You are viewing v1 of 1. Go to the current version · All versions

Ad copy variants within platform limits

Generate distinct, honest ad variants for Google, Meta or LinkedIn that respect character limits and test different angles.

Change note: Initial version

Prompt

Write {{count}} distinct ad variants for {{platform}}.

Product and offer: {{offer}}
Audience: {{audience}}
Primary angle to emphasize: {{angle}}
Must include: {{must_include}}
Must avoid (claims, words, topics): {{avoid}}

Constraints
- Respect these limits strictly: {{limits}}. Count characters for every line and show the count in brackets.
- Each variant must test a different angle (for example: pain point, outcome, social proof, price or risk reversal, curiosity). Label the angle.
- Only make claims that follow from the information above. Do not invent numbers, reviews, or deadlines.
- No clickbait, fake urgency, or misleading promises. Plain language over buzzwords.

Output a table with columns: #, Angle, Headline, Description or body, CTA, Character counts. After the table, add three sentences on which two variants to test first and what result would tell me which angle wins.

Variables

  • {{count}}Number of variantsnumber

    Default: 6

  • {{platform}}Platformselect

    Options: Google Search ads, Meta feed ads, LinkedIn sponsored content, Reddit ads

    Default: Google Search ads

  • {{offer}}Product and offerlong

    Default: Plotline, project planning software for design studios. 14-day free trial, no card required.

  • {{audience}}Audiencetext

    Default: Agency owners and project leads at small design studios

  • {{angle}}Primary angletext

    Default: Avoiding overbooked weeks and missed deadlines

  • {{must_include}}Must includetext

    Default: The words 'free trial' and 'no card needed'

  • {{avoid}}Must avoidtext

    Default: Superlatives like 'best', and anything implying guaranteed results

  • {{limits}}Character limitstext

    Default: Headlines up to 30 characters (3 headlines), descriptions up to 90 characters (2 descriptions)

Example output

| # | Angle | Headline | Description | CTA | Counts |
|---|---|---|---|---|---|
| 1 | Pain | Overbooked Again? [21] | See clashes before deadlines slip. Free trial, no card needed. [60] | Start free trial | ok |
| 2 | Outcome | Plan Without Clashes [20] | Build your schedule from the real team calendar. Free trial. [59] | Try it free | ok |
| 3 | Social proof | Used by 140 Studios [19] | Design studios plan projects in Plotline. No card needed. [57] | See how it works | ok |
| 4 | Risk reversal | Free Trial, No Card [19] | Plan one real project in 14 days. Cancel nothing, pay nothing. [62] | Start free | ok |
| 5 | Curiosity | Where Is Your Team Free? [24] | One view shows who is stretched next month. Try it free. [57] | See the view | ok |
| 6 | Comparison | Beyond the Spreadsheet [22] | Replace project-plan spreadsheets with a live schedule. [55] | Start free trial | ok |

(The table shows the first headline and description for each variant; the full response would list all three headlines and both descriptions per variant.)

Test first: #1 (pain) and #4 (risk reversal). If #1 wins on click-through and trial starts, the problem framing resonates; if #4 wins, the barrier is commitment, so lead with the trial terms in other channels too.

Notes

Stating character limits and counts forces compliance, and labeling the angle makes results interpretable. Check final counts in the platform editor, since models can miscount.