Most of what makes a product good is the thing you didn't ship.
I want to talk about that, because the conversations around product tend to be about what got built, the features, the roadmap, the release notes, and almost never about what got removed from a roadmap on a Tuesday afternoon after a week of sitting with it. Removal is invisible work. It doesn't show up in the changelog. It doesn't appear in the press kit. Nobody tweets about a feature that isn't there. But in the arc of a small app that needs to stay small to be good, the cuts are the shape of the product.
Here are the ones I remember most clearly.
The budget tracker.
For about a month, Cue was going to help you budget your gift cards. The idea was: you've got $47 on this Target card, $30 on that Starbucks card, and maybe the app helps you decide which to use first based on expiration dates, spending habits, a little optimization algorithm. It was a good idea. It also wasn't the app I was building.
The app I was building is a memory, the thing that remembers you have the card when you're near the store. Adding a second job ("and here's how you should spend it") mixes a warm helpful voice with a preachy optimizer voice. Those two voices cancel each other out. The first one is what the app is for. The second one is what a different, worse app would do with the same data.
I cut it. I don't regret it.
The receipt scanner.
This one hurt a little. The pitch was: you scan a gift card and you scan the receipt when you use it, and the app tracks your real balance over time. It's a small, useful thing. It's also a completely different product's worth of UX, a camera flow for receipts, an OCR pipeline for line items, a merchant-matching layer, a manual-correction UI for when the OCR is wrong, and an edge case for every promotional coupon applied at checkout.
And at the end of all that work, what the user gets is a better guess at their current balance. Which they could get, more accurately, by tapping "Update Balance" and typing a number after they paid.
The feature was, on paper, a nice-to-have. In practice it was a long stretch of engineering work to save the user five seconds, once per redemption. Those five seconds are not why the app exists. I cut it.
The marketplace.
At one point I was going to let people buy and sell partially-used gift cards inside the app. There was a whole mechanic for it, a listing flow, an escrow layer, a transaction fee, a dispute process. It was going to be the business model, at one point. The free app would become a marketplace.
I spent too long designing this before I noticed what I was doing. I was turning a quiet, private memory app into a financial platform with strangers in it. Every feature required to make the marketplace safe would have made the app louder, less private, more transactional, more like every other app I was trying not to be. The app I was designing was a worse version of the app I'd already started, wearing a new hat.
The real reason to cut it was simpler than any of that. I looked at the paywall I'd already drafted, one Pro subscription, no marketplace, no fees, no transactions, nothing hostile, and I asked myself which app I'd rather use. Not which had the bigger TAM. Which I'd rather use. It wasn't close.
The social feed.
"What if you could see what gift cards your friends have, and trade them?" This one I killed fast, but it still came up. Almost every consumer app in 2025 tries to grow a social graph because that's how the last decade of apps grew. I don't want CardCue Pro to grow that way, because gift cards are a private object, the card itself, the balance, the merchant, the expiration, all of it, and any version of "let your friends see your cards" punctures the thing that makes the app worth trusting in the first place.
The whole value proposition of Cue is quiet. A social feed is the opposite of quiet. Kill.
I'm listing these cuts because I want to push back on a narrative I keep hearing from other founders I talk to, which is some version of "we keep finding new things the product could do." That framing is a trap. Yes, you will keep finding things the product could do. The question is not could, it's should. And the way you answer should is by asking whether the new feature competes with the existing ones for the user's sense of what the app is. If yes, cut it, or do it in a completely different product.
The test I eventually settled on, and I steal this partly from the Complete press kit where I wrote out the product thesis.does this feature make the user trust the app more, or less? Trust is the whole asset. Every feature either builds it or spends it. A budget optimizer spends trust (now the app is judging me). A receipt scanner is neutral at best. A marketplace spends a lot of trust (now I'm in a financial transaction with a stranger). A social feed spends almost all of it.
The features that build trust are the ones I kept. A scanner that works in 400ms and never touches the cloud. A geofence that fires once per trip and only at the right moment. A widget that shows the one card you need. A settings screen with copy that reads like a person. A voice that can be turned off completely. A design system that makes the whole app feel like one object.
Those are the features I kept because the user would notice if they disappeared. The cuts are the ones that pointed somewhere else, even when they were individually good.
A product manager friend of mine used to say that every feature has three costs: the cost to build it, the cost to maintain it, and the cost to explain it. The third one is the one that kills small apps. Every feature you ship has to be introduced to the user somehow, has to be discoverable, has to not confuse the feature next to it, has to fit inside the mental model the user has built. An app with thirty features has thirty things to explain. An app with seven features has seven.
CardCue Pro, if I counted, has maybe ten things it does. Most of them you never think about because they don't announce themselves. That's the point. The features I cut would have made the number bigger. They would have made the app a little harder to describe in a sentence. They would have made the press kit longer. They would not have made the app better.
The things you don't ship are how the app stays good. That's the whole post.