How to start a one person business with AI

The 2026 AI business model, in 29 categories and 346 points

Everything below is free and there is no signup on this page. It is the whole method for how to start a one person business with AI, in 29 categories and 346 points, written for one person with no staff, no investors and no inventory to speak of.

29 categories. 346 points. Free, and there is no signup on this page.

What is on this page

  1. Which one person business model actually works, and which ones quietly turn into a job? 13 points
  2. How do I find my target audience when "small business owners" is not an answer? 14 points
  3. How do I validate a business idea in a week, before I build anything? 13 points
  4. What is brand positioning, and why would somebody buy from me instead of the other twelve? 11 points
  5. How do I create an offer people actually buy, and what is the math underneath it? 14 points
  6. How do I price my product when I have no idea what it is worth? 13 points
  7. How do I get customers when I can only run one channel properly? 14 points
  8. How do people find things now that AI search answers the question before they click? 11 points
  9. Which content is worth making, and which posts are a wasted Saturday? 13 points
  10. How do I build an email list before I have anything to sell? 13 points
  11. When am I ready to run Facebook ads, and what will it cost me to learn on them? 14 points
  12. How do I borrow somebody else's audience without paying for it up front? 11 points
  13. How do I write copy that sells when I am not a copywriter? 13 points
  14. Where exactly am I losing people between a stranger seeing me and somebody paying? 12 points
  15. How do I sell something over a thousand dollars without sounding like a pitch? 11 points
  16. How do I get testimonials and proof when I have barely any customers yet? 11 points
  17. What happens to my business if a platform removes me tomorrow? 10 points
  18. What should I automate first, and how would I even know an automation had broken? 11 points
  19. What is AI genuinely good at in a business, and where will it embarrass me? 13 points
  20. Which three numbers do I look at every week, out of the thirty I could? 11 points
  21. What do I actually have to do about security when I am the whole IT department? 11 points
  22. Where is my money going, and is any of what I sell losing money? 13 points
  23. Do I need an LLC, a business bank account and bookkeeping before I start? 12 points
  24. Which rules can genuinely get me fined, and which ones is everybody ignoring? 12 points
  25. How do I keep the customers I already have instead of hunting new ones? 12 points
  26. What do I hand off first when I am the whole team, and what do I never hand off? 11 points
  27. How do I know when to quit something instead of pouring in another six months? 10 points
  28. How do I keep going for years without burning out or getting bored? 10 points
  29. Could I ever sell this business, and what would somebody actually pay for? 9 points

Category 1 of 29  ·  Which business model actually works for one person

Which one person business model actually works, and which ones quietly turn into a job?

A one person business model works when doubling sales does not double your workload. Digital products, memberships, licensing, templates, newsletters and high ticket advisory with a hard cap on client count all pass that test. Staff, held inventory, physical logistics and open ended hourly work all fail it, and that is the whole difference between a business and a job with better branding.

  1. The constraint is not money or ideas. It is your hours and your attention. Pick a model that does not consume them linearly.
  2. Good solo models: digital products, courses, software, templates, memberships, newsletters, licensing, affiliate and sponsorship, print on demand, high ticket advisory with a hard cap on client count.
  3. Bad solo models: anything with staff, inventory you hold, physical logistics you touch, a service that bills by your hour with no ceiling, or a marketplace that needs two sides to show up.
  4. The test: if you doubled sales tomorrow, would your workload double? If yes, it is a job with better branding.
  5. Recurring beats one time. A membership at fifty a month with two hundred people is a real income with a customer count you can actually serve alone.
  6. Digital delivery means zero cost per extra sale. That is the single biggest advantage a solo operator has, and it is why margins matter more than volume.
  7. Service work is not the enemy. It is the fastest cash and the best research. Just cap it, price it high, and use it to fund the product.
  8. Beware anything requiring same day human response at scale. One person cannot cover a support queue and build at the same time.
  9. Beware anything where a single account ban ends the business. Pick a model where you own the customer relationship.
  10. AI changes which models are viable for one person. Content volume, research, coding, design, and customer replies used to require hires. They do not now.
  11. AI does not make a bad model good. It makes a good model faster.
  12. Stack two or three models on one audience rather than running two audiences. One audience, several ways to buy.
  13. Decide early whether you want a business that grows or one that pays you well and stays small. They are built differently, and both are legitimate.

Category 2 of 29  ·  Who exactly you sell to

How do I find my target audience when "small business owners" is not an answer?

How to find your target audience: name one person with one problem, not a segment. If you cannot say what they typed into a search bar last week, you do not have a market yet. The problem needs an existing budget line and enough urgency that it gets solved this month rather than someday.

  1. Name one person, not a segment. Not "creators," not "small business owners." One person with one problem.
  2. If you cannot say what they typed into a search bar last week, you do not have a market yet.
  3. The problem needs an existing budget line. People pay to remove a pain far more readily than to reach an aspiration.
  4. Urgency decides everything. A today problem sells. A someday problem sits in an abandoned cart.
  5. Can they actually pay, and can they decide alone? Selling to someone who has to ask a boss doubles your sales cycle.
  6. Find where they already gather. A subreddit, a Facebook group, a Discord, a trade forum, a conference, a comments section.
  7. Existing sellers are proof of buyers, not a wall. An empty niche is usually empty for a reason.
  8. Talk to twenty of them before you build. Twenty real conversations beats six months of guessing.
  9. Ask about their last purchase in this area, what it cost, and what disappointed them. Do not ask whether they like your idea. People are polite and it will cost you a year.
  10. Watch what they do, not what they say. Buying behavior is the only honest answer.
  11. Reachability matters as much as need. If getting in front of them is expensive, the business is expensive.
  12. Prefer a customer who buys again. Acquiring a new buyer is the most expensive thing you do.
  13. Write the person down on one page and check every decision against them. Solo operators drift, and this is the anchor.
  14. Pick a group you can stand to be around for five years. You will be reading their complaints every day.

