Build What Users Do, Not What They Say

The features users asked for

Across our first 30 user interviews, the top requests were consistent:

1. Receipt scanning — "point my camera at the bill, fill the expense automatically"
2. Multi-currency — "handle our Bali/Bangkok trip in baht and rupees"
3. Complex splitting rules — "four people pay for hotel, three for fuel, two for snacks"
4. Recurring expenses — "rent and WiFi every month, automatic"

We built all four. Two months later, we looked at usage data.

What the data showed

Less than 3% of active users touched any of those four features.

• Receipt scanning: ~1.8% of users tried it once. Most didn't come back.
• Multi-currency: ~0.4% of expenses used non-INR.
• Complex multi-payer rules: ~2.1% of expenses had any custom split logic beyond "equal".
• Recurring expenses: ~2.7% of users set up at least one recurring rule.

Meanwhile, equal-split + UPI settle + group balance view were touched by 95%+ of active users every week.

Why users ask for what they don't use

Three reasons we've identified, in roughly this order:

1. Aspirational use cases. Users imagine a future trip to Bali where they'll need multi-currency. They don't plan around their actual weekly Goa-trip-with-college-friends pattern, which is INR-only.
2. Interview psychology. When asked "what would you want?", users default to listing more things. Saying "just keep the basic thing fast" feels like a non-answer.
3. Feature comparison shopping. Users have seen Splitwise, Splitkaro, and Tricount. They list features those apps have because the comparison is fresh in their mind.

The reframe

We deleted (or hid) most of those four features. Receipt scanning we kept but moved deep in the menu. Multi-currency we paused. Recurring expenses we kept but downgraded from "hero feature" to "power feature."

We doubled down on the three features users actually used:

• Equal split — but with custom-amount escape hatch when needed.
• Settle-up via UPI — one tap, opens GPay/PhonePe/Paytm with amount pre-filled.
• Group balance view — see who owes whom at a glance.

Three features. That's the whole product. We say it on the homepage. We say it in the App Store listing. The discipline of saying "these three things, done well" is itself the differentiator against Splitkaro's feature sprawl.

The lesson

Run user interviews to find pain. Don't run them to get a feature wishlist. The wishlist is asking for confirmation; the pain is what you actually solve. After every interview, ask "what did you actually do this week, and where did the friction live?"

Then ship the features that remove the friction users encountered, not the features they imagined needing. The data — usage, retention, NPS — will tell you within 4 weeks whether you got it right.