Prompts to plan a campaign
Last updated
A campaign is a bet on language, placed across channels. The plannable part: mine what the audience already says, force genuinely different concepts (levers, not brainstorms), write the channel plan with kill criteria, and attack the whole thing before budget does. Four prompts, one sequence — for product launches, fundraising, internal change or events alike.
Order matters: audience language first. Concepts built before the mining step are guesses in nicer fonts.
The four templates
Before concepts: what do the people you must move actually say — donors, employees, customers, attendees?
Below is raw material from my audience: [REVIEWS / SURVEY COMMENTS / INTERVIEW NOTES / SLACK THREADS / FORUM POSTS] about [TOPIC/CATEGORY]. Extract: 1) the 7 most charged recurring phrases — quoted exactly; 2) top pains and top hoped-for outcomes in THEIR wording (no marketing translation); 3) the objections/fears that keep appearing; 4) the words THEY use for what we call [YOUR INTERNAL TERM] — the vocabulary gap; 5) 5 message hypotheses, each pairing one pain quote with one outcome quote. Rule: nothing invented — if the material is thin on something, say where. Material: [PASTE 30+ SNIPPETS]
You need 5 genuinely different campaign directions to choose from — not one idea in five outfits.
Campaign goal: [CONCRETE, E.G. 500 SIGNUPS / €50K RAISED / 80 % POLICY AWARENESS]. Audience: [WHO, FROM THE MINING STEP]. Their strongest pain/outcome quotes: [PASTE 3]. Constraints: [BUDGET CLASS, CHANNELS AVAILABLE, TONE LIMITS, TIMEFRAME]. Generate 5 concepts, one per lever: magnify (the outcome dramatized to its extreme), reverse (flip the category cliché), substitute (one real person's story carries everything), combine (pair the promise with the biggest objection, head-on), minify (the whole campaign in three words). Per concept: name, big idea in one line, hero execution, how the audience quote appears in it, and the risk that kills it. Use their quoted language in at least 3 concepts.
Concept chosen. Now: what runs where, what it costs, and — crucially — when you stop.
Act as my [CAMPAIGN LEAD]. Build the channel plan for: [CONCEPT ONE-LINER]. Goal + primary KPI: [NUMBER BY DATE]. Budget: [TOTAL]. Duration: [WEEKS]. Audience reachable via: [WHAT YOU KNOW]. Format — table: channel, job in the funnel, message angle (from the concept), format/asset needed, budget share, primary metric with week-2 target, KILL criterion (the number at which we stop this channel and reallocate). Then: 1) the asset production list with effort (S/M/L); 2) the 3 assumptions this plan bets on, each with its cheapest week-1 validation; 3) what we deliberately do NOT do, so scope creep has a wall. Ask me up to 3 questions first.
Everything's ready. Spend twenty minutes being your own worst critic — cheaper than week one doing it for you.
Attack my campaign plan as a skeptical [CMO / BOARD MEMBER / WORKS COUNCIL — WHOEVER APPROVES]. No compliments. 1) Reality check: what does this campaign ACTUALLY promise, and can we keep it? 2) Where does the logic break — which audience assumption, which channel bet? 3) What is the plan avoiding (a hard number? a weak asset? the fact that [X]?) — and what does avoidance cost? 4) What would a world-class campaign in this category do that we're not doing? 5) Priority list: what to CUT or fix before launch — not additions. 6) End with the one uncomfortable question about this campaign. Plan + key assets: [PASTE]
Frequently asked questions
Does this work for non-marketing campaigns?
Yes — the sequence is audience-agnostic: internal change programs mine Slack and survey comments, fundraising mines donor conversations, event campaigns mine last year's feedback. Only the placeholder content changes; the mining → levers → plan → critique spine holds.
Can AI predict which concept will win?
No — and distrust any confident prediction. What it CAN do is force real variety (levers), keep concepts tethered to audience language, and stress-test logic. Which concept wins is what the week-1 validations in the channel plan are for.
What if I have no audience material to mine?
Then that's the first campaign task, not a blocker: 5 short interviews, a survey with one open question, review-mining in your category. Concepts without any audience language are guesses — even 20 real quotes shift the quality visibly.
How detailed should the channel plan be pre-launch?
Detailed on the first two weeks (assets, targets, kill criteria), sketched after — you'll reallocate based on the validations anyway. Plans that schedule week 6 precisely are fiction with a spreadsheet.