Category 3 of 29  ·  Proof of demand, before you build a thing

How do I validate a business idea in a week, before I build anything?

How to validate a business idea in about a week: read five signals, then take money. The five are search volume, marketplace sales, competitors still paying to run ads, active communities, and how long the incumbents have survived. Then presell, because a deposit is strong proof and a waitlist is weak proof.

  1. Demand either exists or it does not, and you can find out in about a week.
  2. Look at five signals: search volume, marketplace sales, paid competitors still running ads, active communities, and how long the incumbents have survived.
  3. Saturation is proof of buyers. The way through a crowded market is the thing everybody in it does badly, not an empty market nobody wants.
  4. Judge the quality of competition, not the quantity. Twenty bad competitors is an opening. Two great ones is a wall.
  5. Read the one and two star reviews of everything selling in the space. That is your product roadmap and your sales copy in one place.
  6. Presell. Money is the only validation that counts. A waitlist is weak proof, a deposit is strong proof.
  7. Run a landing page with a real price and a real buy button before the thing exists. Refund anyone who buys if you decide not to build it.
  8. Find who already failed at this and read why. Failure post mortems are more useful than success stories, and nobody reads them.
  9. Set your green light criteria before you look at the data, or you will talk yourself into it.
  10. Do not validate on friends, family, or your own warm audience only. They are being kind.
  11. Time box the research. Two weeks maximum. Research is the most comfortable form of procrastination available.
  12. Use AI to compress this. It can read a thousand reviews and pull the recurring complaints in an hour. Then you verify a sample yourself, because it will occasionally invent a pattern that is not there.
  13. A red verdict is a good outcome. It cost you two weeks instead of a year.

Category 4 of 29  ·  Positioning: why you and not the other twelve

What is brand positioning, and why would somebody buy from me instead of the other twelve?

Brand positioning is the answer to why somebody buys this instead of that. Without one, price becomes the only answer available and you lose that fight. Position against a named alternative, in one sentence a stranger can repeat back to you.

  1. Positioning is the answer to "why would I buy this instead of that." If you do not have one, price becomes the only answer, and you lose.
  2. Position against a named alternative, not against the void. "Instead of a two thousand dollar course, this."
  3. Pick a wedge: narrower audience, faster result, simpler method, different worldview, or a format nobody else offers.
  4. Being second with a better angle beats being first. First movers do the expensive market education for you.
  5. A real point of view is the strongest moat a solo operator has. Opinions cannot be copied by someone who does not hold them.
  6. Do not compete on price unless your costs are structurally lower. You will just be poor and busy.
  7. The category you claim decides who you get compared to. Choose it deliberately.
  8. One sentence, tested on a stranger. If they cannot repeat it back, it is not positioning, it is a paragraph.
  9. Positioning shows up in the name, the price, the first line of the page, and what you refuse to sell.
  10. Being for someone means being not for someone else. Narrow is not a limitation, it is the whole mechanism.
  11. Revisit it when the market moves. AI moved a lot of markets in the last two years and a lot of positioning statements are now describing something everyone gets free.

Category 5 of 29  ·  The offer, and the math under it

How do I create an offer people actually buy, and what is the math underneath it?

How to create an offer: it is the outcome, the speed, the risk removal, the price and the terms together, never just the product. Three numbers sit under it. What a customer is worth over their whole life with you, what one costs to get, and how many months it takes to earn that back. Payback period is the one that governs how fast you can grow, because ratios do not pay bills and timing does.

  1. The offer is not the product. It is the outcome, the speed, the risk removal, the price, and the terms, together.
  2. Sell the outcome. Deliver the mechanism. Nobody wants a course, they want the thing on the other side of it.
  3. Risk reversal moves conversion more than almost any copy change. A guarantee shifts the risk to you, which is where it belongs if the thing works.
  4. Watch guarantee abuse. Thirty days is standard, sixty converts better, and both attract a small percentage who were never going to do the work.
  5. Bonuses should solve the next problem the buyer will hit, not pad the page.
  6. Know what a customer is worth to you over their whole life with you. Not one sale. All of it.
  7. Know what it costs you to get one, counting your time honestly.
  8. The number that actually governs how fast you can grow is payback period. If it takes five months to earn back what you spent to acquire someone, you need five months of cash sitting there. Ratios do not pay bills, timing does.
  9. Contribution margin per sale, after processing fees, delivery costs, affiliate payouts, and refunds. That is the real number.
  10. Refunds are a margin line, not a customer service issue. Track the rate weekly.
  11. One offer, mastered, beats five half built. Solo operators die of too many products, never too few.
  12. Test the offer before you buy traffic. A bad offer with good traffic is an expensive way to learn nothing.
  13. Recurring revenue changes every calculation in this section. Build it in if the model allows.
  14. Have a version of the offer you can deliver at ten times current volume without hiring anyone. If you cannot, cap the sales.

Category 6 of 29  ·  Pricing

How do I price my product when I have no idea what it is worth?

How to price your product: on the value of the outcome, not on your costs or your hours. A low price does not make the sale easier, it makes it cheaper, and it selects for the buyers who complain most and refer least. Three tiers is the default, and most solo operators are underpriced for years.

