On the perils of mistaking a launch for an accomplishment
Why is it that so many more people remember the landing of the Apollo 11 mission on the moon than its take off? The Apollo launch burned 2 million liters of kerosene and liquid oxygen, generating 7.5 million pounds of thrust in a truly spectacular manner, but the landing of the Eagle module on the moon was the far bigger accomplishment.
Like the Apollo 11 launch, launching a product is a clear milestone that takes a huge amount of work and coordination by multiple teams. Think of some of our memorable launches — launching Gmail on April Fools’ Day, the Chrome comicbook, and Psychicsearch — these launches were natural moments of celebration.
But the catharsis of that moment can blur our perspective and lead us to miss a fundamental point: a launch is not itself a meaningful goal. It didn't matter that we launched these products successfully -- what actually mattered was that they were successful products that thrilled our users over time. Meaningful goals demonstrate material, positive impact on the utility of our products, the health of our business and happiness of our users. An overemphasis on launches incentivizes unripe products, fragmentation of efforts (too many arrows), and undermines the obsession to continuous improvement that is essential to truly excellent products.
We propose switching the language used for describing achievements from launches to landings. Landings occur once a measurable notion of success has been achieved -- happier users, delighted customers and partners, more efficient and robust systems. Landings become what we work for, what we celebrate, what we reward.
Landings, unlike launches, are not self-evident and require explicit definition and discussion. Do we measure success by growth in daily active users? Retention? Reduced tail latency? User satisfaction metrics? Profitability? How do these metrics differ across products (search has a different success metric than apps than platforms…)? Although it is not always obvious what the landing metric should be, forcing those decisions to be made early is essential to success. It is far more important to know "what does it mean to land" than "what does it take to launch".
Focusing on landings also makes it easier to have productive discussions about launch timing. Photos decided to postpone their launch by 6 months until they felt confident they had the right polish for a successful product. That was the right call: Delaying a launch can speed up landing.
Importantly, shifting our culture towards landings lets us highlight accomplishments that don't even involve introducing anything new. When the landing criterion is clear, it allows us to properly reward our individuals and teams who do work that improve existing products significantly, such as engineers that anticipate problems and actually simplify the solution during code reviews or teams that focus on critical systems, like Blaze, GWS or the Chrome updater, where landing means making important improvements (including reducing complexity) to existing systems. You can land without ever launching!