A 3× lift in entry-to-quote, achieved across three releases — and one screen killed at 1% before it could scale.
I led the design rebuild of Acko's new car insurance journey. The product was converting 21% of users to a quote and losing the rest — half the pipeline was researchers with delivery dates weeks away, and the other half was being actively pushed back by dealers at the showroom. The product had no answer to either.
I rebuilt the journey around a clear position against dealer misinformation, a structural fork based on purchase intent, and a graduated rollout that used 1% traffic as a live design experiment. Three releases over three months. E2Q stabilised at 65% — a 3× improvement on the baseline.
New car buyers are one of motor insurance's most valuable segments — clean history, brand-new vehicle, high renewal rate. Acko's pricing was genuinely better than dealer offers. None of that advantage was converting into policies.
Between booking a car and collecting it — typically two to six weeks — dealers have daily touchpoints with the buyer. Acko had zero. Dealers had financial incentive to push their own insurance partners, and they used every touchpoint to do it. When one internal user mentioned Acko while collecting his car, the dealer dropped his quote from ₹1 lakh to ₹35,000 on the spot. Not because Acko was overpriced — because the dealer was protecting their margin.
The product had no presence in any of those moments and no rebuttal once the user came back to it. Acko was losing a segment it should have owned.
PRD said E2Q was 44% overall, 66% with misclicks removed. That looked like a funnel problem — too many steps, too much friction. The signals underneath pointed somewhere different.
Two signals made this clear: Q2S: 11% same-day. 0.5% for "not sure". A 10× gap driven entirely by how close someone is to needing insurance. The funnel can't fix this. ~50% of quotes from far-delivery users — the biggest cohort, almost zero conversion. No early-buy hook meant no reason to commit before the showroom visit.
The product needed to do three things the old journey couldn't. Each became a structural decision — not a screen-level fix.
User enters new car insurance journey → booking status (booked / finalised but not booked / still deciding) → three paths:
| Path | What they get |
|---|---|
| Booked | Direct to MMV + plans — get out of the way. Route them to purchase. |
| Finalised | Lock-in incentive first — give them a reason to commit before the dealer visit. |
| Deciding | Education + comparison — build intent. No pressure to buy. |
The fork wasn't a feature. It was the structural argument that one product can't serve two users with completely different needs — and the booking status modal became the spine of the new journey.
Three decisions that shaped the journey. Each pushback changed something structural. The first set up the tone, the second set up the architecture, the third caught what would have scaled badly.
Pushed for activist copy, not defensive
Early drafts answered dealer claims politely — useless to a user being told the opposite by someone in front of them. Pushed for confident, fact-led, specific. Named garages near the user, real settlement numbers, direct rebuttals to specific claims. Approved through leadership review.
Forked the journey at booking status
A user who hasn't booked yet doesn't know their dealer or delivery date. A user picking up tomorrow does. Asking both groups the same questions created friction for the ready buyer and irrelevance for the researcher. The booking status modal became the structural spine — every screen downstream changed because of it.
Killed the landing page at 1%
Built an activist landing page to address dealer myths. At 1% rollout, scroll data showed 99% of users ignored it — they hit the CTA at the top, the content below was invisible. Carrying it into a 50% rollout would have added friction for everyone and helped no one. Cut it.
E2Q went from 13% at 1% to 34% at 50% after removing the landing page. The most useful signal from the 1% phase wasn't what worked — it was what didn't, fast enough to fix.
Dealer rebuttals were integrated into the plans page itself, not a separate screen — specific named garages, real settlement data, direct answers to the claims dealers were making at the showroom.
Address what buyers actually hear in the room — not generic reassurance.
Fact-led proof users can hold up against misinformation at the dealership.
Named garages near the user, numbers that land, direct rebuttals to specific claims.
Key surfaces from the new car insurance journey: booking status entry, forked flows, plans page, dealer selection pre-payment, and checkout.
Three releases over three months. Each one had a hypothesis and a decision point — not just a traffic split.
1% — February 2026. E2Q at 13%, worse than baseline. Landing page had a 1% scroll rate. Plan naming confusion (1+3 vs 3+3 future scope leaked into UI). Killed the landing page, fixed the naming, did not scale until resolved.
50% — March 2026. E2Q at 34%, above baseline for the first time. Fork holding. Activist positioning landing. Dealer selection pre-payment working. Confidence to scale.
100% — April–May 2026. E2Q stabilised at 59–65% and continued improving. mweb — previously the weakest channel — consistent at ~55%. App at ~75–80%. 3× the baseline.
Q2S held at ~5% across all stages — and that's expected. Quote-to-sale is downstream of when a user's car arrives, not how good the journey is. Knowing what design cannot fix is part of the work.
Three things I'd change if I ran this project again. Each would have moved us faster or taught us more.