  1. Price on the value of the outcome, not on your costs or your hours.
  2. A low price does not make the sale easier. It makes it cheaper, and it attracts the buyers who complain most and refer least.
  3. Price sets your customer. Charge more and you get people who take the work seriously.
  4. Anchor. A higher tier nobody buys makes the middle tier look reasonable.
  5. Three tiers is the default for a reason. Most people take the middle one.
  6. Payment plans raise the price people will accept and raise your default rate. Both are true. Price them accordingly.
  7. Raise prices earlier than feels comfortable. Most solo operators are underpriced for years.
  8. Discounting teaches your audience to wait for the discount. Once trained, they are hard to untrain.
  9. Test price changes on new traffic only, and give it enough volume to mean something.
  10. Handle currency and tax at checkout properly. It is a conversion issue as much as a legal one.
  11. Grandfather existing customers when you raise. It costs little and buys a lot of goodwill.
  12. Free is a channel, not a price. Know exactly what the free thing is supposed to lead to, or kill it.
  13. High ticket suits a solo operator better than most people admit. Ten sales at three thousand is a better year than three hundred at a hundred, and it is far less support.

Category 7 of 29  ·  Distribution is the whole game

How do I get customers when I can only run one channel properly?

How to get customers as one person: one channel, mastered, before you touch a second. Making things got cheap and getting seen did not, so the product is the easy half now. Thin across five channels is invisible on five channels.

  1. Making things got cheap. Getting seen did not. Assume the product is the easy half now.
  2. One channel, mastered, before you touch a second. Thin across five is invisible on five.
  3. The channel has to match the offer. A hundred dollar product does not sell on a sales call, and a ten thousand dollar service does not sell from a Reel.
  4. There are four ways to get in front of people: organic content, paid ads, other people's audiences, and direct outreach. Pick the one you can actually sustain alone.
  5. Volume with feedback. Post, read the numbers, adjust, repeat. The people winning have more attempts, not better instincts.
  6. Consistency beats intensity. A year of weekly beats a month of daily followed by silence.
  7. Make it native to the platform. Cross posting the same file everywhere is why most of it dies.
  8. On short form, the first two seconds decide everything. On long form, the title and thumbnail do.
  9. Search driven content compounds quietly. Social content spikes and dies. You want both, and you want to know which is which.
  10. You are building a body of work, not a stack of posts. The library is the asset.
  11. Measure what drives money, not what drives follows. They are frequently different pieces.
  12. Give a channel six months before judging it. Give it up entirely if it has had a fair year.
  13. When something works, go deeper before going wider. Solo operators expand too early, constantly.
  14. AI lets one person produce at a volume that used to need three. Use it for the production, never for the judgment about what to make.

Category 8 of 29  ·  How people find things now

How do people find things now that AI search answers the question before they click?

AI SEO is the shift underneath all of this: assistants answer plenty of questions without sending anyone to a website, so content that only answers a simple question is worth less than it was. Discovery is split across classic search, social search, AI assistants, marketplaces and video, and your customer uses two or three of them. The goal moves to being the source that gets cited, which means original data, first hand experience, and numbers nobody else has.

  1. Discovery is split across classic search, social search, AI assistants, marketplaces, and video. Your customer uses two or three of them and you need to know which.
  2. AI assistants answer a lot of informational questions without sending anyone to a website. Content that only answers a simple question is worth less than it was.
  3. Which means the goal shifts to being the source that gets cited: original data, first hand experience, numbers nobody else has, opinions a model cannot generate.
  4. Structure your pages so machines can read them. Clear headings, the answer near the top, plain language, proper markup.
  5. Brand search is the safest traffic there is. People typing your name cannot be taken from you by an algorithm change.
  6. Video is still expensive to fake well, which makes YouTube and podcasts unusually defensible right now.
  7. Communities are where buying decisions actually get made. Reddit, Discord, niche forums, group chats. You cannot advertise there, you can only be useful there.
  8. Marketplaces are search engines with a checkout attached. Amazon, Etsy, app stores, YouTube. Treat them as channels, not as landlords.
  9. Word of mouth is still the best channel and the least controllable. The only lever is making something worth mentioning.
  10. Ask every customer how they found you, in one field at checkout. It will disagree with your analytics, and it will usually be righter.
  11. Own your name across every platform, even the ones you do not use.

Category 9 of 29  ·  Content that does work

Which content is worth making, and which posts are a wasted Saturday?

A content marketing strategy that works gives every piece one job: reach, trust, or conversion. Reach content answers a question strangers already ask, trust content shows the work including the losses, and conversion content shows the product doing the thing. The common failure is making only trust content and wondering why the audience never grows.

  1. Every piece has one job: reach, trust, or conversion. If you cannot say which, do not publish it.
  2. Reach content answers a question strangers are already asking. It brings new people.
  3. Trust content shows the work, the process, the losses, the specifics. It converts the people reach brought.
  4. Conversion content shows the product doing the thing. Most people never make any.
  5. The common failure is making only trust content and wondering why the audience never grows.
  6. One idea per piece. Two ideas is zero ideas.
  7. Repurpose across formats, but rewrite for each format. A caption is not a script is not an email.
  8. Make things that keep working: tutorials, comparisons, calculators, templates, teardowns. Evergreen beats topical for a solo operator, because you cannot outrun the news cycle alone.
  9. Publish on a cadence you can hold for a year with the flu. Then hold it.
  10. Batch. One recording day, one writing day, one editing day. Context switching is the solo tax.
  11. Read the analytics as feedback, not as a verdict on you.
  12. AI is very good at the second draft, the variations, the descriptions, the repurposing. It is bad at the idea and bad at your voice. Guard both.
  13. A content library with traffic is a saleable asset on its own.

