The idea pleased him a great deal. He spent two weeks on it: research, writing, a filming day, and three nights of editing.
He published it, and nothing moved. No views, no comments, not one question.
A month later, in a review session, he wrote the idea's title on a sheet and read it as if he had seen it in a list of suggestions. And he realised something simple: the title says nothing, and he would not click it himself.
Ten minutes at the start would have told him that.
Why test at all
Because the cost of the wrong idea is not measured in views but in the week you did not spend making a different one. You can execute a limited number of ideas a year, so every weak idea takes a good idea's place.
And testing is not a search for a guarantee, because there is none. It is removing the ideas that reveal themselves in ten minutes.
The rule: test in the cheapest form that carries the core
The cheapest form that carries the idea's core. And the core is the reason somebody would care, not the details of execution.
An idea resting on a surprising fact: its core is one sentence, and you test it with a sentence. An idea resting on a visual scene: its core is an image, and you test it with an image. An idea resting on a personal experience: its core is a story, and you test it by telling it in a minute.
Four tests ordered by cost
1. The title and first line test (ten minutes).
Write the idea as a real title the way it will appear, and under it the first line you will say in the video. Then close the file for a day.
The next day, read them while imagining you have seen them in a suggestions list among ten titles. Would you click?
The cheapest test in the world, and it rejects half of all ideas. And an idea that cannot carry a clear title is usually not an idea but a subject, and turning a subject into a filmable idea is set out in from a vague idea to a filmable one.
2. The audience question test (half an hour).
Post a question in the community tab or a story, and watch its wording: ask about the need, not their opinion of your idea.
Do not ask: "would you like a video about X?" The answer is always yes and it is worthless. Ask: "what is the hardest part of X for you?" or "has it ever happened to you that...?"
And what you measure is not the number of replies but how specific they are. Ten replies with details and stories mean this is a real point. A hundred likes with no words mean nothing.
3. The small publish test (one hour).
Take the idea's core and publish it in the simplest form: a written post, a one-minute short, or an image with text.
The strongest test, because it measures behaviour rather than words. And what you look at:
- Did anybody ask a question seeking detail? The strongest signal there is.
- Did anybody save or share it?
- Did a reply say "I needed this"?
And if you published a short and nothing moved, that does not necessarily mean the idea is bad; it means its core does not carry in a short form. Here you move to the fourth test rather than passing judgement.
4. The one-person story test (five minutes).
Sit with somebody from your target audience and tell them the idea in a minute with no preamble. Then stay quiet.
And watch two things: did they ask a question, and which question? The question they ask is what your video should answer, and if they ask nothing, the idea did not touch them.
This test is more precise than everything before it, and the least used, because it requires leaving your house.
What to measure in each test
| Test | What you measure | Sign of a pass |
|---|---|---|
| The title | Whether you yourself click a day later | An immediate yes |
| The question | How specific the replies are | Stories and details |
| The small publish | Behaviour: saves, questions, shares | A question seeking detail |
| The story | The person's question | A specific question, not courtesy |
Three traps in testing
1. Asking for an opinion rather than a need. People are polite about opinions and honest about a problem that hurts them.
2. Asking people close to you. Your mother and your friends will say the idea is excellent. Ask your audience or people who resemble them.
3. Reading silence as rejection. Silence may be bad timing or weak wording. Rerun the test once with different wording before you judge.
When testing is a waste of time
Do not test in four cases:
- The idea is cheap to execute. If it takes two hours, make it. Testing costs more than production.
- The idea is personal and you want to make it for yourself. That is not subject to a test; make it and move on.
- The idea is in an area where you know your audience thoroughly. Your confidence here rests on twenty previous videos, and it is more precise than a test.
- The idea is tied to a passing event. The time you spend testing is the time you would have published in.
What to do after the test
It passed: record it in your stock along with what the test produced. And what people said becomes material for the video: their questions become stops in the structure.
It failed but there is something in it: keep it and note beside it what you learned. Sometimes the idea is right and the framing is wrong, so you come back to it months later with another title.
It failed with nothing in it: delete it. A list nothing is deleted from becomes a graveyard, and the criteria for deleting are in when do you delete an idea.
It passed better than you expected: then this is not one idea but a series. And how to arrange your stock into lines is the subject of from scattered ideas to a series.
A half-hour session each month
Make testing a habit rather than an event. Once a month, take five ideas from your stock and write a title and a first line for each. Twenty minutes.
Then pick the strongest one or two and test them with a question or a small post that same week.
And after three months of this you have something more useful than a list of ideas: a list of tested ideas, and you know which one you are making next week without thinking.
In short
Before you spend a week on an idea, spend ten minutes on its title, half an hour on a question to your audience, and an hour on a small publish. And measure how specific the replies are rather than how many. And do not test what you can make in two hours, nor what you want to make for yourself, because some ideas need nobody's permission.