fairytale video
Make an ad
Test one creative change in an AI ad
A useful creative comparison starts with one written question. “Which ad is better?” leaves too much undefined. “Does showing the clasp before the creator speaks make the bag’s mechanism clearer?” gives you a change you can describe and a baseline you can preserve.
Record the baseline before changing it
For a hypothetical bag campaign, baseline A begins with the creator holding the bag, then shows its photographed clasp. Version B begins with the clasp already visible. Keep the product variant, creator, scene, approved line, duration and closing instruction the same in the written briefs.
A generated version can vary unintentionally even when those settings remain constant. Review both outputs for product identity, speech and movement before treating them as a meaningful comparison. A new face or changed label is a production failure, not the opening-shot variable you intended to examine.
A small experiment card
- Question: is the clasp understandable earlier with a product-first opening?
- Planned change: the first visual state only.
- Held constant: product, references, words, duration and destination.
- Creative check: reviewers can identify the clasp without an explanation.
- Campaign observation: the metric and comparison period chosen in your campaign tools.
- Possible next decision: retain, revise or gather more evidence.
This is a planning illustration, not a guarantee of improved clicks or sales. Choose the campaign setup separately. fairytale video creates the clips; it does not allocate audience exposure or run a controlled media experiment for you.
Separate production approval from campaign evidence
First confirm that each file is usable. A version with an added spoken claim should be rejected even if it later attracts attention. Then document the conditions under which the approved versions were published: dates, placement, audience setup, destination and spend where applicable.
If those conditions differ, describe the result as an observation with those differences attached. Avoid claiming the opening caused a change just because the second video ran later. A promotion, different distribution or a changed destination can affect the comparison independently.
Decide what the result actually permits
Suppose both versions receive very little exposure. The useful conclusion may be that there is not enough evidence to choose, while a creative review still suggests the clasp is easier to recognize in B. Keep those two findings separate. One concerns visual comprehension; the other concerns campaign response.
Preserve both finished files and their production inputs. A new comparison is another deliberate paid render when new generation is necessary. Use variant records to keep the baseline traceable, a results worksheet to interpret observations and a refresh plan to choose the next single question.