Category 10 of 29  ·  The audience you actually own

How do I build an email list before I have anything to sell?

How to build an email list: start on day one, before you have anything to sell, because followers are rented and email is owned. The reason to join has to be specific and immediately useful, since the word newsletter is not a reason. A thousand people who open is a business, and this list is the single most valuable thing you will build.

  1. Followers are rented. Email and SMS are owned. Build the owned list from day one, before you have anything to sell.
  2. The reason to join has to be specific and immediately useful. "Newsletter" is not a reason. "The exact checklist I use" is.
  3. Deliverability is a technical job you have to do once: your own sending domain, proper authentication, and a slow warm up. Skip it and your emails land in spam forever.
  4. Clean the list. Subscribers who never open are actively hurting the ones who do.
  5. Send regularly or the list goes cold and your deliverability with it. Monthly is the floor.
  6. Segment by behavior, not by demographic. What they clicked beats who they are.
  7. The welcome sequence does more work than everything else combined. Write it once, properly.
  8. Every broadcast should have one job and one link.
  9. Open rates are unreliable now. Watch clicks and revenue per send.
  10. Keep an export of your list somewhere off the platform, refreshed on a schedule.
  11. SMS gets far more attention and carries far stricter rules. Treat consent as sacred.
  12. A thousand people who open is a business. Fifty thousand who scroll past is a number you can post about.
  13. This list is the single most valuable thing you will build. It survives platform bans, algorithm changes, and pivots.

Category 11 of 29  ·  Paid ads

When am I ready to run Facebook ads, and what will it cost me to learn on them?

Facebook ads for beginners, and every other paid channel, amplify what already works and never fix a broken offer. You are ready when organic has proven the offer converts, and not before. Write down your maximum acceptable cost per customer, your kill number and your scale number before you spend a dollar, because solo operators go under from a working campaign more often than a failing one.

  1. Ads amplify what already works. They never fix a broken offer, and they will drain you faster than anything else on this list.
  2. Do not run paid traffic until organic has proven the offer converts. This rule is broken constantly and it is the most expensive mistake available to a beginner.
  3. Know your maximum acceptable cost per customer before you spend a dollar. Write it down.
  4. Start with retargeting. Warm traffic is the cheapest money you will ever spend.
  5. Creative is the variable that matters most now. Platforms handle targeting better than you do.
  6. Test many creatives against few audiences, not the other way round.
  7. The first spend buys data, not sales. Budget for it and do not panic.
  8. Platform attribution overstates itself, every platform, always. Compare against your actual bank deposits.
  9. Watch payback period, not one day's return. A campaign that pays back in ninety days is fine if you have ninety days of cash, and fatal if you do not.
  10. Set the kill number and the scale number before launch, so the decision is arithmetic and not emotion.
  11. Landing page speed and message match. If the ad says one thing and the page says another, you paid for a bounce.
  12. Ad accounts get banned for boring reasons. Read the policies, keep a backup account, and never build the whole business on one.
  13. Never scale faster than your cash allows. Solo operators go under from a working campaign more often than a failing one.
  14. AI makes creative production almost free. That means everyone's creative volume went up, and the bar for stopping a thumb went up with it.

Category 12 of 29  ·  Other people's audiences

How do I borrow somebody else's audience without paying for it up front?

Affiliate marketing for beginners, guest spots, referral programs, bundles and joint launches are all the same move: borrowing trust. It is the fastest route to strangers and it costs money only when it works. Give first, be useful in public before you ask for anything, make partnering easy, and pay on time.

  1. Borrowing trust is the fastest route to strangers, and it costs money only when it works.
  2. Affiliates and referral programs. Make the terms generous enough to be worth someone's attention.
  3. Guest spots: podcasts, newsletters, other people's videos. One good appearance can outperform a quarter of posting.
  4. Collaborate with peers at your size, not above it. People one rung up ignore you, people at your level trade with you.
  5. Bundles and joint launches put you in front of several lists at once.
  6. Give first. Be useful in public before you ask for anything, every time.
  7. Make partnering easy: swipe copy, assets, a link that works, and a dashboard they can check.
  8. Track it honestly and pay on time. Affiliates talk to each other.
  9. Communities are for contribution, not advertising. Six months of being useful outperforms one promotional post, and the promotional post gets you banned.
  10. One good partner beats a hundred cold emails, but you have to send the cold emails to find them.
  11. Be careful renting a list you cannot verify. Ask for their recent open rates and a screenshot, and expect a real answer.

Category 13 of 29  ·  Copy and messaging

How do I write copy that sells when I am not a copywriter?

Copywriting for beginners comes down to one rule: clear beats clever, every time. Write to one person, lead with the outcome they want, and use their exact words lifted from reviews, calls and comments, because you are not writing copy so much as collecting it. Sounding human is now a trust signal, since everything else sounds generated.

  1. Clear beats clever, every time, with no exceptions worth chasing.
  2. Write to one person. Sales copy addressed to a crowd sounds like a press release.
  3. Lead with the outcome they want, not the features you built.
  4. Use their exact words, lifted from reviews, calls, and comments. You are not writing copy, you are collecting it.
  5. Specific beats general. A number, a name, a date, a detail.
  6. Handle the objection inside the copy. The objection they do not voice is the sale you do not make.
  7. Put proof next to every claim. A claim alone reads as a boast.
  8. One clear action per page. Two calls to action halve both.
  9. The headline does most of the work. Write twenty and pick one.
  10. Read it aloud. If you would not say it to a person, cut it.
  11. Sounding human is now a trust signal, because everything else sounds generated. This is a real advantage and most people are throwing it away.
  12. Test the headline, the offer, and the price before you rewrite the body. That is where the movement is.
  13. AI writes serviceable first drafts and terrible final ones. Use it to get to the page, then make it sound like you.

