The Real 2025 Numbers

📈 Volume — what we did in 2025

Our verification service grew throughout 2025. The returning-customer figure is the metric we watch most closely — it tells us how often buyers find us valuable enough to come back for the next purchase.

⚡ Speed — how fast we delivered

🛡️ Reliability — uptime & incidents

Every service-affecting incident — defined as any disruption to a paid service lasting more than 5 minutes — is logged on our Status page and resolved openly. We don't hide outages; we explain them.

💰 Customer protection — refunds & fraud caught

We track refunds in two separate categories, because they mean different things:

Bucket 1 — "Issue refunds" (problems we caused or couldn't resolve)

Of these 45 issue refunds, 6 were for delivery errors we corrected at no further cost to the customer. The remaining 39 were resolved as goodwill or service-issue resolution — covering things like accidental double-charges, late deliveries beyond our SLA, or genuine customer-experience problems where we chose to refund even if not contractually obliged.

Bucket 2 — Manual Search guarantee refunds (the policy working as designed)

Our $35 Manual Search service includes a guarantee: if our researchers cannot locate the auction record, we refund in full. In 2025 we honoured this on roughly ~121 orders — about 9% of manual searches. We track these separately because they are not problems — they are the policy working exactly as promised. The other 91% of manual searches found the record successfully.

A counterintuitive fact about refund rates: low rates only build trust when they're verifiable. "We rarely refund" without numbers means nothing. "0.40% issue refund rate on 11,162 orders, plus an additional ~121 guarantee refunds honoured" — that's the actual signal.

Fraud caught

Our anti-fraud monitoring (see Anti-Fraud Policy) blocked roughly ~145 attempts in 2025 — stolen-card payments, bot scraping, sequential chassis harvesting, and re-sale attempts. Each blocked attempt protects legitimate buyers and keeps our prices low.

🌍 Reach — where our buyers came from

Buyers from around the world used JP Sheet in 2025. Our top 8 buyer countries by order volume:

Beyond the top 8, we served buyers in dozens of other countries — anywhere Japanese cars are imported or sold. The full geographic breakdown is available on request for journalists, researchers and academic use.

📞 Support — how fast we replied

"Support contacts handled" includes WhatsApp, email, and the AI chatbot's escalations. Pure chatbot conversations (resolved without human escalation) are counted separately — typically several times this volume — and not included here.

⚖️ Compliance — privacy & legal requests

We commit to resolving every data rights request (GDPR/CCPA/UAE PDPL) within 30 days, free of charge. See our Data Rights page for details.

For law enforcement requests, we follow our published process — valid legal process only, scope down to what is requested, notify the customer unless prohibited. Details on our Anti-Fraud Policy.

✅ What we promised vs delivered

📝 What we want to be better at in 2026

Honest list — the gaps we noticed in 2025 and are working on this year:

• Chatbot accuracy. Our AI chatbot still gives incorrect answers on edge cases (rare chassis codes, country-specific import questions). We're investing in better training data in 2026.
• Email delivery rate. A small percentage of customers don't receive their report email because of strict spam filters at their provider. We're rolling out improved sender reputation and SPF/DKIM configuration to push this closer to 100%.
• Mobile experience. Some pages render imperfectly on older mobile browsers. WCAG 2.1 AA improvements are rolling out throughout 2026.
• Faster manual searches. Manual search averages currently exceed our 24-hour target on weekend submissions. Adding weekend research-team coverage in Q3 2026.
• Currency display. All pricing shows USD only. Adding live local-currency display (PKR, KES, AED, GBP, BDT, etc.) in Q2 2026.

⭐ Customer satisfaction

Our published reviews page aggregates from Google Business, Trustpilot, and Facebook — all independently verifiable. We deliberately don't display every internal kind word we receive; only what's verifiable on a third-party platform.

🔬 Methodology

All figures come from JP Sheet's internal records — primarily our order database, payment processor records (Stripe + PayPal), and uptime monitoring. Where exact figures aren't available, we use approximations and label them as such with a "~" prefix.

This report is not independently audited. It is published in good faith and reflects our internal records as of the date shown. If you spot anything that looks wrong, please email [email protected] with subject "Transparency Report Question" and we will investigate and correct.