I Killed Nine Business Ideas. The Real Reason Had Nothing to Do With the Market.

I killed nine business directions in a few months. Recruiting. B2C. Code generation. Enterprise. Banks. Compliance. Hardware. Services and contracting. Agent orchestration.

Every rejection came with a reason, and every reason was about the market. This one is taken — three vendors with funding and teams. This one is empty, and empty usually means nobody pays. Solo entry is impossible here: tender, vendor audit, twelve months to the first invoice. The contract value is too small to feed one person. That one needs hiring.

I was pleased with myself. Disciplined selection, no falling in love with an idea, every "no" backed by research.

Then I wrote all nine into one list and looked at them together.

The one thing all nine had in common

Too much convincing, too little building.

Procurement. Vendor audits. Twelve-month deal cycles. Cold emails to people who don't want to read them. Demos for a committee of five where the decision is made by a sixth person who wasn't on the call. Conferences. Networking with people I would choose not to talk to.

None of my market reasoning was a lie. But it arrived second. Something else fired first, and the market rationale caught up afterwards — like a lawyer called in after the verdict.

I said it out loud in that session and tripped over my own sentence:

I get high from building. And it doesn't help me earn, because so far all I do is build.

Building is the reward. Selling is the price. And I was crossing out the price before I started counting.

This isn't a discovery about the market. It's a discovery about the mechanics of my own decisions, and it's worse, because you can change a market and you can't really change that.

The scissors: I can reach people who don't pay, and I can't reach people who do

Who I can reach: engineers, small startup founders, people who write code with an assistant and argue about it in threads. My people. I speak their language without a translator.

Here's my own observation, not a law of the market: software is hardest to sell to people who build software themselves.

An engineer looking at a product doesn't see the result. He sees the work — "a couple of endpoints, a queue, one table, I'd build that over a weekend." The price reads not as payment for value but as a markup on something he can do himself. The question stops being "is this worth the money" and becomes "is this worth my two weekends" — and the answer is almost always no, because he likes his weekends.

I'm that buyer. I pay for Claude, Spotify, and a couple of servers. Everything else I either built or found free and made peace with the friction. So when I say "my audience pays poorly," I'm mostly talking about myself, not about them.

Who does pay: enterprise, banks, large corporate buyers. I can't get there solo. Not for lack of knowledge — seven years in enterprise Java, integrations, migrations, regulated environments. I know how those systems are built and why they look the way they do.

What's missing isn't knowledge. It's the entry format: a legal entity with a track record, references, liability insurance, and a person who attends meetings for six months.

That's the scissors. The people I can reach pay poorly; the people who pay well are out of reach. So I need a bridge — a buyer technical enough to speak my language, but spending company money rather than his own. Software houses. Systems integrators. Agencies. Technical businesses with 10 to 100 people.

That's the first positive specification I've produced in this entire search. Before it, I had a growing list of what I wouldn't do — which turns out to be not a strategy but a way to never start.

Boring is where the money is. I'm still not going there.

Now the uncomfortable part.

Interesting problems attract engineers. Engineers build products. Products pile up, compete, and margins collapse. Look at any developer tool category: twenty options, half of them free, because their authors were interested rather than commercial.

Boring B2B repels those same people. Waste tracking. Tachograph compliance. Subsidy applications. Shift planning. Nobody goes there for fun — which is exactly why the money is still sitting there for whoever can tolerate it.

I know this. I wrote it in my own research notes. And every time I reached a niche like that, I found a reason it wasn't a fit.

The most interesting case is recruiting. I have a product for recruiting agencies and I still run it: I know their processes, their tools, their complaints, and I have people to talk to. That's access to an industry most people don't have and can't buy.

And I still didn't pick recruiting as the territory everything else gets built around. The product lives its own life — that's one decision. Making that audience mine for the next five to ten years is a different one, and to that I said no. Not because the people are bad. Because what they talk about among themselves doesn't interest me.

The difference between "I run a product for an industry" and "I make that industry my territory" looks bureaucratic until you count it in years. The first is a project. The second is what you read in the morning, who you sit next to at conferences, and what you think about in the shower. My edge was in the second one, and I gave it up deliberately. That's an expensive decision and I won't pretend it was rational.