Category 14 of 29  ·  The buying path

Where exactly am I losing people between a stranger seeing me and somebody paying?

Your sales funnel is the actual route from stranger to buyer, and you find the leak by writing down every step, because there are more of them than you think. Every step loses people, so measure the drop off at each one and fix the worst one rather than the easiest one. Check the whole path on a phone yourself, at least monthly.

  1. Map the actual route from stranger to buyer. Write down every step. It is usually longer than you think.
  2. Every step loses people. Know your drop off at each one and fix the worst.
  3. Fewer steps is usually better, with one exception: expensive things need more steps, not fewer.
  4. Checkout: fewest possible fields, mobile first, a processor people recognize.
  5. Order bumps and upsells at checkout are the cheapest revenue increase available. One extra line on the page.
  6. Abandoned cart follow up recovers a meaningful slice, and takes an hour to set up once.
  7. Deliver instantly and confirm clearly. Confusion in the first ten minutes becomes a refund in three days.
  8. The first twenty four hours after purchase set the refund rate. Make sure they get one win fast.
  9. Free trial, paid trial, or money back guarantee. Each pulls a different buyer. Pick deliberately and describe it accurately.
  10. Make cancellation genuinely easy. Hard cancellation generates chargebacks, which threaten your ability to take payments at all.
  11. Test one thing at a time or you learn nothing.
  12. Check the whole path on a phone yourself, at least monthly. Most of your traffic is there and most solo operators only ever look at the desktop version.

Category 15 of 29  ·  Selling to an actual human

How do I sell something over a thousand dollars without sounding like a pitch?

How to close a sale above roughly a thousand dollars: in a conversation, not on a page. Ask far more than you talk, diagnose before you prescribe, and state the price plainly and then stop, because the pause is doing the work. A prescription without a diagnosis is a pitch, and it feels like one.

  1. Anything above roughly a thousand dollars still sells in a conversation, not on a page.
  2. Ask far more than you talk. The person who asks the questions is running the conversation.
  3. Diagnose before you prescribe. A prescription without a diagnosis is a pitch, and it feels like one.
  4. Qualify out fast. Your scarcest resource is hours, so a polite early no is a gift to both of you.
  5. State the price plainly and then stop talking. The pause is doing the work.
  6. Almost every objection is money, time, trust, or need. Know which one you are actually hearing.
  7. Never chase. Follow up on a set schedule, then let it go. Chasing reads as need, and need repels.
  8. Decide on the call where you can. Written proposals kill momentum.
  9. Record your calls and listen to them. It is unpleasant and it is the fastest improvement available.
  10. Say no to bad fits. One wrong client can eat a month of a solo operator's year.
  11. Most of the money is in the follow up, and almost nobody does it.

Category 16 of 29  ·  Trust and proof

How do I get testimonials and proof when I have barely any customers yet?

How to get testimonials that carry weight: named, with real photos and a specific outcome, because anonymous praise is worse than none. Polish stopped being evidence of anything, since anyone can produce a beautiful page in an afternoon. What still works is your face, your voice, your track record, case studies with real numbers, and details only somebody who actually did the thing would know.

  1. Polish stopped being evidence of anything. Anyone can produce a beautiful page in an afternoon now.
  2. What still works: your face, your voice, your track record, and details only someone who actually did the thing would know.
  3. Testimonials with real names, real photos, and specific outcomes. Anonymous praise is worse than none.
  4. Case studies with numbers. One real number beats a page of adjectives.
  5. Show the work in public, including what failed. The willingness to name a loss reads as the only honest voice in the category.
  6. Third party proof carries more weight than anything you say about yourself: reviews, ratings, press, being cited by someone else.
  7. Consistency over years is the proof that cannot be faked quickly, which is exactly why it works.
  8. A real address, a real support channel, and a real refund policy do quiet, heavy lifting.
  9. Never overstate. One provable claim beats five vague ones, and a single caught exaggeration takes the rest down with it.
  10. Answer criticism in public, calmly, once.
  11. If you use AI in the work, do not build the brand on pretending otherwise, and do not make it the story either. Talk about what the customer gets.

Category 17 of 29  ·  Your own home base, and platform risk

What happens to my business if a platform removes me tomorrow?

How to build a website for your business, and why it outranks any platform you post on: own the domain, the site and the list. Everything else is a tenancy, and any platform can remove you tomorrow with no explanation and no appeal that works. Spread your single points of failure, keep a second payment processor already open, and keep customer contact details outside the store platform.

  1. Own the domain, the site, and the list. Everything else is a tenancy.
  2. Assume any platform can remove you tomorrow with no explanation and no appeal that works.
  3. Spread your single points of failure: payment processor, email service, hosting, domain registrar. Never let one company hold two of them if you can help it.
  4. Have a second payment processor account already open. Opening one under pressure takes weeks you will not have.
  5. Export everything on a schedule. Customers, content, files, financials.
  6. Do not build your only product inside somebody else's ecosystem unless the economics are so good you can accept losing it.
  7. Keep customer contact details outside the store platform. If the shop goes, the customers should not go with it.
  8. Read the terms of the platforms you depend on. Especially the parts about account termination and money holds.
  9. Know the appeal path before you need it, and keep records that would support one.
  10. Once one traffic source works, start the second. Not before, and not much later.

