Fewer visitors, more test rides, every one counted.
A used-motorbike marketplace in Malaysia lifted test-ride conversion per visitor by 40% in 22 days, while its traffic fell.
- Two-sided used-motorbike marketplace · Malaysia
- June to August 2026
- Event-level data, checked with the client
Most buyer journeys ended in a WhatsApp chat, and most went cold there.
The marketplace sells used motorbikes in Malaysia and Vietnam. Buyers browse listings, then message the sales team on WhatsApp or book a test ride, and the sale closes offline. The site had about 500,000 page views a week.
In 22 days, 5,048 people showed booking intent on WhatsApp and 400 booked a test ride. 81% of those intents were anonymous, with nothing on file to follow up.
Added on top of what already ran
Active from 26 June 2026, after a 22-day measured baseline.
Recommendations on live listings
On product pages and in banners, retrained on the marketplace's own data in July.
Triggered email
Abandoned-browse and abandoned-sell emails replaced bulk sends to an old, quiet list.
Web push from zero
A way to reach anonymous visitors again: 7,315 subscribers in 22 days.
Only bookings count
First-party identity across devices, and attribution set so that only test-ride bookings carry value.
Baseline against Releva active
Equal 22-day windows. Rates per visitor, so the result holds while traffic moves. Before and after, not a holdout.
| Measure | 4 to 25 June | 26 June to 17 July | Change |
|---|---|---|---|
| Test riders per 1,000 visitors | 1.65 | 2.32 | +40% |
| WhatsApp intents per 1,000 visitors | 22.8 | 27.8 | +22% |
| Test-ride bookings, on 20% fewer visitors | 182 | 203 | +11.5% |
| Recommendation click-through | 31.5% | 39.0% | clicks +154% |
| Push sends | 1,093 | 14,763 | 13× |
Where the attributed value came from
1 July to 6 August 2026. The value of the bikes on test-ride bookings that followed a Releva click.
| Channel | What ran | Share |
|---|---|---|
| Recommendations | Blocks on live listings, 36.6% click-through | 47% |
| Banners | Triggered banners with recommendations: 190 clicks in the baseline, 5,349 once live | 31% |
| Triggered email | Abandoned browse and abandoned sell; 36% open on abandoned sell, against 2.9-4.6% for bulk sends | 19% |
| Web push | Launched from zero on the main site | 3% |
Daily attributed value was 45% higher after the proof window than during it, as the model retrained.
Why the numbers held
Clicks are worth zero. A click counts only if that person then books a test ride. WhatsApp intents carry no value in the report.
Exposed against not exposed. Product viewers who saw a recommendation booked test rides at 6.5 times the rate of those who did not. They chose to look, so part of the gap is intent.
Bike-level match. One in eight test rides was for the exact bike the buyer had clicked in a recommendation, a median of about nine hours earlier.
Next. Recommendations reached about 4% of page views, and 1.9 million searches in nine weeks were not personal yet. The next step is personalized search and recommendations on more pages, with a holdout group.
Get the method and setup notes
How each number on this page was counted, the definitions, and what was switched on in the store, in order. Enter your work email and the notes open; we also send you the link.
Thanks. Open the method and setup notes.
We use your email to know who reads our case studies, and to send the link. Privacy policy
Read next
Vehicle marketplaces
Turn marketplace browsers into booked test rides, and prove it in counted bookings.
Read the use case →ProductSales agent
Search, recommendations, chat and banners on your site, decided for each visitor.
See the Sales agent →SolutionKnow what drove each sale
Orders you can open, counted once, and a holdout that shows the lift.
See the solution →
See what your funnel loses between the listing and the booking
Thirty minutes, with your analytics open. We will map where intent goes cold.
Book 30 minutes
You'll talk to one of us.