Post-purchase survey on website

Post-Purchase Survey Implementation on Website

Post-purchase survey provides direct feedback from buyer at moment of highest loyalty. Properly implemented survey yields data for funnel optimization, reveals traffic sources Analytics misses (word-of-mouth, podcasts, offline ads), and reduces returns via proactive expectation management.

Where and When to Display

Three placement points with different effectiveness:

1. Success page (immediately after payment). Highest conversion — 30–50%. User still "in the moment", hasn't closed tab. Show 1–2 questions max.

2. Email within 3–7 days. Lower conversion (5–15%), but can ask more questions about actual product experience. Works for physical goods.

3. Inline in personal account. Non-intrusive reminder on next visit. Works for SaaS and subscription products.

Questions: What to Ask

Classic set for e-commerce:

1. How did you hear about us? (single select)
   □ Google / Yandex
   □ Social media (Instagram, TikTok, VK)
   □ Friend/colleague recommendation
   □ YouTube
   □ Podcast
   □ Repeat customer
   □ Other: [field]

2. What influenced your purchase decision? (multi select)
   □ Price
   □ Reviews
   □ Brand reputation
   □ Specific feature/specification
   □ Competitor comparison

3. How likely to recommend us? (NPS, 0–10)

Traffic source question is most valuable. Google Analytics, Meta Pixel, UTM tags lose up to 40% attribution (iOS 14.5+, Safari ITP, VPN, direct visits). Survey fills gap.

Technical Implementation

Data model:

CREATE TABLE surveys (
    id          BIGSERIAL PRIMARY KEY,
    order_id    BIGINT REFERENCES orders(id),
    user_id     BIGINT REFERENCES users(id),
    answers     JSONB,
    nps_score   SMALLINT CHECK (nps_score BETWEEN 0 AND 10),
    source      VARCHAR(100),
    completed   BOOLEAN DEFAULT false,
    token       UUID DEFAULT gen_random_uuid(), -- for email link without login
    created_at  TIMESTAMP DEFAULT NOW(),
    completed_at TIMESTAMP
);

answers in JSONB stores arbitrary structure without migrations when questionnaire changes.

API endpoints:

POST /api/surveys          — create survey on order creation
GET  /api/surveys/{token}  — get questions (public, via token)
POST /api/surveys/{token}  — save answers

Success page component (React):

const PostPurchaseSurvey: React.FC<{ orderId: number }> = ({ orderId }) => {
  const [step, setStep] = useState(0);
  const [answers, setAnswers] = useState<Record<string, unknown>>({});
  const [dismissed, setDismissed] = useState(false);

  const questions = [
    { id: 'source', type: 'single', label: 'How did you hear about us?', options: [...] },
    { id: 'nps',    type: 'nps',    label: 'How likely to recommend?' },
  ];

  const handleAnswer = (questionId: string, value: unknown) => {
    setAnswers(prev => ({ ...prev, [questionId]: value }));
    if (step < questions.length - 1) {
      setStep(s => s + 1);
    } else {
      submitSurvey({ ...answers, [questionId]: value });
    }
  };

  if (dismissed) return null;

  return (
    <div className="survey-card">
      <button className="dismiss" onClick={() => setDismissed(true)}>×</button>
      <QuestionRenderer
        question={questions[step]}
        onAnswer={handleAnswer}
      />
      <ProgressDots total={questions.length} current={step} />
    </div>
  );
};

Step-by-step progress instead of long form lifts completion rate 20–35%.

Email Survey

Survey token embedded in transactional email. Link format /survey/{token} opens page without login needed:

// Laravel Mailable
public function build() {
    return $this->markdown('emails.survey')
        ->with([
            'survey_url' => route('survey.show', $this->survey->token),
            'order'      => $this->order,
        ]);
}

Page /survey/{token} uses same React component but with expanded question set (5–7 vs 2).

Dashboard and Analytics

Minimal dashboard:

• Source distribution — pie chart by answers->>'source'
• NPS trend — line graph NPS per week
• Completion rate — COUNT(*) WHERE completed / COUNT(*) * 100

SQL for source distribution:

SELECT
    answers->>'source' AS source,
    COUNT(*) AS cnt,
    ROUND(COUNT(*) * 100.0 / SUM(COUNT(*)) OVER (), 1) AS pct
FROM surveys
WHERE completed = true
  AND created_at >= NOW() - INTERVAL '30 days'
GROUP BY 1
ORDER BY 2 DESC;

NPS Segmentation and Follow-up

Trigger on Detractor:

// In observer after survey save
if ($survey->nps_score <= 6) {
    SupportTicket::create([
        'user_id'  => $survey->user_id,
        'subject'  => 'Low NPS follow-up',
        'priority' => 'high',
        'meta'     => ['order_id' => $survey->order_id, 'nps' => $survey->nps_score],
    ]);
}

Timelines

Total: 2.5–3 working days.