Category 18 of 29  ·  Systems and automation

What should I automate first, and how would I even know an automation had broken?

Business automation starts by doing the job by hand first, writing the process down while you do it, then automating the repeatable middle and keeping the judgment at each end. You know it broke because you built the alarm, and without one a silent failure looks exactly like everything working. Anything that sends, publishes, deletes or charges gets a human check before it fires.

  1. Do it by hand first. You cannot automate a process you have never run, and you will automate the wrong version.
  2. Write the process down while you do it. That document is what you hand to AI or to a contractor later.
  3. Automate the repeatable middle. The judgment at each end stays yours.
  4. Every automation needs an alarm when it breaks. A silent failure looks exactly like everything working.
  5. Prefer boring tools that have existed for years over clever ones that launched last month.
  6. Do not automate before the offer is proven. You will build a beautiful machine for making something nobody wants.
  7. One source of truth for your data. Two spreadsheets that disagree is worse than one that is out of date.
  8. Naming conventions and a folder structure, decided once. Future you will thank present you weekly.
  9. Review your automations quarterly. Dead ones cost money and, worse, they report success.
  10. The honest test on any automation: does it save more of your hours than it costs to maintain?
  11. Anything that sends, publishes, deletes, or charges gets a human check before it fires. That rule has saved more solo businesses than any tool has built.

Category 19 of 29  ·  AI, honestly

What is AI genuinely good at in a business, and where will it embarrass me?

AI for business collapsed the cost of production and did not touch the cost of judgment. It is good at first drafts, research sweeps, summaries, code and variations at volume. It is bad at knowing what is worth doing, at taste, at your voice, and at verified facts, and it will invent a statistic with complete confidence. Everyone has access to the same models, so the edge is the workflow you built around them and the context you feed them, never the model itself. This business runs on Claude, and the part that took the work was not picking the tool, it was writing down the context and the process it gets handed every time.

  1. It collapsed the cost of production. It did not touch the cost of judgment. Everything that follows comes from that.
  2. Genuinely good at: first drafts, research sweeps, summarizing, code, data cleanup, repetitive production, translation, variations at volume.
  3. Genuinely bad at: knowing what is worth doing, taste, your voice, verified facts, and anything where being confidently wrong is expensive.
  4. Check everything before it touches a customer. It will invent a statistic, a citation, or a feature with complete confidence.
  5. It lowered the floor for your competitors too. Assume your product can be copied within weeks, because it can.
  6. Which pushes all the durable value toward the things it cannot copy: your brand, your distribution, your relationships, your proprietary data, and your judgment.
  7. Everyone has access to the same models. Your edge is the workflow you built around them and the context you feed them, never the model itself.
  8. Do not put customer data or credentials into tools you have not read the terms on.
  9. Disclose where the law requires it, and where a customer would feel misled if they found out. Those are two different tests and you have to pass both.
  10. AI handles routine support well and angry customers badly. Route emotion to yourself.
  11. Build so you can swap vendors. Anything hard wired to one company's feature set is a dependency you did not need.
  12. The real solo advantage: AI gives one person the output of a small team without the payroll, the management, or the coordination cost. That is the whole thesis of the one person business in 2026.
  13. The real solo risk: producing far more than you can judge. Volume without a filter is just faster noise.

Category 20 of 29  ·  Knowing what actually worked

Which three numbers do I look at every week, out of the thirty I could?

Pick three KPI numbers and look at them weekly. For most businesses those are traffic, conversion rate and revenue per customer, plus repeat rate if you have recurring income. Keep a plain weekly log of what you changed, because without it you will never know which change did what, and be able to say your one current constraint out loud.

  1. Pick three numbers and look at them weekly. Not thirty.
  2. For most businesses those are traffic, conversion rate, and revenue per customer. Add repeat rate if you have recurring income.
  3. Set analytics up before launch. Retrofitting attribution is miserable and mostly impossible.
  4. Attribution is imperfect no matter what you spend on it. Ask customers directly at checkout and weigh that heavily.
  5. Look at cohorts, not totals. Totals hide the fact that this month's customers behave differently to last month's.
  6. Do not optimize a step that is not the bottleneck. Find the constraint, fix the constraint, find the next one.
  7. Small numbers lie. Do not declare a winner off forty visitors, however much you want to.
  8. Track leading indicators, not just revenue. Revenue tells you about last month.
  9. Keep a plain weekly log of what you changed. Without it, you will never know which change did what.
  10. A dashboard nobody opens is worse than a notebook you actually write in.
  11. Know your one constraint at any given moment and be able to say it out loud.

Category 21 of 29  ·  Security, because you are the whole IT department

What do I actually have to do about security when I am the whole IT department?

Small business cyber security comes down to three things you do once. Two factor authentication everywhere using an app or a hardware key rather than text messages, a password manager with unique passwords, and a backup you have actually tested by restoring it. Keep the admin email separate from the public one, and lock the domain at the registrar.

  1. Two factor authentication on everything, using an app or a hardware key. Not SMS.
  2. A password manager, with unique passwords everywhere. This is not optional at this point.
  3. Lock the domain at the registrar and use a separate email address for that account only.
  4. Keep your admin email separate from the public one you put on your website.
  5. Phishing attacks target exactly the services you depend on and they look real. Never act on an urgent email about your account. Log in directly instead.
  6. Backups that you have actually tested by restoring one. An untested backup is a hope.
  7. Remove access the day a contractor finishes. Access sprawl kills quietly.
  8. Never keep payment credentials or API keys in shared documents or chat logs.
  9. Support impersonation is common. No real platform asks for your password or your two factor code.
  10. Have a plan for the day it goes wrong: who you contact, what you restore, what you tell customers.
  11. If you take payments, understand that a compromised account can cost you the ability to take payments at all, not just the money.

