Following the publication of my article on the Startup Uncertainty Principle, I had a lot of fascinating conversations with founders who have gone through the same journey and experienced the same impossible decisions early on. To recap, the idea of the Startup Uncertainty Principle is probably best summarized by this line from the original article: “Given that a startup doesn’t know which bet will pan out, there will always be a tradeoff between its desire to strengthen its commitment to any particular idea and its desire to preserve optionality and keep exploring new avenues.” In other words, there is always a choice to be made between doubling down on any particular idea or keeping doors open for more iteration and exploration.

Some of the stories and reflections and advice of the founders I spoke to are below. Others will continue sharing their perspectives in the coming days and weeks, so I anticipate a follow up to this post with more stories too.

Do you have a perspective to share? Reach out to me on Twitter or LinkedIn!

People often perceive successful startups as having a linear journey but in reality most experiment heavily along the way, especially in product, to expand their offering and better serve their audience. At Product Hunt we often had one or two ‘moonshots’ in development at any one time. Many of those experiments didn’t work, but that’s the default outcome when introducing something new. For this reason, we typically limited the scope of the MVP to something achievable within a few months with only one full-time engineer collaborating with me, a designer, and a community or sales leader. I’m a fan of parallelizing big swings and doubling down on projects that have the strongest market pull; however, it requires sound judgment in knowing when to push forward and when to move on.

In the first iteration of Aviary, I was completely in love with our high level premise: we would build the world’s first online collaborative creative platform where every piece of art could be built in the browser and remixed by strangers, with full attribution intact. It was really cool, but pretty damn niche and we struggled for years to gain product market fit. 4 years in and we were getting only 50,000 edits a day and it still felt like pushing a boulder uphill to get more users on the platform. With the release of the iPhone, I realized the next ten years of creativity was moving out of the browser and into people’s pockets… and we needed to completely rethink our collaborative tools strategy and refocus on enabling the next generation of creative mobile apps.

The board and I were hesitant. We’d raised lots of investor money around that original pitch. So what we decided to do was keep the core product alive for now, and shutter all ancillary projects to provide the resources for a rapid exploration of how we could approach the mobile market. After actually launching the product, we found instant product market fit: we passed 50,000 edits a day in less than a month… something that had taken us 4 years to accomplish with the previous product. With our stats spiking like a skateboard ramp, we could now safely focus all resources on this new direction.

My main takeaway was a hard lesson: a founder has to both love the business they start like it’s their baby; and be fully prepared to abandon that baby if there’s a better way to succeed. That’s a really, really hard lesson to learn. The silver lining for me has been founding my new company BurnerPage around that principle, by making it easy for founders to rapidly experiment and learn what works.

I like the way Donald Rumsfeld framed this problem: known-unknowns vs. unknown-unknowns. At our startup, known unknowns are things like pricing for a category-defining product; whether there will be a recession in the next 12–18 months affecting the macro environment both on revenue and funding side; is your TAM estimation model bonkers and built on sand (it is); what target feature set is optimal to achieve product-market fit, and how do we figure out if we got it wrong; and if we turn up the dial on burn rate because we are pretty sure we are the verge of breaking out and achieving product-market fit, is that going to be something we regret. Unknown-unknowns are things like startup competitors still in stealth mode; a key employee who is currently interviewing and will resign next week, and will we be able to convince her to stay once that comes to light; an investor thinking to pull out; a design partner who is about to exit the market but hasn’t let you know that yet; and more than anything else just the questions that you should be asking but you’re not, because you haven’t thought to worry about them.

Wei Boyang once said, “Worry is preposterous; we don’t know enough to worry.” But every startup CEO worries anyway, because the not knowing enough, the uncertainty, is screaming at you every day: the thing that is going to blow me up tomorrow is not even on my radar as a problem.

I think my learning with previous startups is that picking the right opportunity trumps everything else. So it’s worth going broad in the product discovery process early on. Even at the expense of forward momentum. I think most startup mistakes I’ve personally made have come down to building the wrong thing by settling too early.

With Arro, we’re trying to stay comfortable being uncomfortable. It can be frustrating at times as execution is often easier than exploration. We’re trusting the process though and we’re backing ourselves to find the right opportunity.

