Buy the App That Says No: A Five-Minute Test for Software Restraint

Watch a software demo and you'll notice the shape of it. Confetti. A cursor flying across the screen. A number going up. Nobody spends ten minutes explaining what the product refuses to do, because refusal doesn't screenshot well.

Which is exactly why refusal is the better thing to shop for.

Capability got cheap

A couple of years ago, an inline AI assistant was a differentiator. Now it's a checkbox in a starter template. The features that used to signal ambition — the chatbot, the integrations, the real-time collaboration — now mostly signal that a team shipped what everyone else shipped.

That's not a knock on features. It's an observation about information. When everyone can do everything, what a company chooses not to do becomes the most legible thing about them.

An anti-feature isn't a missing feature

Missing features are gaps. Anti-features are decisions to leave the gap alone on purpose. A notes app with no cloud sync by default. A budgeting tool that won't link your bank. A writing app that won't send your drafts to a model. A password manager that won't add a browser extension that phones home.

The distinction matters because intent is the thing you're evaluating. Ask "why doesn't this sync?" and listen to the answer. "We haven't built it yet" is a roadmap reply. "Because sync is a liability we don't want to own for this kind of work" is a product reply, and it tells you the team has an actual theory of the person using it.

The five-minute script

Four questions, asked inside the trial rather than on the sales call. None of them require a technical background.

What's the default? Not "can I opt out" — what happens if I do absolutely nothing. Defaults are the product. Everything below them is the escape hatch.

What breaks when the network dies? Turn off Wi-Fi mid-task and watch. Useful breakage tells you what runs locally. Everything going grey tells you the app is a browser tab wearing a trench coat.

How do I get my stuff out? Export to CSV, Markdown, or JSON during the trial, not someday. If export takes a support ticket, you've found the real feature: retention by inconvenience.

How do I cancel? Open the billing page and look for the button. If it's a phone call during business hours, they've told you which side they're on.

Then one quiet observation: open the settings screen and count the toggles labeled personalization. Every one of them is a place your data flows that you didn't ask for.

Red flags, in plain English

• "We may share information with trusted partners." A sentence where no noun does any work.
• Training on your content by default, with the opt-out buried three menus deep.
• No export option beyond a PDF.
• A "Do Not Sell or Share My Personal Information" link that opens a form instead of performing an action.
• Mandatory arbitration with a class-action waiver on a six-dollar-a-month product.
• A free trial that requires a card, renews at full price, and sends no reminder before the charge.

US buyers should scan the footer first. If the app sells to Californians, the CPRA language should be there and it should actually work. If there's nothing at all, that's a signal about how much thought went into the parts of the product you can't click on. The same goes for cancellation: with regulators paying closer attention to subscription traps than they did a few years back, a company that still makes leaving hard is making a choice about you.

Restraint costs money, and that's the point

A tool that refuses to sync has to be good enough to use without sync. A tool that won't run your notes through a model has to earn its keep with editing, speed, and keyboard shortcuts. Restraint only reads as premium when the thing underneath is excellent. Otherwise it's an excuse with good copywriting.

So the test isn't "does this app have fewer features." It's "does this app have fewer features and still feel finished." That combination is rare, and worth paying for, because it usually means someone said no to something a competitor said yes to — and knew why.

The fair objection

Sometimes you genuinely want the sync, the assistant, the integrations, and you'd rather have them than a philosophy. Fine. Nobody is arguing you should pick a worse tool to feel principled.

The argument is about defaults and disclosure. Ship the feature. Just let me see what it costs me in access, in data, and in the difficulty of walking away. Consent that arrives pre-checked isn't consent, it's formatting.

The apps that get this right are easy to spot once you start looking. They volunteer what they don't do before you ask. They make export a button instead of a favor. They cancel you in two clicks without a survey asking why.

Run it before your next renewal

Before the next charge hits your card, spend five minutes doing the opposite of a demo: pull the plug, try to export, try to leave. Whatever survives that is the product you're actually buying.

The tools I've kept longest have all been a little stubborn. Not because stubbornness is a virtue on its own, but because it's evidence that somebody made a decision on my behalf and got it right.