Skip to content

Don't Build a Mine Before You Struck Gold

Flo Crivello
Flo Crivello
6 min read
Drawing of a young Asian man wearing a dark hoodie, scratching his face with his left hand, looking puzzled. Behind him, there's a massive mining operation with large machinery, excavators, and dump trucks moving earth.
I drew this myself, obviously

The best analogy I’ve heard for startups is that they’re like looking for gold.

Not because of the adventure, or the camaraderie, or the riches awaiting you on the other end.

But because a gold-seeking expedition has two very different phases: first, you look for gold; then, you build a mine. And startups are really in the business of doing the former — Steve Blank even goes as far as to define a startup as a company that is in that stage: once it's found gold (or a "repeatable business model"), it turns into just a business.

The analogy yields a few powerful lessons:

  1. Market is king. If you're on a site with no gold, it doesn't matter how good of a digger you are, or how "persistent" you can be — you're wasting your time.

    Conversely, it didn't take a genius to strike gold during the gold rush in California in the 1850s.

    That's certainly been my experience over the years: I've seen folks I thought were brilliant waste their time chasing a market that doesn't exist; and many a "clown car crashing into a goldmine."
  2. Know whether you’re digging old or new ground. Old ground is ground that everyone has been able to dig for a while.

    New ground is land that just opened — there was a tech shift (mobile, cloud, AI…), or a new segment appearing in the market with more buying influence (developer tools), or new regulation coming into effect, etc.

    It’s generally better to dig new ground — you’ll have no incumbent to beat. That’s why VCs ask “why now?”

    But that’s not to say that startups can’t be built on old ground. It’s just that the playbook is different — mainly, speed to market matters a lot less there, and quality a lot more, since you’re already super late anyway.
  3. Do not build a mine before you struck gold. Everyone is just so eager to build a mine — grow their team, set up processes, define OKRs, etc… — because 1/ they feel like that's what successful businesses do 2/ it’s a clear and easy task, whereas looking for product / market fit is hard and ambiguous.

    This is immensely important, and some research like the Startup Genome Project even finds that this kind of “premature optimization” is the #1 cause of startup death. Three big problems with building a mine on the wrong site:
    1. Operating a mine is a full-time job. You spend your days on operations, project syncs, and all the random drama that comes with a couple dozen humans being in the same room.

      These can feel like progress, but are actually a huge waste of time, energy, and attention.

      And if you spend your time on these things, you won't have any left to actually look for gold, which, unfortunately, only you as a founder can do. Researching markets, talking to prospects, wrapping your head around the problem space — all that is a full-time job too, and some more.
    2. The mine will anchor your search space. Even if you somehow find the time to look for gold, with a mine on the side, you will still be psychologically anchored to your previous idea, and mistakenly search in its surroundings.

      Elad Gil warns about this:

      "[Changing] direction in a market you are already in […] is the most common type of pivot, and the most likely to end badly. [F]ounders worry too much about sunk cost and the industry knowledge they have built. […] By staying the same bad market, the company may be doomed despite the pivot."

      Naval Ravikant has described this as being in a helicopter, looking down into a dense, infinite jungle. Once you find a spot you like, you land the helicopter and start digging. But if you don't find anything, you don't start roaming around the jungle: you climb back into the helicopter!
    3. You can't really move a mine.

      If you hired a team of 12 diggers after landing, you’re gonna have a hard time fitting everyone into the helicopter. Even if they all fit, this kind of move is so painful that you’ll resist the conclusion that there’s no gold here for much too long.

What To Do When You Built a Mine Over The Wrong Spot

You done did it. You raised a lot of money and hired way too much and now are realizing that there's no market here. What do you do?

I'm actually painfully familiar with this pretty awful position. I started Teamflow in March 2020, as the pandemic was just starting. The reasoning was simple:

  • Every company was forced into remote overnight;

  • Remote has a lot going for it (no commute, can hire anyone, etc.), but also a ton of downsides (makes it hard to work together) 1

  • Wouldn't it be neat if we could have all the upside and none of the downside?

  • Let's build a virtual office for remote teams to work together!

Things went very well when everyone went remote — we quickly got close to $1M ARR — and very poorly when people returned to the office. Curves that had been going up and to the right were now flat at best, even as the product improved.

At first, I resisted the conclusion that the market had vanished.2

Elad Gil is the one who ultimately convinced me, pointing out: "it's not just you — you're in a category that's failed."3 He was right. We'd been heads down digging, trying to get the growth to restart. But as I perked my head up and looked at the 50+ competitors in the surroundings, I saw no one pull ahead.

These were good, talented teams, with lots of funding. And there were a lot of them. None had made it: we were all digging the wrong ground.

In the end, we were lucky that this coincided with the greatest tech revolution in decades: the rise of AI. We'd played with GPT-3, built a "meeting recording bot," and quickly realized that we could get LLMs to not only generate text for us, but to take action, becoming full-fledged agents – before there was much talk about agents, really. That led us to Lindy, which we pivoted into.

This is the short, pretty version — reality was obviously infinitely more messy and painful, and the adventure is ongoing. As we pivoted and went back to R&D mode, I had to let go of more than two thirds of the team. I felt awful.

But "experience is what you get when you don't get what you wanted"4 and I sure got plenty of that. Here's how I'd summarize my advice for anyone who built a mine before striking gold:

  1. Don't fuck around.

    If you've concluded that there is no there there, you have to make big, bold moves, and make them fast. In hindsight, I delayed the inevitable for easily 6 months too long.5

    One big, stupid driver for waiting was not wanting to destroy the team's morale. I now realize that morale doesn’t matter — winning does. Your job as CEO isn't to make people happy, but to build a successful company.

    Jim Collins urges people to "face the brutal facts, yet never lose faith." It's important to keep things in this order.
  2. Cut to the bone.

    There isn't room for a lot of people in the helicopter — cut to a skeleton crew of 4-6 people.

    Do it all at once — the worst way to cut a rat’s tail is piecemeal. If you have a doubt on whether someone should go, it means they should (obviously, offer months of severance, and spend days tapping your network to help these people find new jobs — you’re the one who screwed up).
  3. Neglect the mine.

    Once you’ve cut, aim to deliberately neglect even the little that's left.

    That's very hard for founders to do, because they tend to be hyper-conscientious over-achievers. No one wants to do a bad job, especially not them.

    But as someone who built a mine before striking gold, you have to fire yourself as manager, and resume working full-time as a gold prospector. The mine will be mismanaged — but the mine doesn't matter.

    In my case, that meant completely clearing my calendar once I'd decided we were working on agents. I spent the time hacking on prototypes with the team, reading papers, and researching both the technology and the market.

This all sucks, but there’s a way through — some of the best companies, from Slack to Instagram to Twitter, are born from pivots. If it sounds painful, it’s because it is — and the best way to avoid the pain is not to build a mine before you’re certain you've struck gold.

  1. I've since concluded that the downsides vastly outweigh the advantages, especially for startup that are pre-product / market fit.
  2. Covid generally sent a lot of false signals to a lot of companies — Clubhouse is probably the most famous example — which led to quite the hungover in subsequent years.
  3. Huge thanks to Elad for the candid advice that probably saved me 1-2 years.
  4. Randy Pausch
  5. This seems to be the case of all big decisions, for everyone.