How to pick a vertical SaaS one person can sell

The hard part of a vertical SaaS is not the build. It is whether one person can find the buyer and get paid.

Vertical SaaS is software for one kind of work, sold to one kind of business. A dental office. A window-washing crew. A restaurant. The buyer has a job title and a mess they already pay someone to handle. One person can sell it when they can name that buyer, find a few dozen of them without ads, and say a price out loud.

A lot of writing under that phrase is for investors: public companies, market size, and the next Toast. Those companies sell vertical software with a team. You would be selling a smaller slice of a trade by yourself. Indiecity is for one person or a team under twenty. This guide is that smaller job.

What vertical SaaS means when you have to sell it

Horizontal software does one job for many kinds of businesses. A calendar. A generic contact list. A tool “for teams.” The words on the page could belong to a law firm or a bakery.

Vertical software does a job the way one trade already does it. The screen uses their nouns. The steps match the Tuesday they already have. A yoga studio’s week is classes, teachers, and memberships. A blank calendar makes the studio translate its own work. That translation is the gap.

Vertical is the market: one trade. The company that can sell it alone is the small shape in what a micro SaaS is. Storemapper was that shape, a locator for merchants. The buyer you cannot sell to is “any modern knowledge worker,” Tringas’s phrase for a product that could help anybody and gets used by nobody. You cannot email “anybody” on a Tuesday. You can email office managers of dental practices. If your sentence still says “SMBs” or “knowledge workers,” you have a category, not a vertical you can sell.

A generic scheduling link is cheap to build in 2026. The trade-specific version is still a sales job. The nouns do not find the buyer for you. They give you something specific to say when you write.

Toast is vertical SaaS. It is the wrong plan

The software that matters in a trade is the last program an owner turns off at the end of the day. Dave Yuan uses that test at Tidemark, and Toast is his picture. Toast started as another point-of-sale system for restaurants. That system became the record the restaurant runs on, and Toast added more products around it. Toast is a public company. That path takes a team, a sales organization, and years. It is a true vertical. It is not a plan for your next two weeks.

Yuan also writes that investors used to tell vertical founders their market was too small. For you, a market you can actually reach is what makes a sale possible. Do not kill a trade because an investor would call it small. Do not keep a trade you cannot find.

Tringas used that kind of product as a real example. In his chapter on finding ideas, he pointed at Mindbody, software built around a yoga studio’s classes, teachers, and payments, instead of a generic calendar the studio has to bend into shape. Mindbody is not a one-person company. It shows how deep “one trade” can go. It is not a weekend build.

A narrow trade can be an easier fight. It is still a sale

Rob Walling splits this by the kind of company you are building. In his video on vertical and horizontal SaaS, he says a huge horizontal market fits a plan that needs venture capital or a public company, because the market is large and the growth has to be fast. For a smaller company, his read is that vertical software is often easier to grow, and easier for a later buyer to acquire, because rivals inside a trade usually have less funding and less sophisticated marketing. He says that is not true in every case.

Less sophisticated rivals are not the same thing as an easy sale. You still find the buyer, say a price, and support the product yourself. The verticals one person can sell are usually a business cost, not pocket money. If you are still choosing between a consumer app and a trade, settle that in B2B versus B2C for one person before you pick the industry.

A standalone subscription product is still the slow step. You build it and find the buyers yourself, and the revenue takes a while. That is Walling’s stair-step warning from 2015. In 2025 he added that some people can skip earlier steps if they already have the skill, the time, and the money. If you have never sold software, choosing a vertical does not remove that ramp. It stops you from spending the ramp on a buyer you cannot name.

Some ideas should fail these checks. That is the point of running them.

Six ways to throw the idea out

Throw candidates out faster than you collect them. A mediocre idea is already hard when you are the only person who can sell it. The checks below are the practical version of that filter. Fail one, and you do not yet have a vertical one person can sell.

1. They already spend money on the mess. A spreadsheet they maintain for free is a harder sale than a tool, a freelancer, or a lost job they already pay for.
2. The group has a name you can say to a stranger. Dental office managers. Window-washing companies. Not “SMBs.”
3. You can find them. If you cannot picture an hour of search producing real names, you do not have a channel.
4. The improvement is obvious against what they pay now. You will not have a sales team to explain a small gain.
5. The sale and the support fit your week. A security review is a company sale.
6. You can stand the buyer. You are about to spend years in their inbox.

The spend check is the one people skip because the idea feels clever. A business that handled the mess for free still has to find the right person and wait for a budget before they pay you. Sell where money already moves. If you need the audit, use a SaaS idea from a problem you already pay for and how to tell if people will pay.

Obvious means five times cheaper or five times better. You will not have a sales team to explain a 25 percent improvement. Storemapper launched at $5 a month against at least $1,000 for a custom locator. Those figures are in the same chapter. Do not copy the $5. Copy the pattern: the buyer already had a painful bill, and the new bill was obviously smaller.

You have to be able to find them. Lawyers you can search and visit. Window-washing companies list themselves under that phrase in ordinary directories. A group you cannot locate fails even if the software idea is tidy. Traveling vacuum salesmen are the example in that chapter. McKenzie’s Stripe Atlas guide makes the same test concrete: office managers of dental practices, and an hour of looking in one city. He used Topeka. The city is not the lesson. The hour is. If your trade cannot survive that hour, it cannot survive a month of selling.

You also need one free way to the first 25 customers, and a thought about where a later couple of hundred might gather: a directory, a forum, a platform they already use. The first five Storemapper customers were freelance clients. The next ones came one by one from Shopify discussion boards. If you cannot name the first room, fail the idea. If you cannot name the second room, keep the idea and write the gap down. You do not need the second room to send the first ten notes. You need it before you treat the trade as a living for years.

