$125k MRR, no team, no investors. Jason Zigelbaum runs Zigpoll completely alone.

You don’t need another AI tutorial. You need an opportunity worth acting on.

Most people are still figuring out what AI can do. Others are already using it to launch side hustles, build businesses, and create new income streams.

Get instant access to 200+ proven AI money-making ideas for 2026, plus the free daily newsletter trusted by 1M+ professionals.

Analytics can tell you what happened. The part they cannot tell you is why.

Jason Zigelbaum spent years on the agency side of e-commerce, watching brands invest heavily in ad spend and data tools and then guess at the part that mattered most - what was actually going on in the minds of the people not buying, abandoning carts, or churning. He found that gap irritating enough to build a solution for it.

That solution is Zigpoll, a survey and customer feedback platform built initially for e-commerce, and now used by SaaS teams as well. It captures customer sentiment at the exact moments when people will answer honestly - post-purchase, on exit, after delivery - and turns those responses into decisions rather than reports.

Jason runs it completely alone. No co-founder, no funding, no sales team. It took about two years to find traction. Since then it has doubled in revenue every year. He started 2026 at just over $1M ARR and is closing June at $125k MRR - roughly a $1.5M run rate, representing a 44% increase in the first six months of the year.

Here is how he built it, what he got wrong, and what running a seven-figure solo SaaS actually looks like day to day.

Most product ideas come from observation at a distance. Jason's came from watching the same problem play out across dozens of client businesses over several years.

His agency background meant he had already seen e-commerce brands hit the same wall repeatedly: strong analytics, visible numbers, and complete uncertainty about the human reasoning behind them. A cart abandonment rate could be measured to three decimal places. The reason behind it was almost always a guess.

Before Zigpoll, Jason built a couple of other SaaS products. Shopify acquired one of them - an app called Metafields Manager - and he sold the other. With those businesses providing a small but stable income and requiring only a few hours a week to maintain, he had the financial runway to build something new without taking outside money or splitting equity with anyone.

The first version of Zigpoll was built with a code editor and time - nights and weekends for months. He is technical, so the build cost was almost entirely the opportunity cost of hours not spent elsewhere. The original focus was not even post-purchase surveys. That use case emerged from early customer demand, and he followed it. The same pattern repeated for exit-intent surveys, CRO surveys, and order delivery feedback. The product was shaped by what customers actually needed rather than what he had originally imagined.

Zigpoll runs on a deliberately straightforward setup:

The philosophy behind it is consistent with how Jason runs the business generally: as a solo developer, every hour spent on infrastructure is an hour not spent on product. Leaning into reliable third-party tools wherever possible - and choosing the boring, well-supported option over the clever novel one - keeps maintenance overhead low enough that one person can manage it without things breaking in the night.

The most expensive mistake Jason made was not a technical failure or a missed market opportunity. It was a misunderstanding about who his best customer actually was.

For a long time, he pictured his core user as a single in-house e-commerce team. He priced and built for that person: standard SaaS packaging, with integrations and AI features gated to higher-tier plans. On paper, a reasonable approach.

In practice, it was quietly punishing the segment growing fastest inside his user base: agency operators running Zigpoll across ten or twenty client stores simultaneously. These users needed integrations not as an upgrade incentive, but as basic table stakes - every client store they managed already used the tools Zigpoll was gating. By placing those features behind a paywall, Jason was charging his most valuable customers extra for the things they needed most, and slowing down the behavior that was spreading the product organically.

He did not see it until he read the onboarding data closely. The signal had been there for months. When he fixed it - moving integrations to the standard plan and rebuilding the pricing structure around the agency operator segment deliberately - revenue per account climbed 24% without a single price increase.

The technical choices reflect the priorities of a small team moving fast with limited infrastructure overhead:

Simple, well-documented tools throughout. AI coding tools work best with standard frameworks, and keeping the stack orthodox meant faster iteration.

No single launch moment made the business. No viral post, no Product Hunt spike. Zigpoll grew through a small set of channels compounding quietly over time.

His summary of what works: find the channel where the platform handles distribution for you and become the best option on it. Then identify which customers refer you without prompting and build relentlessly for them. Word of mouth is the only growth channel that gets cheaper as you scale rather than more expensive.

The most sustained theme in how Jason talks about the business is listening to customers. He is worth quoting carefully on this, because his framing is more specific than the general advice suggests.

His first point: most founders confuse hearing with listening. Founders who are active on Twitter ask Twitter. Founders in a Slack group ask the group. The vocal minority gets mistaken for the market. Real listening is quieter and more structured. It means asking people who are using or leaving, at the exact moment when they can tell you something true, and reading the answers carefully enough to act on them.

His second point: the most valuable feedback is rarely a feature request. Asking a customer "how did we do?" returns a rating that tells you almost nothing. Asking "what almost stopped you from buying?" returns a list of every unresolved objection from people who bought anyway. The question matters. The timing matters more - a thank-you page survey right after checkout teaches more than an email sent three days later, because you caught the customer while the reason was still present.

His third point: you need less data than you think before acting. Founders tell themselves they will act once the sample is large enough, so the survey sits and the decision waits. Qualitative feedback does not work like a conversion test. When forty of your first fifty responses say the same thing, you do not need eight hundred. You need the nerve to go fix it. Collecting is never the hard part. Acting on what you hear, especially when it contradicts what you were excited to build, is where most founders stall.

The near-term target is $2M ARR. Jason describes it as close enough to feel real, far enough that he cannot coast to it. He thinks about it most mornings.

The larger goal is expanding Zigpoll beyond e-commerce. He built on Shopify first because that was the world he knew, but he built a feedback engine, and every business that needs to understand its customers has the same underlying problem. SaaS companies guess why people do not convert and why they churn, exactly as e-commerce brands do. The positioning and the tooling can serve both, and that expansion is already in progress.

What he is clear about is the way he wants to get there: solo, bootstrapped, and changeable on an afternoon without asking anyone's permission. The freedom to run the business that way is, in his words, the entire point. Reaching $2M without giving any of it up would mean more to him than a larger number achieved someone else's way.

You can follow his thinking on LinkedIn and X, where he shares the real numbers and mechanics behind the business. The product itself is at zigpoll.com.

The next wave of millionaires will be people who figured out how to make AI work for them.

The window to get ahead is still open. But not for long.

Here are 200+ proven ways to make money with AI in 2026.

Sign up for Superhuman AI, the free daily newsletter read by 1M+ professionals, and get instant access to all 200+ ways to profit from AI this year.

Claim your free list