Are you spiraling?

↓

Joy sparked in my web dev heart as I played with the collapsable tree of XML inside my Firefox browser window.

Every node was there! And the forecasted latitude and longitude of the storm that would become Hurricane Frances had updated within minutes of the official change published to the web by the National Hurricane Center on their own RSS feeds.

Within a few more minutes, that same information would flow through all of the standard media channels. As their newscasters narrated, millions of viewers would take a few steps towards their glowing TV's to try to understand exactly what this meant for them. Was the storm closer now? By how much? Had the path shifted south again? Millions of furrowed brows would erupt in unison, not quite sure what it meant for them.

My ugly little tree was spring-loaded with answers. Every one of those concerned viewers had a location. Using that location and the data in the feed, one of those servers from Amazon's new Web Services offering could instantly calculate a distance-to-the-storm for every visitor, at scale, delivering insights faster and more precisely than they could get it from their news anchor. I could also show them how the forecast had shifted, taking advantage of the new graphical capabilities of web browsers.

It would be 2 years before the app I was building could do all that. And 4 years before we'd need scale.

But that first test? It was a heist. A cheap, I-can't-believe-they-left-the-safe-unlocked, web-crawling, friend's-server-borrowing heist.

I just had to take the data and run.

Prometheus Meets World

Kayak had launched earlier in 2004. Their heist: "every flight option is right here, just filter!"

A few years later, I'd hear about Dropbox's: "just click save and watch the changes appear over there!"

Eventually, our site would add weather alerting, and I'd learn about Twilio's: "just run that line of code and a text message will appear on your phone! Try it!"

These magic tricks were the delightfully easy-to-understand demonstrations of defensible, value-creating cores. Someone, somewhere had had a fire-stealing moment. An oh-my-Zeus, this-is-now-possible moment.

After that, their mission became clear: take it to the world.

Not So Fast

Many startups never have a fire-stealing moment, but are very good at convincing themselves, and investors, that they do. The sooner these journeys come to an end, the better.

A lot of honest founders do seize something, then elect to Leroy Jenkins themselves and their Olympic torch straight into the bear-infested woods. Must go faster?

My approach with both startups has been a spiral. It takes a lot longer, and I've never perfectly executed it, but the main idea is to take that spark, start a campfire, run a perimeter, and see if you find anything promising nearby.

My first campfire with Stormpulse was the XML feed. I shared it with friends, saw their reactions, and knew that although I could probably sell this to someone, it wasn't going to be worth my effort. Yes, someone could probably have turned that into a pure-play data business, but it wasn't going to be me. Most of the people I knew needed a UI.

So I put that fire out, kept what I had learned, packed up, and started a new fire nearby. How about an interactive weather map, powered by that data? I scanned my horizon and the reaction was much more promising; but what I had built wasn't for them 🐙. I had fallen so in love with the product I had lost myself a bit. It was too complex.

So we put that fire out and started one last one (wood is expensive!). What if it was still beautiful, but eye-rollingly simple? And what if you didn't have to remember to visit our website, but could instead access it through the site you already visit?

This one worked.

Success is the Slow Path, Usually

There's tension in this approach, which is whether you're trying to optimize for success or speed to success.

This sounds strange, but it explains how "fail fast!" is commonly interpreted. Some people simply do not want to spend years to find the right combination of location (market) and fire (value). Believing there must be an idea that will get them to success sooner, they abort the mission entirely and re-spawn elsewhere.

My interpretation of "fail fast" is to be willing to put out the fire that isn't working — the small flame that warms you, but isn't ever going to be worth it, then take what you learned, and start a new one nearby, building off what you've learned.

Yes, it's hard to stomp out your progress and pack up camp. Outwardly, it looks like you really don't know what you're doing ("is this a pivot?", "why can't you be like Leroy?!"), and admitting you were wrong: "we're closer, but this isn't the site we're looking for."

Inwardly, you know you know the market better than you did yesterday. Abandoning the mission and starting over with an entirely new idea? Sure, maybe you'd get lucky; maybe wouldn't have to keep doing this. Maybe the first iteration would be it.

But maybe you've decided living to get lucky isn't how you want to live.