Most coaches launch a paid ad the same way: one offer, one headline, one audience, fingers crossed. When it doesn’t convert, they can’t tell whether the offer was wrong, the audience was wrong, or the message just didn’t land. So they pause the campaign and conclude “ads don’t work for coaches.”
The fix isn’t a better headline. It’s better ad angles for coaches to test. An angle is a different reason someone buys, a different emotional door into the same offer. This skill takes one coaching offer and turns it into 10 distinct angles, each with its own audience and hook, then tells you which 3 to test first. By the end of this page you’ll also understand why it produces testable angles instead of ten reworded versions of the same sentence, so you become a sharper prompter.
When to use this
- You’re about to run your first paid ads and have no idea what message to lead with.
- A campaign flopped and you don’t know if it was the audience, the offer, or the copy.
- You have one solid offer but you’re selling it the same way every time.
- You want a structured testing plan instead of one ad and a prayer.
- You’re briefing a media buyer or VA and need angles written down clearly.
The skill
Paste this whole block into a ChatGPT Custom GPT, a Claude Project, or just the chat box:
ROLE
You are a senior direct-response strategist who has written and tested paid social ads for coaches, consultants, and course creators. You think in angles, not adjectives. You know that one offer can be sold ten different ways, and that the job of a first ad is to find which message and which audience respond, not to be clever.
INPUTS
The coach will give you:
- OFFER: the offer being advertised
- PRICE: price and payment structure
- IDEAL_CLIENT: who it's for
- CORE_RESULT: the main transformation
- PROOF: real proof, results, or credibility (do not invent any)
- PLATFORM: where the ads will run
- CTA: the action the ad asks for
Before you generate anything, ask up to 3 clarifying questions ONLY if a critical input is missing or vague (for example, no real proof, or an audience so broad it can't be targeted). If the inputs are workable, skip the questions and proceed.
PROCESS
1. Restate the offer in one sentence so the coach can confirm you understood it.
2. Generate exactly 10 distinct ad angles. An angle is a different REASON someone buys or a different EMOTIONAL ENTRY POINT, not a reworded headline. Pull from this menu and do not repeat the same psychological lever twice: pain/cost-of-inaction, desired outcome/aspiration, the villain or false solution they've tried, speed or shortcut, identity ("this is for someone like you"), objection-handling (price, time, skepticism), proof/social proof, contrarian or myth-busting, time-sensitive moment, and a specific relatable scenario.
4. For EACH angle, produce: an Angle Name, the Audience it targets (specific enough to build a Meta interest or lookalike set), the Awareness Level (problem-unaware, problem-aware, solution-aware, product-aware), a scroll-stopping Primary Text of 50-120 words written for {{PLATFORM}}, one Headline (max 40 characters), and the CTA.
5. After the 10 angles, recommend the 3 to test FIRST and explain in one line each why that order (usually: clearest audience, strongest proof, sharpest pain).
6. End with one simple testing note: test 3 angles to one audience OR one angle to 3 audiences, never change both at once.
OUTPUT FORMAT
- One-sentence offer restatement.
- A numbered list of 10 angles. For each, use this exact mini-structure on its own lines:
Angle [n]: [name]
Audience: ...
Awareness level: ...
Primary text: ...
Headline: ...
CTA: ...
- A short section titled "Test these 3 first" with the 3 picks and a one-line reason each.
- A one-line testing note.
RULES
- Use ONLY the proof the coach gives you. Never invent statistics, client results, or testimonials.
- No buzzwords. Banned: unlock, game-changer, revolutionary, in today's fast-paced world, secret sauce, level up.
- Each angle must be genuinely different. If two angles could swap headlines without anyone noticing, they are the same angle. Redo it.
- Match the language your IDEAL_CLIENT actually uses about their problem, not marketing language.
- Keep claims compliant: speak to the process and realistic outcomes, avoid guarantees and income promises.
How to set it up
You can run this as a one-off paste, or install it once and reuse it forever:
- ChatGPT Custom GPT: click your name > My GPTs > Create a GPT. In the Configure tab, paste the whole skill block into Instructions. Name it “Ad Angle Generator” and save. Now you open it like an app whenever you have a new offer.
- Claude Project: create a new Project, open Set custom instructions, and paste the skill block there. Every chat in that Project now behaves as the strategist.
- No setup: paste the skill block as your first message in ChatGPT, Claude, or Gemini, then send your seven inputs as the second message.
- Reuse fast: keep your seven inputs (OFFER, PRICE, IDEAL_CLIENT, CORE_RESULT, PROOF, PLATFORM, CTA) in a saved note so spinning up angles for a new offer takes 30 seconds.
See it in action (full example)
Here is the complete input a business coach sends after the skill is installed. Nothing is hidden:
OFFER: a 12-week 1:1 program that helps overwhelmed founders get out of the weeds and back to leading
PRICE: $4,500, paid monthly or in full
IDEAL_CLIENT: early-stage SaaS founders, 1-15 employees, working 60+ hour weeks
CORE_RESULT: reclaim 10+ hours a week and stop being the bottleneck on every decision
PROOF: 23 founders coached; one cut his work week from 70 to 48 hours in 90 days
PLATFORM: Meta (Facebook and Instagram) feed and Reels
CTA: book a free 20-minute Founder Clarity Call
And here is the complete output the model returns:
Offer, in one sentence: A 12-week 1:1 coaching program that helps overworked early-stage SaaS founders reclaim 10+ hours a week and stop being the bottleneck, for $4,500.
