1. Validate before you build
The cheapest mistake you can avoid is building a product nobody asked for. Spend two weeks on validation before you write a line of code. Yes, two whole weeks.
The 3 questions every idea must pass
- Is the pain frequent? Daily > weekly > monthly. Annual problems aren't products.
- Is it expensive? Time, money, or both. Free problems = free products.
- Are people already hacking around it? If they're not, they don't care.
2. Build the smallest possible MVP
Your MVP should embarrass you. If it doesn't, you over-built. Ship the one workflow that delivers the core outcome, nothing else.
3. Positioning beats features
Two AI products with identical features will get wildly different traction based on how they're positioned. Positioning is the single most important thing to get right.
The positioning formula
For [specific audience] who [specific pain], [product] is the [category] that [unique outcome], unlike [the obvious alternative].
4. Launch the right way
Launch isn't a single event, it's a sequence. Spread across 7–10 days. Optimize for feedback in week 1, conversions in week 2.
- Day 1: LaunchLoop + Show HN, feedback first.
- Day 2: Product Hunt, PR moment.
- Day 3–4: Niche communities (Reddit, Slack, Discord).
- Day 5–7: Personal channels + DM round.
- Day 8–10: Recap thread + post-mortem blog post for SEO.
5. Compound momentum after launch
The launch isn't the goal, momentum is. Founders who win pick one channel and compound on it for 6+ months before adding a second.
- Ship a public update weekly. Visibility ≠ noise; consistency = trust.
- Relaunch on LaunchLoop every 4–6 weeks with a real new wedge.
- Talk to 5 users a week, every week. Forever. No exceptions.
- Write one deep SEO article per month. Year 1 = SEO foundation.