A lot of people, including me, talk about when to launch an MVP. Let’s talk about when we shouldn’t.
I know startup is mostly about moving fast and breaking stuff. The drive to make quick decisions and take big risks is what allows companies to grow in leaps and bounds instead of just incrementally.
But I don’t have to tell you that when you move too fast or break too much stuff, the cops show up. Metaphorically. Chain reactions from one big mistake can take down great ideas, strong teams, and solid companies.
Now, a minimum viable product (MVP) is the moving-fastest and stuff-breaking-est version of a product that a company can ship to market. We should always be a little bit fearful when launching an MVP. But in some cases, that fear means we need to pull the plug and take a big step back.
I’ve got a ton of experience launching MVPs. I own my share of failures, frustrations, and stupid mistakes. So here are the top warning signs that send me back to the drawing board before I embarrass myself and everyone around me, listed in order of earliest to latest in the MVP process.
The process isn’t defined
Stop me if you’ve heard this one before. Oh, and stop your MVP launch if I’m describing yours.
You’ve defined the problem and the solution. You know the customer. You can visualize the tech for the final product and exactly how awesome it will be. You even have the tools in place to set up a manualized version of everything so you can test in a controlled environment.
For most of us, me included, what I just described above are all the ingredients necessary to throw an MVP out to the world.
Except sometimes for me there’s one nagging thing: I have no idea what to do with the customer once I land them. I can close them, I can onboard them, but I’m clueless as to what any of the next steps are once they’ve paid.
I’ve been in such a scenario a few times. And I’ve learned that it’s always a dealbreaker for an MVP.
To give a more concrete example: I have a friend who is a wizard with T-shirts. Every so often, maybe once a year, he gets a great idea for a T-shirt. And God love him, he designs it, sells it, gets tons of revenue and amazing feedback, and then… nothing. He makes a couple thousand bucks and continues his real life until the next great T-shirt idea hits him.
Moral: Make sure you’re not just building a profitable hobby. Before you launch, do a gut check to make sure you’re not setting yourself up to be overwhelmed by a ton of false positives that lead to nowhere.
The tech isn’t ready
I’m going through this now. I need tech for my MVP and I don’t know how much. This is like not knowing how much bungee cord you need for a jump.
In my case, a lot of the MVP can be done manually, but there’s just no way to limit the customer segment to a number low enough that I feel good about the risk of knocking over my manual portion. If that happens, I either have overstressed employees or unhappy (maybe irate) customers.
The former you can’t do lightly, the latter you just can’t do. Product suicide.
The problem is that there are way too many other priorities for me to just turn on the tech faucet. I can get it to drip, so I’ve been designing patches to get that fallover risk down. But the relationship between my patch and the risk of failure is just not quantitative enough for me to feel confident that my design will result in X% less risk. And since I can’t shrink the customer segment any further, going over X% means shutting the whole thing down.
Moral: Until we can properly build the tech to limit our risk of total failure, whether that tech limits the test, automates more of the manual parts, or injects flexibility and modularity into the test, we’re not in a position to launch the MVP.
The product is fundamentally flawed
Early in my career I had my world blown up twice because of an MVP. In both cases, the MVP was a wholesale pivot from an existing successful product to a next-level iteration of that product. And in both cases, the mistake was building for a model that the technology couldn’t support.
In case #1, we had developed technology for a new way to create an existing consumer product. However, being an existing product and somewhat of a commodity, the market was walled off by a few incumbent players who had, as they often do, made the barrier for entry into the market prohibitively high.
I wanted to continue to develop our own nascent market, one where we were already the major player. That approach landed on deaf ears, and we launched our MVP into the traditional market. We got crushed because we couldn’t produce the same artificially high quality that the consumer had grown to accept in the traditional market. Three years later, the new market took off, and today, the traditional market is dying.
In case #2, the company wanted to follow a product trend that promised more automation, more self-service, and higher margins. Having built the existing product from the ground up, I knew instinctively that the tech behind the product would need to be entirely rewritten to accommodate the new model. But we ended up following the least-advisable move-fast-and-break-stuff strategy — “We’ll launch the MVP and figure out the tech as we adopt the customers.”
This rarely works. It’s essentially making promises your tech can’t keep.
Moral: You can launch an MVP if you don’t know whether or not your tech will scale, but you can’t launch an MVP if you don’t know whether or not your tech will work.
The testing isn’t complete
Marketplaces are almost completely digital now, whether we’re selling digital products or physical products or even services. Digital marketplaces allow us to experiment, but the markets themselves don’t often allow second chances.
To use a dumbed down example: If our product is purple and we realize early in the launch that our customer base would have preferred green, we need to be able to make the product green without stopping, recalling, recoloring, retesting, and relaunching.
Chances are we can’t do that without losing the market in the process. But thankfully, digital marketplaces do allow for segmentation, and we can slice and dice our launch audience down to a hand-picked, pilot-sized number. Then we test.
Moral: Our product testing needs to go beyond the standard internal, alpha, and beta testing, and into a public pilot. And that public pilot needs to test every aspect of the customer experience, from how they’re introduced to the product to how quickly they become a repeat customer.
The numbers aren’t there
One of the hardest things an entrepreneur can ever do is pull the plug on a good MVP. One of the easiest things an entrepreneur can do is keep a good MVP running. But to be honest, we can’t live with good. We need great.
Sometimes the entire buildup to the MVP launch can go right, even the pilot, and the results can show a lot of promise. But for some reason those same numbers hint at mediocrity, stagnation, or worst-case, amazing numbers that don’t translate into revenue.
For each startup in my career where I’ve had a qualified success or failure, I’ve had two more that weren’t either success or failure. They were duds, failed at or before liftoff, never saw the light of day. These weren’t just ideas jotted down on napkins, these were full-blown MVPs that I couldn’t push past that final step on the road to greatness.
Moral: We can’t launch if we don’t believe we will be an unqualified and massive success out on the market. And that’s a hard metric to measure, because by our very nature as entrepreneurs, we don’t take a project on unless we believe in it wholeheartedly from the beginning. It’s hard to deny that belief, to admit we were wrong, to give up on an idea. But if we don’t see greatness in the numbers, we need to pull the plug.
Live to fight another day
The silver lining on the cloud of halting an MVP launch is that it doesn’t mean the story is over. More than one of my successful startups came from reborn versions of MVPs that never got off the ground.
In some cases it might be a few weeks, in others it could be a few years, but opportunity never dies, it just needs to find the right context. Whether or not we go back to the drawing board, that’s always up to us.