Guides
Long-form reads on where ideas come from, what kills them, and what to do with the one that survives.
01 · Tech · Illustrated
AI business ideas you can run alone, where AI does the work you check
AI as the tool, not the product. 16 ideas where the customer pays for a checked result, and how often the AI-dependency objection was named: about 17% of the ideas our gate dropped.
02 · Starting a business · Illustrated
How to build a business that does not depend on you every day
The same business at 300 customers needs 34.5 support hours a week with hands-on onboarding and 6.9 with self-serve setup. Design that difference in before you build.
03 · Validation tools · Illustrated
How to write a business proposal, and when to screen the idea first
Summary, problem, result, scope, timeline, price, what you need and how to say yes: the 8 sections of a business proposal, plus the check a software project needs before you quote.
04 · Validation tools
A ChatGPT prompt that tries to kill your business idea
A prompt built from the 12 questions our own idea check asks, with rules that force PASS, FAIL or UNKNOWN, forbid invented competitors and prices, and end the idea on one FAIL — plus three ways to check the answer in five minutes.
05 · Side income · Illustrated
A freelance side hustle, and how to turn one repeat request into a product
Freelancers already see what paying clients ask for. Log every request, sell the one that repeats as a fixed package, and check it before you build a tool around it.
06 · Starting a business · Illustrated
How to run a business alone without the week breaking
We run WhittleOS as one person and publish our risk register: 10 risks, 1 open and 7 being tested. The rhythm, the limits and the preparations that keep a solo business running.
07 · Problems · Illustrated
How to spot a business opportunity in what people complain about
A trend tells you what is talked about; a complaint tells you what someone would pay to stop. Four signs of a real opportunity, and a dated look at our pain library, where the largest topic held 631 problems on 2026-09-29.
08 · SaaS · Illustrated
How to start a SaaS business alone, one check at a time
Pick a recurring job, reach the buyer without a sales call, do the price math, and sell the first subscription before the code. 7 stages, each with its stop.
09 · Starting a business · Illustrated
How to start a small business by selling before you spend
$2,000 spent before the first sale, at $20 profit per sale, is 100 sales before you have earned anything. The spending order that keeps a small business out of that hole.
10 · Side income · Illustrated
Money-making ideas: which are businesses and which are jobs
Most money-making ideas sell your hours: 13 of the 16 here do. What the same work can become as a business, from 4 of our lists, and the 3 that stay jobs.
11 · Startup ideas · Illustrated
Online business ideas you can run without sales calls
A niche store, a course, a membership, a productized service: 10 online businesses that sell in writing, and 6 that look online but start with a meeting.
12 · Side income · Illustrated
Passive income for beginners: avoid the product that needs you
With 200 customers, 2 minutes of help each is light and 30 minutes each is 23 hours a week. The support trap, and how to spot it before you build.
13 · Side income · Illustrated
How to make passive income online with a small product
Online income often stops because someone else controls the traffic. 5 sources of buyers, who controls each, and how to depend on them less.
14 · Side income · Illustrated
Passive income streams, sorted by how passive they really are
The streams that need the least work are the ones paid for with money. 16 streams sorted by upkeep, with what each still asks of you once it runs.
15 · SaaS · Illustrated
SaaS product examples, from household names to one-person products
10 horizontal products that do one job for any business, 8 vertical ones that do every job for one industry, and where the one-person versions sit.
16 · Starting a business · Illustrated
Service business ideas, and the unpaid work that comes with each
Most services pay from the first job. With 20 clients a month, the heaviest unpaid work in our table is 9.2 hours a week. 16 ideas and the work around each.
17 · Side income · Illustrated
Side business ideas that fit around a full-time job
An accountant building cash-flow models for nonprofits, a mechanic inspecting used cars on days off: 16 side businesses that reuse your day job's skill, and a 10-hour-week test for the price.
18 · Side income · Illustrated
The best side hustles for students who can code
Websites for clubs, bots for servers, data cleanup for labs, bug bounties. 15 side hustles for student developers, most starting from a group a student already belongs to.
19 · Starting a business · Illustrated
The most successful small business ideas, and why no industry is safe
Five-year survival across 19 US sectors, from BLS data: even the best sector lost 38.2% of its new locations. The starting habits we would copy, read from 5 famous failures.
