.