We’ve brought a software mindset into my current company, Square Roots. But it’s much more complex to contemplate when your “product” is not an app, but in this case a network of multi-million dollar indoor farms, where we’re growing living plants in highly controlled, food-safe and people-safe environments. In software if you get it wrong, you can roll it back. In food if you get it wrong, people can get sick. Which you absolutely cannot do. So there’s definitely a desire for more stability, consistency and following standard processes in farming. i.e. the less uncertainty the better! Nevertheless, within the farms, we’ve created a modular platform that enables a reasonable degree of experimentation day to day, and we’ve also been able to engineer workflows so that a single facility can support multiple channels to market. For example, from our newest farm just outside Chicago, we can grow, harvest and pack food to service restaurants, while at the same time grow, harvest and pack food to sell to grocery retail stores. (Chefs look for different characteristics than consumers — so the difference is big). With our modular platform, we can keep getting better to service one market, while experimenting quite happily as we contemplate the other with new ideas. For example, this could mean configuring environmental conditions in certain modules in our farm to grow plants with different morphologies, tastes or textures. Thinking about Nir’s phrase: “there’s a trade off between strengthening [a startup’s] commitment to any particular idea and its desire to preserve optionality and keep exploring new avenues,” we’re able to do both. But again, I think that’s because we’ve brought the “software mindset” into a completely different vertical.

We got a little lucky on this one in that product/market fit (of a sort) dropped into our laps. We are focused on products for older adults and when we started we had learned from our corporate jobs that existing participant recruitment solutions (which rely on online advertising to build their panels) were systematically biased in their sampling and have difficulty recruiting the right people at the pace of product development. So we built out our own network of research participants, which allowed us to learn and iterate quickly. When a former colleague reached out and offered to pay us to use our network and run a project for her older-adult focused product, we realized that we’d stumbled upon a solution to an important product development barrier. So we pivoted and started building out a research services business.

Every day at an early stage startup is a decision point. You only have a limited set of resources, be it money or team or time. And you have to decide how best to spend it. But early on, you don’t know how best to spend it. You have to make assumptions. Most will be wrong, like Nir mentioned in his article. But some will be right. And the magic of building a successful startup is recognizing those right decisions early, and doubling down on them. At Anchor, we had a core value around moving fast. It wasn’t just about being better than the competition. It was also around failing fast and learning fast.

Following the publication of my article on the Startup Uncertainty Principle, I had a lot of fascinating conversations with founders who have gone through the same journey and experienced the same impossible decisions early on. To recap, the idea of the Startup Uncertainty Principle is probably best summarized by this line from the original article: “Given that a startup doesn’t know which bet will pan out, there will always be a tradeoff between its desire to strengthen its commitment to any particular idea and its desire to preserve optionality and keep exploring new avenues.” In other words, there is always a choice to be made between doubling down on any particular idea or keeping doors open for more iteration and exploration.

Some of the stories and reflections and advice of the founders I spoke to are below. Others will continue sharing their perspectives in the coming days and weeks, so I anticipate a follow up to this post with more stories too.

Do you have a perspective to share? Reach out to me on Twitter or LinkedIn!

People often perceive successful startups as having a linear journey but in reality most experiment heavily along the way, especially in product, to expand their offering and better serve their audience. At Product Hunt we often had one or two ‘moonshots’ in development at any one time. Many of those experiments didn’t work, but that’s the default outcome when introducing something new. For this reason, we typically limited the scope of the MVP to something achievable within a few months with only one full-time engineer collaborating with me, a designer, and a community or sales leader. I’m a fan of parallelizing big swings and doubling down on projects that have the strongest market pull; however, it requires sound judgment in knowing when to push forward and when to move on.

In the first iteration of Aviary, I was completely in love with our high level premise: we would build the world’s first online collaborative creative platform where every piece of art could be built in the browser and remixed by strangers, with full attribution intact. It was really cool, but pretty damn niche and we struggled for years to gain product market fit. 4 years in and we were getting only 50,000 edits a day and it still felt like pushing a boulder uphill to get more users on the platform. With the release of the iPhone, I realized the next ten years of creativity was moving out of the browser and into people’s pockets… and we needed to completely rethink our collaborative tools strategy and refocus on enabling the next generation of creative mobile apps.

The board and I were hesitant. We’d raised lots of investor money around that original pitch. So what we decided to do was keep the core product alive for now, and shutter all ancillary projects to provide the resources for a rapid exploration of how we could approach the mobile market. After actually launching the product, we found instant product market fit: we passed 50,000 edits a day in less than a month… something that had taken us 4 years to accomplish with the previous product. With our stats spiking like a skateboard ramp, we could now safely focus all resources on this new direction.

