// advice
Things we've learned building consumer products. Not rules. Not a framework. Just patterns we've seen hold up across multiple products, markets, and years. Take what's useful. Ignore the rest.
Ship before you're ready.

The version you're embarrassed by is the version that teaches you the most. Every week you spend polishing is a week you're not learning from users. The product in your head is a fantasy. The product in the market is data.

Measure one thing at a time.

Dashboards with 40 metrics are dashboards nobody reads. Pick the one metric that matters most right now. Make it visible. Make it impossible to ignore. When it moves, you'll know why. When you track everything, you learn nothing.

Talk to the users who leave, not the ones who stay.

Your power users will tell you what to add. Your churned users will tell you what's broken. The second group is more valuable. They represent the majority you're losing. A 15-minute call with someone who cancelled will teach you more than a month of analytics.

Make the default the right answer.

Settings are where products go to die. Every toggle is an admission that you didn't make a decision. The best products decide for the user, and they're right 80% of the time. The remaining 20% is the price of simplicity. Pay it.

Your competitors don't matter as much as you think.

You'll spend 100 hours thinking about competitors for every 1 hour a user spends comparing you. Users don't switch products because a competitor is slightly better. They switch because your product stopped being worth the effort. Focus on your own retention, not someone else's features.

Charge more.

You're almost certainly underpriced. The fear of charging too much is universal and almost always wrong. Users who pay more are more engaged, more retained, and more vocal advocates. Low prices attract low-intent users who churn at the first inconvenience. Test a price that makes you uncomfortable. You'll be surprised.

Kill features faster than you add them.

Every feature has a maintenance cost, a cognitive load on users, and a surface area for bugs. If less than 10% of users engage with a feature, remove it. The product gets simpler, the codebase gets lighter, and nobody emails you about it. The features you're most attached to are usually the ones users care about least.

Hire slow. Fire fast. But mostly, stay small.

Every person you add increases coordination overhead quadratically. A team of 3 that communicates well will outship a team of 12 that needs process to stay aligned. The right question isn't "who should we hire?" It's "can we avoid hiring?" AI tools make this more true every month.

Retention is the only metric that compounds.

Acquisition is a tap you turn on and off. Retention compounds. A product with 60% D30 retention and slow growth will eventually beat a product with 20% D30 retention and fast growth. Always. Math doesn't negotiate. Fix retention before you spend a single euro on acquisition.

The best time to ask for feedback is never. Watch instead.

Surveys lie. NPS lies. Users say they want feature X and then never use it. Instead: watch session recordings. Read support tickets. Look at where users rage-tap. The gap between what people say and what people do is where all the insights live. Behavior is the only honest feedback.

Write everything down.

If a decision isn't written down, it wasn't made. If a process is in someone's head, it's a liability. Write the decision, the reasoning, and the expected outcome. When the outcome differs, you have something to learn from. Verbal agreements are forgotten by Friday. Written decisions compound into institutional knowledge.

Don't optimize what you shouldn't be doing at all.

Before making something faster, ask if it should exist. Before automating a process, ask if the process is needed. The most productive thing you can do is delete. Delete the meeting, delete the feature, delete the report nobody reads. Efficiency applied to the wrong task is still waste.