20 · Validation tools
How to test demand with a landing page before you build
Four things decide whether a landing-page test means anything: one promise, a button that asks for a commitment, a pass mark written before launch, and enough visitors for silence to count. How to set each one up alone, with no audience and no calls, and where a fake door test stops being honest.
21 · Starting a business · Illustrated
34 easy businesses to start—and the first risk to test
Easy to open is not easy to sell. 16 local services and 18 desk-based hypotheses, the desk-based ones paired with the first relevant gate; a dated census of 1,728 screened candidates shows why that gate comes before setup.
22 · Side income · Illustrated
34 easy side hustles, from babysitting to small digital jobs
Easy means the first paid job is close and cheap to set up, not that buyers arrive easily. 16 everyday gigs and 18 digital routes show the boundary; 4 workload models show how follow-up can consume the week.
23 · Starting a business · Illustrated
20 home business ideas to test before you buy equipment
20 home business ideas across 4 kinds of home work, from a pre-ordered bakery batch to vacation-rental co-hosting. Each starts with one paid order, not a shopping list.
24 · Starting a business · Illustrated
How to write a business plan: prove who pays first
A plan earns its later sections only after the buyer section closes. Answer 6 questions, stress-test a 5,000-dollar monthly target across 4 visible price inputs, then run 12 deal-breakers.
25 · Starting a business · Illustrated
15 “most profitable” business ideas, ranked by customer burden
Price multiplied by customers is revenue, not profit. This dated public run supplied 26 price hypotheses; 15 distinct ones are ranked at a 5,000-dollar target, with 4 missing profit inputs kept visible.
26 · SaaS · Illustrated
15 one-person SaaS companies with public revenue
A dated list of 15 one-person software operations. 12 reported active MRR and 3 reported no active subscriptions; every row links to its payment-connected public page.
27 · Side income · Illustrated
18 side hustles from home, split by what you keep selling
The practical split is not online versus offline. It is whether the next sale needs another block of your labor. Start with 18 named routes, then stress-test support before calling one scalable.
28 · Starting a business · Illustrated
How to start a business from home without buying first
Home is an operating constraint, not evidence of demand. Use 12 live idea checks, 4 asynchronous paid asks, and a 5-step commitment order before you spend.
29 · Starting a business · Illustrated
How to start your own business when you plan to stay solo
A solo business is an operating contract, not a smaller org chart. Use 10 deal-breakers, 4 transparent support scenarios, and the dated platform-risk record before the first build.
30 · Starting a business · Illustrated
What business should I start? Use your constraints before a list
A 5-question paper filter drawn from the 9-field founder profile. It will not name a winning business; it will remove operating shapes your life cannot carry.
31 · Side income · Illustrated
How to make passive income by earning the low-work version first
A product becomes low-touch only after the work has been designed out. Compare 4 transparent operating models against a $3,000 monthly target; 3 land in the light support band under their stated assumptions.
32 · Starting a business · Illustrated
How to start a startup: the first month before a larger build
A 30-day startup plan should reduce uncertainty in order: complaint, hypothesis, costly action, delivery. The public sample run began with 69 candidates and kept 5; your first month should narrow too.
33 · Startup ideas · Illustrated
New business ideas: what builders launched in 2026, and who noticed
New ideas are not scarce. Of 11 product launches in the 2026 threads our runs read, 9 drew fewer than five points. Here is each problem, what was built, and what an idea needs besides being new.
34 · Side income · Illustrated
Passive income ideas, sorted by how much support they create
Six product shapes hold the same $49 price and $4,000 monthly target constant. Under the published planning assumptions, 3 cross into the heavy solo-support band.
35 · Side income · Illustrated
Side jobs versus a side business: choose the work you actually want
Side jobs can create income quickly but each sale usually needs another hour. Compare 6 skill routes and two transparent models for a $3,000 monthly target before deciding whether a reusable business fits.
36 · Startup ideas · Illustrated
20 Business Ideas That Survived a Filter Most Ideas Fail
Business ideas that came out of 25 real screening runs, not a brainstorm: each with the buyer, the price the run proposed, the verdict it got, and the test that would kill it.
37 · Starting a business · Illustrated
How to start a business: test the buyer before the paperwork
A business does not begin with a logo or an LLC. It begins when a specific buyer will take a costly action. Use the 12 idea checks, then do the paperwork the real operation needs.
38 · Starting a business · Illustrated
How to start an LLC: check the idea before you file
An LLC can be the right structure and still sit around a business nobody wants. Run 4 market questions first, then use official state guidance for the filing and the full 12-check screen for the idea.
