On this page
This is a production worksheet. It creates comparable creative versions, not three guaranteed winners.
Check whether the opening is the problem
Watch the current video without looking at performance numbers first. Is the opening understandable? Does it lead into the product action? Is the result readable? A weak middle or an unrelated destination can make an opening change irrelevant.
If you have performance data, compare the same definitions and similar delivery conditions. A drop in results can come from audience, placement, offer or other changes. Do not label every decline “creative fatigue” and replace footage blindly.
Lock the part that explains the product
Choose a recording that shows one complete action and its result. Save a clean version of that section with its caption and sound decisions. Duplicate the project before adding new openings.
Write down the offer and closing instruction as well. If those change between versions, you will not be comparing only the opening. Keep the destination consistent when evaluating the edits under similar conditions.
Create three genuinely different entries
Use a question, a result-first shot and a reaction-led opening. These are different structures rather than three nearly identical captions. They can all lead into the same product action if that action has a clear payoff.
| Version | First shot | Suggested line | Connection to the demo |
|---|---|---|---|
| A: question | A curious reaction | “Where did I save that?” | The demo performs the search |
| B: result first | The correct saved item already open | “Here is the note” | The demo shows how it was found |
| C: small shortcut | A restrained surprised expression | “There is a shorter way” | The demo shows the supported path |
Refreshing a tired opening: change one thing
If an existing ad needs a new start rather than three fresh ones, decide what feels wrong before exporting anything. A wording change tests a different way into the action. An expression change adjusts tone. A result-first cut changes the structure.
Make the cut replaceable: keep the opening text on its own layer, tie its end to the product cut and save a clean version first. When you swap the reaction, set a new in-point instead of freezing or stretching the expression to fill the old slot.
| What feels wrong | Try changing | Keep fixed first |
|---|---|---|
| The promise is vague | The opening line | Reaction and demo |
| The tone is exaggerated | The expression | Message and product action |
| The result arrives too late | Opening structure | Offer and destination |
| The face crop feels awkward | Composition or source clip | Text and timing where possible |
| The screen is unreadable | The demonstration itself | Do not call this only a hook refresh |
Make the handoff feel deliberate
Each opening may need a different in-point even if the demo stays the same. Leave a clear beat for the question, then bring in the product action. Do not force a long expression into a short slot by making the motion unnatural.
For the result-first version, avoid repeating the entire result so many times that the middle feels redundant. Show enough to create context, then use the demonstration to explain the process. If that changes the total length, note it as part of the treatment.
Keep a version sheet with the exports
Record the exact source, line and timing. A teammate should be able to reproduce the version without searching old messages. Keep a baseline export so you can inspect accidental changes in crop, audio or text.
| Field | What to record |
|---|---|
| Version ID | A question / B result / C reaction |
| Source | Reaction filename or product source |
| Opening duration | Actual time before the fixed demo |
| Text | Exact line and where it ends |
| Changed variables | Everything different from the baseline |
| Review | Phone-size check and any issues |
Know what this comparison can tell you
If the expression, line and structure all differ, you are comparing opening packages. That can help choose the next version, but it does not isolate which small detail caused a result. To study the wording alone, run a narrower comparison with the same source and timing.
Use the result to decide what to make next, and keep uncertainty visible. A few early clicks or views do not establish a universal rule about a face, word or hook type.
Questions & answers
Do I need three different creators?
No. You can vary the line or structure with the same creator. Choose differences that match the question you want to answer.
Are three versions enough to prove a winner?
The number of exports does not establish statistical reliability. Results depend on delivery, sample size and the size of the difference.
Does a fresh opening guarantee better performance?
No. It gives you another creative version to evaluate. Results still depend on the offer, audience, product evidence and delivery.
Find the opening for your next edit.
See the public samples, then unlock the library when you’re ready to work with the originals.
Explore the library