This is where a vertical stops being something one person can sell. Tringas liked small businesses of about three to ten people, and he flagged that note as less reliable than the rest. The part that still decides your calendar is the committee. Large buyers meant layers of approval and a stream of custom requests. A freelancer can say yes today and still be too price-sensitive to fund the support. If the first step is a security review, you are picking a different company than the one you can run at night.

Competitors are not an automatic fail. A market with no product and no service is often a market with no demand. Ten nearly identical products is a reason to move on. The promising middle is a consultant doing the job slowly and expensively, or software people already hate and already pay for. Storemapper came from that middle. Clients kept asking for the same locator, at $1,000 to $2,000 a time, which Tringas describes in the chapter on finding ideas. The repeated bill was the product clue. If your clue is the same, turning client work into a productized offer is the closer guide.

The chapter on what makes a good idea also tells a story without a product name. A programmer was hired by one window-washing company to schedule trucks and take payment on site. The company offered to introduce him to other customers and to a forum for that trade. One question about the perfect window-cleaning software drew hundreds of replies. A year later, in that account, the product was a full-time income and the founder was hiring. Treat it as an illustration of a group you can find, not as a revenue figure you can audit. The useful detail is the directory and the forum, not the ending.

Last check, and people rush it. If you hate the buyer, the niche is a bad job even when the checks above pass. Lawyers can be easy to find and still the wrong life if you cannot stand the conversations. You are not picking a market in the abstract. You are picking who gets your Tuesday for a long time.

Start from work you can already talk about

A filter still needs a first candidate. If you already do the work, start there. A Ruby developer who serves Ruby developers is using the only early advantage they have: they know how those people talk, and who they know. A trade you do not belong to throws that away.

That is not “only build for your own job.” If you do not belong yet, the way in is paid work in the trade. Storemapper started because the same locator kept showing up on client invoices. A folder of screenshots from an industry you have never sat in is not that, and it will not survive the first call.

Belonging has its own trap. People who build software tend to share the same problems, and a lot of those problems already have products. A tool for founders is often a wide product with a narrow label. Nomad List worked because Levels was actually moving cities and needed cost and internet speed, not because he wanted a product for founders. If your week looks like every other builder’s week, go sit with a trade full of people who are not builders. A problem you already pay for is the shorter path when the receipt is already yours.

Belonging gets you the first sentence. It does not prove they will buy. That proof is in the room where they already complain.

Read the room before you name the product

Read one room before you name the product. You are looking for three things. A mess they describe in their own nouns. Money that already moves: a tool, a freelancer, a penalty, a lost job. A place you could return to next month without buying an audience. If the threads are other builders showing demos to each other, you are in the wrong room.

The hours, the rooms, and the spend test, including when you have no followers, are in how to pick a profitable niche with no audience. Use that. Come back when the room is one trade and the thing you would sell is software. Reading gives you their words. It does not get a card charged. When you have names, customer interviews with zero users are the next conversation, and they are a different job from a sales call.

One sentence, then thirty names

When the notes repeat, write the offer in one line before you open a code editor. Who pays, with whose money, to stop which mess, at what price. If the line still needs a paragraph, you are not ready to sell it. Describe it in one sentence, or draft it in One sentence and then delete the empty words.

I help [job title] at [kind of business] stop [mess in their words]. They already pay for [tool, freelancer, or lost work]. It is [price]. I can find them in [directory, forum, or town], and they can say yes without a committee.

Then fill the sheet the way founder sales with no SDR already describes: company, person, email, people you know first. Thirty names you could actually write to. Not thirty job titles. If an hour of search cannot produce them, the vertical already failed the check above.

The note is its own piece of work. When the person already knows you, follow how to do founder sales with no SDR. When they are a stranger, use how to write a cold email that books a SaaS demo. The first ten payments, if the list is real, are how to get your first 10 paying customers without ads. Do not invent a second sales system while you are still choosing the trade.

Two weeks, then keep the trade or widen it

Pick one candidate and run it. Switching trades mid-week is how the filter never finishes.

1. This week, day 1: write the one-line offer. If you cannot name the job title, stop and pick a tighter group.
2. Days 2 and 3: read one real room. Notes only. Pains, products, prices. No pitching.
3. Days 4 and 5: thirty names. Company, person, email. If the hour of search fails, the vertical fails.
4. Next week: one page, one price, ten human notes. Count replies and payments.
5. If the notes are silence, change the buyer or the problem. Do not add a feature to avoid the silence.

The fear underneath the pick is that the trade is too small and you will be stuck. Keep one trade on the page long enough to learn. A backup list of five audiences is how you never commit. Widen later, if you cannot create demand or the group you can reach is too thin. On that 2025 episode, Walling’s picture was construction, the work he grew up around: electrical contractors on the page, not every contractor, and stay open if a neighboring trade starts asking. The question to answer is what problem you solve, and for whom. The product name comes after that.

He widened Drip that way. It started with SaaS founders, a group he could reach. Course creators kept showing up, so he added them. Widen because a neighboring trade asked. Adding dentists, gyms, and every messy team during the first week walks you back to horizontal software with a vertical adjective on it.

If the money in the sentence is a hobby budget, go back to the B2B choice before you build. If you cannot say the number, price it from what they already spend, not from how long the build felt. If you are about to disappear into a repo for months, validate the idea before code. You are allowed to fail this test. A failed test is cheaper than a finished app for a buyer you still cannot find.

When someone in the trade pays, send the story. When the product has a name, put it on the map. You can join Indiecity while the thirty names are still a spreadsheet. The room will not pick the trade for you.