The ratio is a parameter, not a character flaw

My first instinct was to file myself under lazy. Real founders sell. You're just hiding in a terminal from the unpleasant part.

Half true. But useless as a conclusion: "you're not good enough" produces no action except guilt on a schedule.

The useful version is different. Not wanting to sell is not a character trait. It's a parameter. It has a value, you can name it as a number, and then you can work with it instead of being ashamed of it.

My value is 70/30. Seventy percent of effort into what I build, thirty into explaining, selling, and negotiating. Not zero. Zero would be a lie — a product without sales is a hobby with a git repo. But not thirty-seventy either, because at that ratio I quit for certain, and quitting after eight months costs more than never starting.

From that point on, 70/30 is a selection criterion rather than a wish. A direction that demands the inverse gets rejected immediately — no month of research, no market rationale invented after the fact. That's the actual saving: I used to spend weeks justifying a decision I had already made in the first minute.

Next to the number sits the rest of the specification, equally operational. Fragmented market, not horizontal. Money already being spent, with existing vendors and a budget line. A buyer who doesn't run tenders. Digital product. Solo entry, in pieces. The ceiling is bootstrap-level rather than unicorn-level — and I keep it as a range rather than a figure, because it depends on the market: the same product sells for several times more in the US or UK than in Poland, at identical build cost.

Four tools that work with the ratio instead of against it

A diagnosis without a method is just well-formulated whining. So here's what actually treats it. None of it is mine — it's all written down by people who walked this road. My contribution is only the order you apply it in when you don't like selling.

Tool 1: Pick the channel before the product

Brian Balfour, dissecting why good products fail to grow, puts it bluntly in Product Channel Fit Will Make or Break Your Growth Strategy: products are molded to fit channels, not the other way around. The channel is a given, it has its own mechanics, and the product has to fit inside them.

For someone with my ratio this is the main lever. The usual order — build it, then figure out how to sell it — ends in a warehouse of finished products with no buyers, because the selling step gets postponed and "later" never arrives.

The correct order: first pick a channel that works without your personal presence — search, a marketplace, a directory, an integration inside someone else's product, a reseller partner — and only then design a product that sells in that channel.

The practical test for a usable channel: the buyer shows up with a formed intent. He's already looking for a solution; you don't have to convince him he has a problem. Everything that requires creating demand is convincing work, and no amount of code will shrink it later.

Tool 2: Don't start with SaaS

Rob Walling's Stair Step Method of Bootstrapping lays out three steps. Step one: a simple one-time-sale product riding a single existing distribution channel — a plugin repository, an app store — where discovery is already built in and not your problem. Step two: repeat and scale that model until it replaces your salary. Step three: only then a standalone product with recurring revenue, where higher lifetime value finally pays for real marketing.

It's usually explained as a way to learn how to build products. For us it has a different meaning.

Step one exists precisely so you can take money from strangers without selling to them personally. You find out whether you can make something people buy before you sign up for the role of salesperson.

Tool 3: Reframe the conversation itself

Rob Fitzpatrick's The Mom Test shows that talking to the market is not pitching. You don't describe your idea. You collect facts about the other person's life: what they did last time, what they use now, what they pay for it, what broke, what they already tried and with what.

The difference isn't cosmetic. A pitch is persuasion, and engineers hate it for good reason — you're asking someone to agree with you. An interview is data collection, the debugging mode you already live in.

Roughly half of that thirty percent of "convincing" turns out to be interviewing, once you stop confusing the two.

Tool 4: Many small bets instead of one big one

Daniel Vassallo's Small Bets portfolio logic: several modest products, each with a small entry cost and a small convincing cost. One failure doesn't erase a year.

For someone whose bottleneck is sales, there's a second benefit. You stop betting everything on a single conversation you can't afford to lose — and a conversation you can afford to lose is a conversation you're much better at.

Put together, the order is: name your ratio as a number → pick a channel that fits that number, before the product → make the first product ride someone else's distribution → run conversations as fact-finding, not as pitches.

None of this makes selling unnecessary. It makes the volume predictable — and predictable is something you can tolerate.

---

I picked this path over the "break yourself" path. Right now I'm testing the bridge segment — that technical buyer spending company money. Whether it works, I don't know yet. But I've stopped explaining my rejections with the market.