---
title: The Feature We Almost Built That Nobody Needed
description: "A feature we almost over-built for ourselves, not travelers — and the kill criteria we use before shipping to anyone."
canonical: https://nowah.xyz/blog/feature-we-built-nobody-used
lastModified: "2026-08-07T08:32:23.036Z"
---

# The Feature We Almost Built That Nobody Needed

A feature we almost over-built for ourselves, not travelers — and the kill criteria we use before shipping to anyone.

The most expensive features are the ones built for the team’s taste, not the traveler’s job. We have almost shipped those. Private testing is where they die.

## Launch-day silence (the failure mode)

## The excitement and the blind spot

![Illustration for this section](https://pics.nowah.xyz/website-media/founder-011-img-1.webp)

The idea came from our own experience. When comparing trips, we wanted to see everything laid out cleanly: flight costs, hotel costs, total duration, number of stops, walkability scores, weather forecasts. We assumed that if we wanted this, users would too.

We skipped the part where we asked them. That was the blind spot.

We spent three weeks building, refining, and polishing. The team was energized. Every standup had demos of new comparison features. We were solving a real problem in an elegant way. Or so we thought.

## Launch day silence

We nearly shipped the feature with a small in-app prompt directing users to try the comparison view. The first day, a handful of people clicked through. The second day, fewer. By the end of the first week, the usage graph was essentially a flatline.

Near zero adoption in two weeks. The most polished feature we had built was the least used.

## The post-mortem

![Supporting diagram](https://pics.nowah.xyz/website-media/founder-011-img-2.webp)

We did what we should have done [before building](/blog/five-things-before-building-ai-product): we talked to users. We scheduled ten calls and asked people about their decision-making process.

The answer was consistent and humbling. Users did not want to compare options themselves. That was the whole point of using an AI agent. They wanted the agent to compare for them and recommend the best one. Building a comparison dashboard was solving a problem our own product had already solved better through conversation.

One user said it perfectly: "If I wanted to compare things in a spreadsheet, I would use a spreadsheet. I use Nowah because I do not want to compare things."

We had built for ourselves, not for travelers. We projected our analytical preferences onto people who had specifically chosen a conversational product to avoid that kind of analysis.

## What we changed

The post-mortem led to a process change that has been more valuable than any feature.

Before this failure, our process was: idea, design, build, launch, measure. After this failure, it became: idea, validate with ten-plus user conversations, prototype, validate again, then build.

The validation step is not optional. It is not a checkbox. It is the gatekeeper. No feature enters the engineering pipeline without evidence that real people in private testing want it.

## What we built instead

We replaced the comparison dashboard with enhanced conversational recommendations. Instead of showing users a comparison table, the agent now explains its reasoning more clearly: "Option A saves you two hundred dollars but adds a three-hour layover. Option B is direct and gets you there by dinner. Based on your previous trips, you tend to prefer convenience over savings. I would recommend Option B."

Usage of the conversational recommendation feature was immediate and sustained. Users engaged with it because it matched why they were using the product in the first place.

## The lesson

The most expensive lesson in product development is building something nobody asked for. Three weeks of engineering time, gone. But the process change it triggered has saved us multiples of that investment since.

If you are building a product, especially an AI product, resist the urge to build what you think is cool. Build what users demonstrate they need. And the only way to know the difference is to ask before you build.

---

Nowah is an AI travel agent that searches and books real flights and hotels through conversation — no filters, no thirty open tabs. [Plan your next trip](https://app.nowah.xyz).