Category 22 of 29  ·  Cash and margins

Where is my money going, and is any of what I sell losing money?

Cash flow for small business kills more of them than profitability does, because processors hold funds, ads bill immediately and refunds come out today. Revenue is vanity, and what is left after fees, ads, tools, refunds and tax is the business. Know your monthly cost to exist to the dollar, know what each product actually leaves you after its own costs, and know your runway in months without having to check.

  1. Revenue is vanity. What is left after fees, ads, tools, refunds, and tax is the business.
  2. Know your monthly cost to exist, to the dollar. Most solo operators are bleeding a few hundred a month on tools they forgot about.
  3. Know the contribution margin on each product. Some of what you sell is probably losing money.
  4. Cash timing kills more small businesses than profitability does. Processors hold funds, ads bill immediately, and refunds come out today.
  5. Understand chargebacks and reserves before you have your first one.
  6. Set tax aside as the money arrives, in a separate account. Never at year end.
  7. Upfront payment and recurring revenue are the two things that let you sleep.
  8. Keep a runway number in months and know it without checking.
  9. Audit subscriptions quarterly. Cancel anything you did not open this month.
  10. Do not confuse one good month with a business. Look at rolling three month averages.
  11. Pay yourself something from the beginning, even if it is small. A business that never pays you is a hobby with paperwork.
  12. Watch the refund rate as a leading indicator of everything else.
  13. Solo means low fixed costs. Protect that. It is your entire resilience.

Category 23 of 29  ·  The legal and money setup

Do I need an LLC, a business bank account and bookkeeping before I start?

How to start an LLC or any business structure, plus a business bank account and a card that is not your personal one: do it before the first transaction, not eleven months later, because catching up costs more than doing it. Terms, a privacy policy and a refund policy go on the site before launch. Get an accountant early, since one conversation is cheaper than one mistake.

  1. Separate the business from you: a business structure, a business bank account, and a card that is not your personal one.
  2. Bookkeeping from the first transaction. Catching up eleven months later costs more than doing it does.
  3. Understand your payment processor's risk appetite for your category, and keep a backup ready.
  4. Sales tax and VAT on digital goods depend on where your customer is, not where you are. Use a platform that handles it or you will get this wrong.
  5. Contracts in writing for anything with a client or a contractor. A short clear one beats a long unread one.
  6. Know who owns the work. If you hire out design or code, get the ownership assigned in writing.
  7. Register the trademark on your name before it is worth stealing, not after.
  8. Terms, privacy policy, and refund policy live on the site before launch.
  9. Know where your customer data is stored and who else can see it.
  10. Get an accountant early. One conversation is cheaper than one mistake.
  11. Keep records. You will need them for tax, for disputes, and for the day you sell.
  12. If you operate across borders, get the residency and tax questions answered once, properly, by a professional.

Category 24 of 29  ·  The rules you can actually get hit for

Which rules can genuinely get me fined, and which ones is everybody ignoring?

FTC disclosure near the link, real consent with a working unsubscribe, and advertising claims you can substantiate are the ones that carry a real penalty. Testimonials have to be real and representative, and cherry picking the one outlier is a claims problem rather than a marketing choice. Advertising claims are the one that ends businesses, and platform policies bind you even where the law does not, enforced faster than any law.

  1. Email and messaging law: real consent, a real physical address, a working unsubscribe. Buying a list is the fastest way to destroy your sending domain.
  2. Privacy: consent for tracking, a way to delete data on request, and honesty about what you collect.
  3. Advertising claims: no income, results, or health promises you cannot substantiate. This is the one that ends businesses.
  4. Testimonials must be real and representative. Cherry picking the one outlier is a claims problem, not a marketing choice.
  5. Affiliate and sponsorship disclosure, clearly and near the link.
  6. Subscription rules keep tightening: clear terms at signup, easy cancellation, and renewal reminders in some places.
  7. AI disclosure requirements are arriving in more places every year. Know the rules where your customers are, not just where you are.
  8. Accessibility is a legal exposure on a website, not just a courtesy.
  9. Platform policies bind you even where the law does not, and they are enforced faster than any law.
  10. Licensing on images, music, fonts, and stock. Including the licensing terms of whatever AI tool made your assets.
  11. Age restricted, health, and financial topics carry their own rules. If you are near one, find out first.
  12. When genuinely unsure, pay a lawyer for one hour instead of guessing for a year.

Category 25 of 29  ·  Keeping customers

How do I keep the customers I already have instead of hunting new ones?

Customer retention is the cheapest revenue in the business, and most people ignore it and go hunting strangers instead. Onboarding decides it, so engineer one quick win early, because people who succeed once stay. Track churn and refund rate, ask people why they left and read the answers properly, and build your next product for the customers you already have.

  1. The cheapest revenue in the business comes from someone who already bought. Most people ignore this and go hunting strangers instead.
  2. Onboarding decides retention. The first week matters more than the next six months.
  3. Engineer one quick win early. People who succeed once, stay.
  4. Answer support fast and like a human being. As a solo operator this is a genuine advantage over any company with a ticket system.
  5. Track churn and refund rate. They are the truth about your product, and they are the only feedback that costs you money to ignore.
  6. Ask people why they left and read the answers properly. It is unpleasant and it is the best data you will get.
  7. Win back campaigns work and almost nobody runs them.
  8. Build your next product for the customers you already have. It is the shortest path to more revenue that exists.
  9. Community increases retention when it is genuinely active, and costs you time when it is not. Be honest about which one yours is.
  10. Surprise people with value rather than discounts. Discounts train, value bonds.
  11. One customer who refers is worth several who do not. Make referring easy and mention it.
  12. Fire the customers who cost more than they pay. Solo operators cannot afford to carry them.