My main takeaway was a hard lesson: a founder has to both love the business they start like it’s their baby; and be fully prepared to abandon that baby if there’s a better way to succeed. That’s a really, really hard lesson to learn. The silver lining for me has been founding my new company BurnerPage around that principle, by making it easy for founders to rapidly experiment and learn what works.

I like the way Donald Rumsfeld framed this problem: known-unknowns vs. unknown-unknowns. At our startup, known unknowns are things like pricing for a category-defining product; whether there will be a recession in the next 12–18 months affecting the macro environment both on revenue and funding side; is your TAM estimation model bonkers and built on sand (it is); what target feature set is optimal to achieve product-market fit, and how do we figure out if we got it wrong; and if we turn up the dial on burn rate because we are pretty sure we are the verge of breaking out and achieving product-market fit, is that going to be something we regret. Unknown-unknowns are things like startup competitors still in stealth mode; a key employee who is currently interviewing and will resign next week, and will we be able to convince her to stay once that comes to light; an investor thinking to pull out; a design partner who is about to exit the market but hasn’t let you know that yet; and more than anything else just the questions that you should be asking but you’re not, because you haven’t thought to worry about them.

Wei Boyang once said, “Worry is preposterous; we don’t know enough to worry.” But every startup CEO worries anyway, because the not knowing enough, the uncertainty, is screaming at you every day: the thing that is going to blow me up tomorrow is not even on my radar as a problem.

I think my learning with previous startups is that picking the right opportunity trumps everything else. So it’s worth going broad in the product discovery process early on. Even at the expense of forward momentum. I think most startup mistakes I’ve personally made have come down to building the wrong thing by settling too early.

With Arro, we’re trying to stay comfortable being uncomfortable. It can be frustrating at times as execution is often easier than exploration. We’re trusting the process though and we’re backing ourselves to find the right opportunity.

We’ve brought a software mindset into my current company, Square Roots. But it’s much more complex to contemplate when your “product” is not an app, but in this case a network of multi-million dollar indoor farms, where we’re growing living plants in highly controlled, food-safe and people-safe environments. In software if you get it wrong, you can roll it back. In food if you get it wrong, people can get sick. Which you absolutely cannot do. So there’s definitely a desire for more stability, consistency and following standard processes in farming. i.e. the less uncertainty the better! Nevertheless, within the farms, we’ve created a modular platform that enables a reasonable degree of experimentation day to day, and we’ve also been able to engineer workflows so that a single facility can support multiple channels to market. For example, from our newest farm just outside Chicago, we can grow, harvest and pack food to service restaurants, while at the same time grow, harvest and pack food to sell to grocery retail stores. (Chefs look for different characteristics than consumers — so the difference is big). With our modular platform, we can keep getting better to service one market, while experimenting quite happily as we contemplate the other with new ideas. For example, this could mean configuring environmental conditions in certain modules in our farm to grow plants with different morphologies, tastes or textures. Thinking about Nir’s phrase: “there’s a trade off between strengthening [a startup’s] commitment to any particular idea and its desire to preserve optionality and keep exploring new avenues,” we’re able to do both. But again, I think that’s because we’ve brought the “software mindset” into a completely different vertical.

We got a little lucky on this one in that product/market fit (of a sort) dropped into our laps. We are focused on products for older adults and when we started we had learned from our corporate jobs that existing participant recruitment solutions (which rely on online advertising to build their panels) were systematically biased in their sampling and have difficulty recruiting the right people at the pace of product development. So we built out our own network of research participants, which allowed us to learn and iterate quickly. When a former colleague reached out and offered to pay us to use our network and run a project for her older-adult focused product, we realized that we’d stumbled upon a solution to an important product development barrier. So we pivoted and started building out a research services business.

Every day at an early stage startup is a decision point. You only have a limited set of resources, be it money or team or time. And you have to decide how best to spend it. But early on, you don’t know how best to spend it. You have to make assumptions. Most will be wrong, like Nir mentioned in his article. But some will be right. And the magic of building a successful startup is recognizing those right decisions early, and doubling down on them. At Anchor, we had a core value around moving fast. It wasn’t just about being better than the competition. It was also around failing fast and learning fast.