The quiet part of winding something down

There's a genre of shutdown post where something explodes. A cofounder leaves, a lawsuit lands, the money runs out on a specific Tuesday and there's a meeting where everyone finds out.

That's not what most endings look like and it isn't what this one looks like. Ours is quieter and I think the quiet version is more common and less written about, so here it is.

What it actually feels like#

You have a company that works. Not spectacularly. It has revenue, it has customers who'd be annoyed if it disappeared, and it does a real thing for real people. By most reasonable definitions it's a success — three years, never raised, still here.

And you notice that you're managing it rather than building it. The interesting problems got solved in year two. What's left is execution against a ceiling you can now see clearly, and the ceiling is lower than the amount of life you were planning to spend under it.

Nothing is wrong. That's the disorienting part. There's no crisis to respond to, which means there's no moment that forces the decision, which means the decision can be deferred indefinitely and usually is. Companies in this state can run for a decade.

What I got wrong along the way#

I want to be specific rather than reflective, because the reflective version is worthless to anyone.

I aimed at the top of the market for status reasons. The recognizable names were the ones I wanted on the site, and they're the ones who most predictably outgrow a marketplace. The durable business was in the long tail and I found that out late, largely because the long tail is less fun to talk about.

I let manual matching hide the product's failures for two years. Every introduction I made by hand was a search bug I absorbed instead of filed. The software never got the pressure it needed because I kept relieving it.

I treated "we can't raise" as permanent after it stopped being true. The category became fundable around 2014 and I was still operating on a 2013 assessment. That wasn't discipline. It was not re-checking.

I was early, and I priced early as an advantage rather than a cost. Being first in a category with no name means you pay for the education of the entire market and then someone with a real budget shows up once you've finished paying.

The thing about being early#

Three years is a long time to be early. Long enough that you stop being able to tell the difference between a market that isn't ready and a company that isn't working.

Influencer marketing is going to be enormous. I'm more sure of that now than I was in 2013, and the evidence is that money and named competitors have finally arrived. I built the right thing. I built it with the wrong balance sheet, at the wrong scale, aimed at the wrong end of the market, and about two years before the demand caught up.

Being right early and being wrong are indistinguishable from the inside, and they're also indistinguishable in the bank account, which is the only place it matters.

What I'd tell someone in the quiet version#

The absence of a crisis is not information. If you're waiting for a forcing function, understand that a company like this will never produce one, and that the cost of waiting is measured in years of your attention rather than dollars.

Ask the question directly instead: if I were starting today, with what I now know, would I start this? Not "is this good" — plenty of things are good. Would I choose it again over everything else I could do with the next three years.

I've been avoiding that question for a while by staying busy, which is the most effective avoidance technique available to anyone running a company.