Skip to content
MyDSHMy DSH Field Notes
中文
Independently maintained; not affiliated with DeepSeek. No association with mydsh.dev or other MyDSH sites.

All categories/Marketing and growth

Marketing planning · Level Advanced · Marketing Ads Creative

Ad creative: three variables at a time

Creative underperforms and the cause is unclear.

Prompt

Three questions first: who, how do they solve it now, why switch. Without answers, creative is self-indulgence.

Task: Creative underperforms and the cause is unclear.

[Before you start, do three things]
1. Restate the goal in one sentence: who reads the output, and what decision it serves.
2. If I have not given you the following, ask me. Do not assume:
   - current creative and data
   - platform and bidding
   - combinations already tested
   - target conversion
3. Say how you plan to do it (at most 5 steps) and wait for my go-ahead.

[Go through each of these]
- fix audience to test creative, or the reverse
- one variable per round with enough sample
- track conversion, not just clicks
- say how much data makes a conclusion valid

[Process]
1. Confirm audience, scene and current alternative
2. State one testable hypothesis
3. Design the smallest test and its criterion
4. Allocate budget and channels with reasons
5. Say when this plan would fail

[Output]
Audience, hypothesis, test, allocation, failure conditions.

[Do not]
- No 'raise brand awareness' as a goal
- Never five channels at once
- No performance promises without a baseline
- Never split budget evenly

What to state before it starts

Leave these unsaid and the model fills them with an assumption. Most 'wrong answers' start here.

  • current creative and data
  • platform and bidding
  • combinations already tested
  • target conversion

Go through each of these

Where this kind of work most often goes wrong. Already in the prompt, so you do not have to remember it.

  • fix audience to test creative, or the reverse
  • one variable per round with enough sample
  • track conversion, not just clicks
  • say how much data makes a conclusion valid