39 · Side income · Illustrated
Side hustle ideas that can turn into a product
A side hustle becomes an asset only when income can recur without every sale repeating your labor. Compare 4 stress-test shapes against a 8-hour week, then test the buyer before building.
40 · Startup ideas · Illustrated
What is a startup? The point where an idea meets a buyer
An idea becomes a startup when you organize a real search for a repeatable business. Across 33 retained runs, the gate stopped 51.8% of candidates before deeper scoring; payment is the first signal that reaches price.
41 · SaaS · Illustrated
What is SaaS? A one-person product from buyer to support
Software becomes SaaS when the customer keeps receiving a service and the operator keeps carrying an obligation. SoloPipe makes the definition concrete; a 10-minute support assumption exposes the operating side.
42 · Validation tools · Illustrated
Will people actually pay? The signals that count before you build
Every signal you can collect before building costs the person giving it something, and tells you about as much as it cost. The ladder from praise to payment, what each rung can and cannot answer, and why our own sample run stops one rung short.
43 · Validation tools · Illustrated
An AI rejected your idea. How to tell whether it is right
Before you argue or give up, find out which check failed, whose constraints it was applied to, and what would overturn it. On our own published cases, the same one-liner came back with opposite verdicts under two different founder profiles.
44 · Validation tools · Illustrated
A year of access, or one check: what a part-time founder is buying
Their entry plan is a year. If what you have is one idea and six hours a week, the question is not how many researched ideas you can browse — it is how many real checks you get and when you must commit.
45 · Problems · Illustrated
You built the app and nobody pays. Which of the three is it?
Shipping got cheap and demand did not. Three symptoms hide under one complaint, and the published report on a product of exactly this shape names the two areas that decide it — neither of which is a feature.
46 · Validation tools · Illustrated
How to compare idea validators without fooling yourself
Most comparisons in this category put a free score next to a paid report and call the difference quality. Here is what has to be held constant, why a vendor cannot grade its own competitors, and our result on one idea in full.
47 · Validation tools · Illustrated
Choosing between three ideas when you have ten hours a week
When the scores are a point apart, the number has stopped being the answer. Here is the same ordering function our Compare mode runs, applied to three published cards, with the row it puts first and the row it says to act on.
48 · SaaS · Illustrated
Is your SaaS solving a one-time problem?
A one-time problem billed monthly is an acquisition treadmill wearing a subscription badge. Here is what standing still costs at each level of repeat, and how to tell which shape you have before you build the billing.
49 · Validation tools · Illustrated
From a complaint to a customer: where the two products diverge
A ranked complaint tells you somebody is annoyed. Whether they would pay, whether you can reach more of them, and whether one platform can end it are separate questions — and they are answerable before you build.
50 · Validation tools · Illustrated
Preuve AI and WhittleOS: two published answers to a weak result
The lazy comparison says only we show evidence, and both products advertise it. The real difference is the default action on a weak result: rework the angle, or end it on a failed check. Both are defensible.
51 · Validation tools · Illustrated
A long report or a kept decision: which one you are buying
A long report transfers the analysis to you and keeps the decision. A verdict commits, and is therefore capable of being wrong in a way you can point at. Neither shape is better in general.
52 · Validation tools · Illustrated
Screening a client's idea before you agree to build it
Delivery risk is the one you priced. Demand risk is the one that costs you the second project — and most of it is answerable in an afternoon, from questions somebody else already wrote down.
53 · SaaS · Illustrated
Can your SaaS sell without demos or sales calls?
If you will not take sales calls, some ideas are simply not yours — and you can tell which before building. Four of our twelve published checks settle it, and the scanner that keeps our own output call-free shows what the habit sounds like.
54 · Problems · Illustrated
A rejection dataset, and the exclusions that make it mean anything
Most of the work in a census is exclusion, and exclusions are where a published figure quietly becomes wrong. Here are ours, in order, with what each one costs and a check that the recipe still lands on the totals.
55 · SaaS · Illustrated
Support load by vertical: same price, very different weeks
Support load is mostly a property of what the product touches, not of how well it is built. Six verticals at one price, with the estimates published alongside their reasons so you can argue with any row.
56 · Validation tools · Illustrated
Validating an idea on six to ten hours a week
At eight hours a week, a product in the middle support band consumes the entire budget before you have built or sold anything. Here is what that leaves, and a two-week sprint whose hours are computed from the budget rather than wished into it.
