Claim
Healthy growth portfolios mix many small lead-bullet experiments (button copy, placement, friction reductions) with rare cannonball bets (signup with phone numbers, SMS delivery, fundamental product changes). The trap is that lead bullets are easier to design and ship, so teams default to them and never make the cannonball bets that actually compound.
Mechanism
Lead bullets feel productive because they ship fast and produce a steady stream of small wins. Cannonballs require strategy work, cross-team coordination, and longer runway, all of which look like cost in a weekly metrics review. The org's incentive structure rewards lead-bullet frequency over cannonball impact, and the team gradually optimises for the wrong thing. Without explicit allocation discipline (e.g., 80% lead bullets, 20% cannonballs), cannonballs disappear.
Conditions
Holds when:
- The team can identify which bets are cannonballs versus which are lead bullets.
- Leadership backs the cannonball runway against weekly-metric pressure.
- The team has the strategic capacity to design cannonballs, not just incremental tests.
Fails when:
- The team treats every project as a cannonball and ships nothing.
- "Cannonball" becomes a label for vanity bets with no measurement plan.
- The category is genuinely mature and lead bullets are where the wins remain.
Evidence
"There's a laziness that can creep in where you're just finding a lot of little things because they're easier to come up with."
Adriel cites cannonballs at Facebook: signup with phone numbers, SMS delivery, friend recommendations — each one fundamental product change with multi-quarter payback. Lead bullets cluster around copy, placement, funnel friction.
— Adriel Frederick on Lenny's Podcast, 2026-04-28
Signals
- Quarterly portfolio reviews show explicit cannonball-versus-lead-bullet allocation.
- Cannonballs ship at least once or twice a year and produce step-change results.
- Lead bullets become healthier because they sit downstream of the cannonball bets, not in lieu of them.
Counter-evidence
For very early products, almost everything is a cannonball — there is no funnel yet to optimise. For deeply mature categories, almost everything is a lead bullet. The discipline is most useful in the scale-up zone where both shapes of bet are available.
Cross-references
- Design for the marginal user — the person on the cusp of converting in the worst conditions — what to research before designing a cannonball