What is the most dangerous thing about running a company alone?

Not the lack of money. Not the lack of staff.

It is that nobody tells you that you are wrong.

When Jason Zigelbaum built Zigpoll, there was no co-founder, no sales team, and no funding. Roadmap, priorities, pricing—almost every call could be his alone. That sounds like freedom. The other side of freedom: a bad idea can live a long time with no one to kill it.

In an Indie Hackers interview he said a line that later traveled farther than the revenue figure. After the piece ran, he noted on LinkedIn that what people messaged him about most was not the roughly $125K MRR—it was this:

You are not the smartest person in your own business. Your customers are, collectively.

It sounds like startup broth. Read Zigpoll’s early years backward, and it is not soft at all—because what he first built was not what customers later paid for.

This is not a teardown of how a Shopify survey tool climbs to six-figure MRR. It follows how he stopped trusting “I know what to build.”

At first he only wanted websites to be more fun

Around 2018 Jason started Zigpoll with a simple itch: websites were boring. Why not a light poll like Twitter or LinkedIn—what’s your favorite color today? One tap, a little interaction, a little feedback for the merchant.

It demoed well. A clean modal, a vote card, done. For someone technical who had lived inside ecommerce, the direction felt natural.

The market answered bluntly: people willing to play are not the same as people willing to pay.

On the Honest Ecommerce podcast he admitted the idea was “cool” and still produced little real commercial value. The problem was not zero clicks. After the click, merchants still did not get an answer that mattered enough.

That is where Zigpoll’s first real turn began.

He wired the tool into real work; answers surfaced themselves

He did not sit inventing the next poll feature. He put the product into merchants’ feedback flows and watched what people asked and filled in.

The answers converged. Merchants did not need “which color do shoppers like?” They needed: Why did you buy? Why didn’t you? Why did you leave?

Interaction and those questions are different businesses. One is noise; the other sits next to sales, conversion, and retention. The product drifted toward post-purchase, exit intent, and conversion-reason prompts. “Make the site more fun” became “help merchants find Why.”

Not a lightning positioning flash. More like being pushed, step by step, by users. It became the method he repeats: don’t start by believing your judgment and hunting proof; let users speak, then see if you will admit they are right.

Seeing the direction is still far from traction

Zigpoll did not succeed on day one. Startup Founder Stories’ write-up matches the public arc: within about a week on the Shopify App Store there were installs and paying users. Crop that clip and you get the familiar story—ship, list, grow.

Jason’s version is less smooth. Real traction took about two years.

A first sale proves someone will try. Traction asks why a cohort stays. He watched who expanded usage, who referred, which requests repeated—and found that revenue moved less with features he found clever than with what customers asked for again and again.

Customers together are smarter than you not because each shopper out-knows the founder, but because dozens or hundreds of daily behaviors, summed, beat one person forecasting demand at a desk.

The danger is not being wrong; it is nobody stopping you

That bites hardest for a solo operator. With a co-founder, someone can still say “this direction feels wrong.” Alone, that mechanism disappears. You can believe the roadmap for three months, ship ten features you love, and explain non-purchase as “they just don’t get how good this is yet.” Nobody pulls you back.

So the habit he formed is heavier than “listen more.” It is: go looking for evidence that overturns you. If ten customers repeat a request, stop clinging to your roadmap. If they use what you didn’t plan, admit the plan may be wrong. If one segment clearly pays more, don’t refuse to narrow because “the market should be bigger.”

For one person, that is not a humility pose. It is cost control. You cannot staff ten directions. The real question is not “what else can I build?” It is “what have customers already voted for with behavior?”

The right segment sat in the data for about a year

He later mentioned on LinkedIn that the right segment for Zigpoll had been in signup data for roughly a year before he saw it.

Founders often treat data as outcomes—signups, revenue, traffic. The sharper signal lives in differences between users: who goes deeper, who stays, who refers, who asks again, who pays for more. Average everyone together and the signal vanishes. Split them and you may find the people who love the product are not “all ecommerce merchants,” but one kind of them.

Serving everyone collapses into a market one person can actually serve. That is not making the business small. For a solo founder, focus is often the precondition for scale.

What he changed was not the feature list; it was his place in the product

Looking back, the interesting part is not only what Zigpoll became. It is how Jason relocated “how smart a founder must be.”

First he thought sites should be more fun, so he built polls; the market did not pay. Then the tool entered real workflows and customers kept asking why they bought or left; the product followed. Later, what moved revenue was what users demanded, used, and spread—his job shifted from “what do I think we should build?” to “what have users already said is worth building?”

A few words of difference. For a solo company, it can be two years and a pile of expensive guesses.

So “customers are smarter” is not modest decoration, and it does not mean “stop thinking.” It means the opposite: one of a founder’s hardest skills is still letting facts kill an idea you love.

No co-founder to brake you. No sales team narrating why deals die. No PM forcing another validation pass. What you have is putting your judgment in front of real users—and accepting a result that may sting: you were not as smart as you imagined.

Once you accept that, the answers customers leave behind can finally become the roadmap.