57 · Validation tools · Illustrated
Before you pay an agency to build your MVP
An agency will build what you specify, on time, to spec. None of that touches the three things that kill most ideas — and all three are decidable before the contract is signed.
58 · Validation tools · Illustrated
Validating with no audience and no ad budget
With nobody to show it to, the first question is not what to test but how many people have to see it before the answer means anything. The grid is larger than it feels, and your analytics will understate it.
59 · Validation tools · Illustrated
When a free idea score is enough, and when it stops being
Below the line where a decision has a cost, precision is waste and a free read is the right instrument. Above it, three specific things start to matter, and none of them is a bigger number.
60 · Validation tools · Illustrated
What idea validation costs, across the tools that sell it
Unit prices make these tools look comparable and they are not: one bills annually, one bills monthly, most sell a one-off report. What a buyer actually pays is the minimum cheque, and the range across the category is wide.
61 · SaaS · Illustrated
Buying a micro SaaS: what the listing does not tell you
A running business answers fewer of the idea questions than it looks like: trading history is evidence about the past, and half the list is about the future. The first one it does answer is hiding in plain sight in the price.
62 · Problems · Illustrated
When to shut it down, decided before you are attached
Nobody makes this decision well in the moment, because by then the sunk cost is you. The workable version is a condition written in advance, in numbers, where somebody else can see it.
63 · Startup ideas · Illustrated
50 small business ideas, tagged by the time and calls they need
A list of 50 small-business hypotheses with a buyer, editorial fit tags, a public evidence pool, a costly first action and a reason to stop before building.
64 · Startup ideas · Illustrated
E-commerce startup ideas from merchant pain — and which ones we would kill
A complaints-first e-commerce idea list that keeps the off-topic rows, native-platform baseline, and rejected concepts visible instead of manufacturing eight winners.
65 · Startup ideas · Illustrated
Low-cost startup ideas for solo developers, ranked by time to evidence
Low cost should mean cheap to disprove, not cheap to code. These real idea cards are ordered by the first buyer action worth waiting for.
66 · Problems · Illustrated
Search intent vs. search volume: how we chose what not to publish
We put three tempting keywords through four checks. One was dropped, one became an experiment, and the smallest remained a conversion asset.
67 · Tech · Illustrated
AI startup ideas: the wrapper question is not what kills them
Every AI idea gets the same objection — a vendor will ship this as a feature. It is a fair question and it is not the one that does the killing. On the run we publish unedited, the missing problem behind the idea accounted for far more of the drops.
68 · Startup ideas · Illustrated
Am I the right founder for this idea? What the answer costs
Founder-market fit is usually discussed as credentials or encouragement. The decidable version asks whether you can build, explain, sell and support this idea — and in this product it is the one analysis where a fact about you, not the idea, can end it.
69 · Startup ideas · Illustrated
Business ideas for agencies: where to look without inventing demand
A complaints-first method for finding agency software problems without pretending that one-off records prove three or five ready-made product ideas.
70 · Startup ideas · Illustrated
Business ideas for coaches: what software can and cannot solve
The evidence splits renewal work into four parts. Two are mechanical enough for software; two remain the coaching relationship itself.
71 · SaaS · Illustrated
How to price a micro SaaS: the model, the floor, the first number
Pricing advice for small software is enterprise theory or "charge more" as a slogan. The solo question is different: is there any price at which one person can carry the customers that price requires? At $3,000/month the answer has a floor.
72 · Startup ideas · Illustrated
Your idea is too broad. What a real narrowing has to contain
Every founder reaches "it is too broad" on their own. The useful part is what comes next: what a narrowing has to contain to count as one, why the same idea narrows for one founder and refuses for another, and what an honest refusal looks like.
73 · SaaS · Illustrated
What to build first — and the case where you should not be asking yet
Every tool answers "what should I build first" immediately. This one checks whether you have earned the question, and refuses in most of the states you can be in — including the one you are in right now if you have not evaluated the idea yet.
74 · SaaS · Illustrated
Will your SaaS be support-heavy? It is decidable before you build
The answer follows from the workflow, not from usage data you do not have yet. Here is the ladder from background noise to "support is the job", the one call that matters most, and why the scale on this analysis runs backwards.
