Free business idea generator for a stronger first draft.

The inputs here are a person rather than a market. Nine years as a veterinary nurse, mostly small animal practice, with rota and triage experience. The buyer is independent single site small animal practices, the kind with three or four vets and no corporate group behind them. The costly problem is post operative follow up: someone has to ring every discharge the next day, it is clinically useful, it is nobody's dedicated job, and on a busy day it either gets done badly by a nurse who should be in theatre or it does not get done.

The idea that comes out is written as five lines rather than a pitch. Buyer: independent small animal practices within one county, ten to forty routine procedures a week. Existing cost: roughly an hour of qualified nurse time a day, currently absorbed rather than budgeted, plus the occasional complication caught late. Offer: outsourced next day post operative follow up calls made by a registered veterinary nurse, with a written note back into the practice record and an agreed escalation rule for anything concerning. Price shape: per procedure rather than a retainer, so a quiet week costs the practice nothing. Cheapest test: two practices, one week, done manually by phone with the notes typed into an email, no software at all.

That is a testable idea because every line can be checked against reality within a fortnight. The buyer is narrow enough to name actual practices. The cost already exists and is already being paid in nurse hours, so the conversation is about moving a spend rather than creating one. The capability is one she has on Monday, without hiring or building anything. The delivery model is deliberately manual, because doing it by hand is how you find out what the call really needs to cover before anyone writes software. And the price shape removes the objection you would otherwise spend the first meeting arguing about.

The idea most people write down instead is an app for pet owners to track their pet's recovery, with reminders, photo uploads and a vet messaging feature. It sounds bigger, and it is worse in every practical respect.

Take it apart line by line. The buyer is unclear: owners would use it and practices would pay for it, and those are two different sales. The cost being replaced is not identified, so nobody can say what is currently spent on this. The capability does not exist yet, so the first six months are a build rather than a test, and the build has to happen before anything is learned. The delivery model needs adoption on both sides, owners and practices, which is two adoption problems rather than one. And there is no cheap way to be wrong, because the test is the product. The nurse can run her version next Tuesday with a phone and a notebook and know by Friday whether practices will pay. The app founder finds out in a year.

Stop generating ideas when a customer has already asked you for something. The most reliable signal in early business is an actual request from someone with a budget, and it is remarkable how often people set that aside to brainstorm. If someone has asked whether you can do a thing, and you can, the next step is a quote rather than a list. Idea generation is for when you do not have that, not for when you would rather not do the awkward part of the thing you already have.

Do not use this when the bottleneck is capital, time or a legal constraint rather than imagination. A new list will not tell you which idea you can afford, and it makes the constraint feel less urgent while changing nothing about it. Write the constraint down first, then generate against it: ideas that need no equipment, ideas that fit eight hours a week, ideas that do not need the licence you do not have. An idea that fails your actual constraint is not an idea you have, it is one somebody else has.

Be honest about idea generation as avoidance. If you have produced four rounds of ideas and made no calls, the problem is not that the ideas were not good enough. Testing is uncomfortable in a way that generating is not, because generating cannot fail and a phone call can. The fix is not a better list, it is taking the least bad idea from the list you already have to ten conversations. Most ideas look about the same on paper and only separate once a real buyer reacts to one.

And take extra care in regulated or licensed areas. Anything touching healthcare, finance, food, childcare, transport or professional advice can be a good idea and still be one you cannot legally test the quick way. Before you spend a week on a plan, find out what registration, insurance or qualification stands between you and the first paying customer. That is not a reason to drop the idea, but it changes the cheapest test from ten conversations to a call with the regulator, and knowing that early saves you a plan built on the wrong assumption.

Most written ideas fail because they describe a product rather than a transaction. A usable one names five things: who the buyer is, specifically enough that you could list ten of them by name; what the problem costs them today, in money, hours or risk they already carry; what you would sell, described as the thing they receive rather than the technology behind it; how it is priced and billed; and what the fastest disproving test is. Five lines is about the right length, because anything longer starts smuggling in assumptions you have not checked.

The line people skip is the existing cost, and it is the one that does the most work. If you cannot say what the buyer spends on this problem today, whether that is a subscription, a part time person, an agency retainer or an absorbed hour of expensive staff time, you are not selling into a budget, you are trying to create one. That is a harder sale and a slower one. It can still be the right idea, but you should know that is what you have chosen rather than discovering it after ten meetings that all end in vague enthusiasm and no order.

The instinct after a good idea is to plan the whole thing: the pricing tiers, the name, the site, the systems. Almost none of that survives the first ten real conversations, so building it first is buying answers to questions nobody asked. What you need instead is the smallest thing that would change your mind. For a service, that is usually delivering it manually to one or two customers and charging for it. For a product, it is often a specific offer made to a specific list, where the meaningful outcome is a deposit rather than an expression of interest.

Write the test down with a pass condition before you run it, because otherwise you will read whatever happens as encouraging. Two out of five practices agreeing to a paid trial is a pass. Five polite conversations where everyone says it sounds useful is a fail dressed as a pass. The most common wasted month goes on a test with no threshold, which produces evidence that can be read either way and therefore changes nothing. Decide in advance what you will do if it fails, too: which line of the idea you would change first, and whether you would change the buyer, the problem or the delivery.

An idea is a sentence. An opportunity is an idea where you have evidence that a specific group of buyers has the problem and pays to deal with it today. A business is an opportunity plus a repeatable way to reach the next buyer at a cost lower than what they pay you. Confusing these stages is what makes early planning documents so unreliable, because a full plan written at the idea stage is mostly fiction with a spreadsheet attached, and the spreadsheet lends the fiction unearned confidence.

Different work belongs at each stage. At the idea stage the work is conversations and cheap tests, and the output is a decision about which ideas to drop. At the opportunity stage the work is delivering to a handful of paying customers by hand, and the output is knowing what the service really involves and what it costs you to provide. Only at the third stage does the question become how you reach the next twenty without knowing them personally, which is when marketing spend, systems and hiring start to make sense. Most early businesses that stall are stuck between the second and third, doing good work for people they happened to know and with no answer for where the next one comes from.

Generating ideas is easy; holding a dozen open indefinitely is what does the damage, because each one takes a little attention and none of them get the sustained effort that would resolve them. Give each idea a date and a decision. Two weeks of conversations, then keep or drop, with the criteria written down before you start. An idea that survives three of those rounds has earned real time; an idea that has sat on the list for eight months without a test being run has answered its question already.

Do keep the dropped ones, with a line saying why they went. Ideas fail for reasons that expire: the buyer had no budget that year, the capability was one you did not have, a supplier would not deal with someone your size, the regulation was about to change. Circumstances shift, and a written record of the specific blocker lets you check whether it still applies rather than re-deriving the whole idea from scratch. It also stops you regenerating the same idea every few months and feeling briefly excited about it, which is a surprisingly common way for a list to eat a year.