Raise your hand if you remember the #FailWhale.
If you’re like most people I talk to, it may have taken you a moment, but now you’re waxing nostalgic about the early days of Twitter, when the illustration of a whale in a net suspended by Twitter birds appeared almost daily, indicating the site was down. You might even be googling the many, many articles churned out by the press, listing the number of minutes the so-called Fail Whale had appeared that day/week/month. Twitter falling over was national news and, let’s be clear, it was a MAJOR problem for MONTHS. Twitter is a social network designed to provide real-time communication, and every time the Fail Whale appeared, the service was failing its primary function.
As a founder or product owner, the thought of Fail Whale may also give you anxiety. You might be pulling up your web host to add some more bandwidth as we speak. But please let me remind you of something incredibly important. Twitter is now a public company that earned $3 billion in revenue last year. So before you spend hundreds of thousands of dollars to redesign your website in preparation of someday having so many users that it might crash, knock it the fuck off. Right. Fucking. Now.
Which brings me to the straight-up, number-one phrase I repeat over and over to product owners. The number one killer of startups is solving problems that aren’t problems yet. And, because you need to hear it over and over again,
The Number One Killer of Startups is Solving Problems That Aren’t Problems Yet.
Back when I was running Cloudspace, we frequently fielded calls from companies asking us to help them rebuild their entire website. When we’d ask why, they’d tell us how they had just raised some capital and wanted a new, scalable website to, and I quote, “be ready for a lot of users.” It was always about managing that flood of new users. We’d tell them that AWS is insanely cheap, and they should just throw EC2 instances at the problem. If they ever got to the point where Amazon could no longer support them we’d fly out with champagne to celebrate their site falling over.
Inevitably, they’d take their business to another dev shop who would happily build that shit, solving a problem that wasn’t really a problem and leaving them with $300k less for customer acquisition and six months shorter runway. Few of those companies are still around, btw.
There are only so many things you can work on at any given time. Scarce resources is one of the defining challenges in business. So when you focus on issues that aren’t real, you ignore problems that ARE real, and are killing your business right now. If you don’t solve your current problems, you’ll never need to solve the hypothetical ones. Because your company will be dead.
Why the hell would you worry about your site’s scalability if you don’t know how to effectively acquire customers? Your site falling over someday MIGHT kill your company (probably not). Never growing past eight customers WILL kill your company. Instead of using your beautiful brain and precious funding to solve theoretical scaling issues, perhaps you should go run some customer acquisition experiments. Build a fucking landing page. Buy some AdWords.
This isn’t just a technology issue. Solving problems that aren’t yet problems comes in all colors. We launched OneTrueFan back in the 2010 San Francisco real estate bubble and leased an office for 30 people because we were worried no space would be available when we raised our Series A. Do you know how expensive 8,000 square feet is in the Mission? Do you know how silly I felt when we crash landed the business never hosting more than seven people in the office?
To be clear, I’m not talking about basic functionality. If you’re a bank and you discover a way for someone to hack your site and steal all the money, fix that. Like, right now. BUT, if you run a bank and you’re about to design a process to handle maybe someday losing all the money? Try just not losing their money.
Still don’t believe me? I have science on my side.
The Startup Genome Project researched 3,200 startups in 2011 and found that while they could not predict with any accuracy what companies would be successful, they could absolutely predict which companies would fail — and those were the companies that prematurely optimized. Not only could they predict those companies would fail, they found that 100 percent of them would never surpass 100,000 users, 93% of them would never make more than $100,000 a month, and also showed that companies that did not prematurely scale grew 20 times faster than those that did. Here’s a handy chart. Print that shit and hang it on your wall.
Founders need to stop looking at every potential failure scenario as actual failure, and instead embrace them as proof that they’re moving in the right direction. As I’ve said before, a successful startup isn’t easier, it’s just different hard. You’re always solving problems.
To return to the start, none of Twitter’s current problems have anything to do with the Fail Whale. (For anyone interested, they officially retired the illustration in 2013 — a full five years after it first appeared.) If they had focused on scalability in the early years, who knows if the company would even be around today. Luckily for them, they didn’t. And they are.
— Eric Marcoullier
Have questions about what to prioritize at your company? My name is Eric Marcoullier, and when I’m not publishing blog posts, I coach executives on product and corporate strategy. Figuring out what problems to leave for tomorrow is just one of the many ways I help my clients achieve greater success. If you’d like to chat for a (complimentary) hour and work through some of the questions you’re facing, drop me a note at email@example.com or visit my coaching site at Marcoullier.com.