75 · Validation tools · Illustrated
Do startup idea validators work? We tested ours and published the misses
Vendors say yes, forums say no, and nobody has run a test. We ran 8 real companies with known endings through our own gate under two founder profiles — it caught 4 of 5 flops, and the honest conclusion is still not an accuracy percentage.
76 · Problems · Illustrated
Famous startup failures, and the shape they share
The most quoted failure statistic is undated, unlinked and useless at founding. Here are five companies whose stories are over, the one-liner each could have been described by on day one, and the property 3 of them shared.
77 · SaaS · Illustrated
How many customers you need, and how many one person can carry
Divide the target by the price and round up. Then check the number that decides whether it works: at 15 support-minutes per customer a month, one person carries about 139 customers before support passes 8 hours a week.
78 · Validation tools · Illustrated
How to score a business idea, with the weights published
Every guide on scoring ideas tells you to pick criteria and assign importance. None of them says what they picked. Here is the rubric this product scores with — all 7 weights, the arithmetic, and the part a score cannot settle.
79 · Problems · Illustrated
Is my startup idea already taken? The wrong question, and the one that works
Half the results reassure you that competition is validation and the other half sell you a tool. Neither gives you a method, because yes-or-no cannot be acted on. Here is the question that can, and the 3 decisions it produces.
80 · Problems · Illustrated
Platform risk, counted rather than warned about
Everyone warns about building on someone else's platform and nobody says how often it is actually fatal. Across 706 candidates our gate killed, platform dependency named 19.8% of them — and it is the only check we softened on purpose.
81 · Validation tools · Illustrated
How to validate a startup idea: the checklist, in cost order
Most checklists list good questions in no particular order, so people answer the interesting ones first. Here are the two lists this product actually kills ideas on, what each costs to answer, and which one did the most damage to 69 real candidates.
82 · Validation tools · Illustrated
Validating a startup idea without talking to customers
A conversation gives you an opinion you grade afterwards. A pre-registered numeric bar gives you evidence you cannot re-grade. Here is the method, the five things it can measure, and the much shorter list of what a pass actually proves.
83 · Problems · Illustrated
Where to find real customer problems, and what makes one quotable
The stock answer names a website. The useful answer names a property: can you open the thing the claim rests on? Here is what one published run actually read, the five tiers we grade a source by, and why 43% of its problems carry no link at all.
84 · Validation tools · Illustrated
Why AI says your startup idea is great
Ask a chatbot about your idea and it will help you. That is the failure: helpfulness and judgement point in opposite directions. Here is the mechanism that produces the flattery, and what one dated run of our own idea through 3 tools returned.
85 · Problems · Illustrated
Why startup ideas fail: 36 rejections, itemised
Most idea lists tell you what to build. This is the other half: 36 ideas from one published run, itemised by the check that killed them — and the reason those counts add up to more than the number of dead ideas.
86 · SaaS · Illustrated
Micro SaaS ideas, each tied to a recorded problem and its source
13 micro SaaS directions, one per market, each tied to one of the 1,823 problems our sweeps recorded. Then the two numbers that decide whether one person can run any of them.
87 · SaaS · Illustrated
SaaS ideas: where they actually die
The hard part of a SaaS idea is rarely building it. Here is where they actually die, and the questions that surface it before you spend six months finding out.
88 · Startup ideas · Illustrated
Where startup ideas actually come from
A list of startup ideas is worth less than one idea you can defend. This is how candidates get sourced from real complaints, and what survived a 30-market sweep.
89 · Reddit · Illustrated
Finding startup ideas on Reddit, and what that method can and cannot tell you
Complaint threads are a real source of documented problems and a bad source of conclusions. Here is the method, its failure modes, and which sources our published run actually read.
90 · Tech · Illustrated
Tech startup ideas and the question that now decides them
Technical difficulty stopped being the moat. The question that decides a tech idea now is what remains of it when the capability underneath becomes ordinary.
91 · Y Combinator · Illustrated
Y Combinator startup ideas for Fall 2026, re-read for one person
The same idea can be a line-item for a funded team and a dead end for one person. Where that line falls, which barriers money actually clears, and what happens when the current Fall 2026 list is run across it.
92 · Validation tools · Illustrated
Best startup idea validation tools (2026): what actually matters
Don't pick by feature list. The idea-validation tool worth using is the one that shows its work — it links evidence to a real source, is willing to tell you no, and binds the verdict to your situation.
Comparing tools rather than reading long-form? Check a startup idea against 12 deal-breakers goes through the checks one at a time, and answers covers the short questions.