Category 26 of 29  ·  You are the whole team

What do I hand off first when I am the whole team, and what do I never hand off?

Before working out how to hire a virtual assistant or anyone else, document the process you are handing over, or you will spend more time explaining than doing. Hand off what eats your week, and buy hours on a defined project rather than hiring for an undefined role. Never hand off your taste, your voice or your relationships, because those are the edge.

  1. You will be the bottleneck long after you expect to be. Plan around that rather than resenting it.
  2. Buy hours before you buy help. A contractor for a defined project beats a hire for an undefined role.
  3. Document the process before you delegate it, or you will spend more time explaining than doing.
  4. Delegate what eats your week, and never what is your edge. Your taste, your voice, and your relationships stay with you.
  5. Test any contractor with one small paid project first.
  6. AI is your first hire and it is cheap. Use it for research, drafts, code, editing, and admin. Keep the judgment.
  7. Batch your work by type, not by project. Solo context switching is the largest hidden cost in the whole operation.
  8. Protect a maker block every day where nothing is allowed to interrupt you.
  9. Have someone to talk to about the business. Isolation makes bad decisions look reasonable.
  10. Access control still applies with one person. Offboard contractors properly, same day.
  11. The failure mode nobody warns you about: you become the everything machine, do all of it badly, and stop shipping. Pick the two things that matter this quarter and let the rest be late.

Category 27 of 29  ·  Speed, and knowing when to quit something

How do I know when to quit something instead of pouring in another six months?

The sunk cost fallacy is the most expensive sentence in business, and the cure is a kill number and a date written down before you start, where you will see it. Ship the ugly version, because the market tells you things no amount of planning ever will. Most failures come from taking too long on the wrong thing, not from moving too fast on the right one.

  1. Ship the ugly version. The market tells you things no amount of planning ever will.
  2. Set the kill criteria before you start: what number, by what date. Write it down where you will see it.
  3. Sunk cost is not a reason to continue. It is the most expensive sentence in business.
  4. Ideas are worth nothing. Finishing is rare and that is where the value is.
  5. Work in weekly cycles with something visible at the end of each one.
  6. Make reversible decisions fast and irreversible ones slowly. Most decisions are reversible.
  7. One big bet at a time. Solo operators lose to their own optionality.
  8. Most failures come from taking too long on the wrong thing, not from moving too fast on the right one.
  9. Do not rebuild what is already working because you are bored with it.
  10. Keep a graveyard file of what you killed and why. It stops you restarting the same idea in eighteen months.

Category 28 of 29  ·  You, over the long run

How do I keep going for years without burning out or getting bored?

The solopreneur problem is rarely failure. It is boredom, and it is running out of runway while you wait. Plan in years, measure in months, work in weeks, know your runway in months, and manage energy rather than time, because there are maybe four good hours a day in you and they should not go to admin. Decide what enough looks like, in a number, before you start, or the goalposts move forever and you never get to notice you won.

  1. This takes longer than the people selling courses about it say. Plan in years, measure in months, work in weeks.
  2. Know your runway in months. Decisions made from panic are always worse decisions.
  3. Manage energy, not just time. There are maybe four good hours a day in you and they should not go to admin.
  4. Boredom kills more solo businesses than failure does. Build in something that keeps you interested.
  5. What compounds: your skill, your name, your body of work, your list, and your catalog. Protect all five.
  6. Comparison will cost you more than competition ever will.
  7. Learn in public. It builds the audience and the credibility at the same time, for free.
  8. One skill compounded for five years beats five skills started.
  9. Your health is a business asset when you are the only employee. Treat it that way.
  10. Decide what enough looks like, in a number, before you start. Otherwise the goalposts move forever and you never get to notice you won.

Category 29 of 29  ·  The business as an asset

Could I ever sell this business, and what would somebody actually pay for?

How to sell your business, or simply build one you could: buyers pay for what is transferable, documented and not dependent on you personally. Recurring revenue, diversified traffic, clean books, documented processes and an owned list raise the price. Everything running through your face, one traffic source, one big customer and messy records lower it.

  1. Online businesses sell. Even small ones. Yours might, and building as though it could changes what you build.
  2. Buyers pay for what is transferable, documented, and not dependent on you personally.
  3. What raises the price: recurring revenue, diversified traffic, clean books, documented processes, an owned list, and low customer concentration.
  4. What lowers it: everything running through your face, one traffic source, one big customer, and messy records.
  5. Keep the accounting real from day one. Nobody buys a business whose numbers cannot be verified.
  6. Contracts, trademarks, and asset ownership all get checked in a sale. Sort them while it is cheap.
  7. Know roughly what businesses like yours sell for. It changes how you think about a bad month.
  8. Build it so it could be sold, even if you never intend to. The same features that make it saleable make it liveable.
  9. A business that runs without you is not just an asset. It is a life you can actually enjoy while owning it.

Where to go next

Reading all 346 points end to end is not the point. Find the three categories you are weakest in and fix those. The diagnostic below asks one question per category, takes about four minutes, and names your weakest three along with the exact points that fix each one.

Two free calculators sit under the money categories: payback period and monthly cost to exist and runway. Both show the arithmetic rather than just the answer.