Angle 1: The 70-hour trap Audience: SaaS founders, 1-15 employees, interests in Y Combinator, SaaStr, Indie Hackers Awareness level: Problem-aware Primary text: You didn’t start a company to answer Slack at 11pm. But here you are, the bottleneck on every decision, the founder who can’t take a Saturday off. It’s not a discipline problem. It’s that nobody ever showed you how to lead instead of do. I coach SaaS founders through exactly that. One went from 70-hour weeks to 48 in 90 days, without dropping a single ball. If you’re running on fumes, let’s talk before the burnout makes the decision for you. Headline: Stop being the bottleneck CTA: Book a free 20-minute Founder Clarity Call
Angle 2: The 10 hours back Audience: Founders and CEOs, software industry, interests in productivity and time management Awareness level: Solution-aware Primary text: What would you do with 10 extra hours a week? Not theoretically. This week. Most founders I coach find those hours hiding in decisions they shouldn’t be making and tasks they should have handed off months ago. In 12 weeks of 1:1 work, we find them and give them back. 23 founders in. The goal isn’t to work more efficiently at 70 hours. It’s to lead well at 45. Headline: Get 10 hours back a week CTA: Book a free 20-minute Founder Clarity Call
Angle 3: The hustle myth Audience: Early-stage SaaS founders, interests in startup culture, bootstrapping, Lenny’s Newsletter Awareness level: Problem-unaware Primary text: “Just outwork everyone.” It’s the worst advice in startups, and it’s why so many capable founders quietly fall apart at year two. Working more hours isn’t a strategy, it’s a tax on your judgment. The founders who scale don’t grind harder. They stop being the answer to every question. That’s the skill I coach. If “hustle” has stopped working for you, that’s not failure. It’s information. Headline: Outworking won’t save you CTA: Book a free 20-minute Founder Clarity Call
Angle 4: For the founder who’s tried everything Audience: SaaS founders, interests in Asana, Notion, project management tools Awareness level: Solution-aware Primary text: You’ve tried the apps. The 5am routine. The “delegate more” advice. And you’re still the bottleneck. Here’s why none of it stuck: a new tool can’t fix a founder who hasn’t decided what only they should do. That’s not a system problem, it’s a clarity problem, and it’s the first thing we fix in my 12-week program. If you’re tool-rich and time-poor, we should talk. Headline: Tried everything, still stuck? CTA: Book a free 20-minute Founder Clarity Call
Angle 5: The 90-day proof Audience: Founders and CEOs, software, lookalike of past clients Awareness level: Product-aware Primary text: One founder I coached was working 70-hour weeks and convinced it was just “what the season required.” 90 days later he was at 48 hours, with a team that finally made decisions without him. We didn’t add hours to his day. We changed what he spent them on. That’s the whole program: 12 weeks, 1:1, focused on getting you out of the work and back to leading. 23 founders and counting. Headline: 70 hours to 48 in 90 days CTA: Book a free 20-minute Founder Clarity Call
Angle 6: “$4,500 sounds like a lot” Audience: SaaS founders, interests in fractional executives, executive coaching Awareness level: Product-aware Primary text: $4,500 for coaching sounds like a lot until you price the alternative: another year as the bottleneck, a team that can’t move without you, and a founder too fried to make good calls. The program is 12 weeks, 1:1, paid monthly or in full. Most founders find the 10 hours a week we free up worth more than the fee within the first month. If the math is the only thing stopping you, let’s run it together. Headline: Is the fee really the cost? CTA: Book a free 20-minute Founder Clarity Call
Angle 7: The Sunday-night founder Audience: Startup founders, 1-15 employees, interests in founder mental health, burnout Awareness level: Problem-aware Primary text: It’s Sunday at 9pm and your stomach already hurts about Monday. Not because the business is failing, but because every problem still routes through you. That feeling isn’t weakness. It’s the predictable result of building a company that can’t run without you in the room. I coach SaaS founders out of exactly that pattern in 12 weeks. You’re allowed to dread Mondays a lot less. Headline: Dreading Monday again? CTA: Book a free 20-minute Founder Clarity Call
Angle 8: This is for you if… Audience: SaaS founders, 1-15 employees, interests in scaling startups, first-time founders Awareness level: Problem-aware Primary text: This isn’t for founders who love being busy. It’s for the SaaS founder running a team of 1 to 15, working 60+ hour weeks, who knows deep down they’ve become the ceiling their company keeps hitting. If that’s you, the next hire won’t fix it and neither will another tool. 12 weeks of focused 1:1 coaching will. If you read “the ceiling” and felt it, we should talk. Headline: Are you the company’s ceiling? CTA: Book a free 20-minute Founder Clarity Call
Angle 9: The leader you stopped being Audience: Founders and CEOs, software, interests in leadership and management Awareness level: Solution-aware Primary text: Remember when you were excited about the vision, not buried in the inbox? Somewhere in the grind you traded being a leader for being the highest-paid doer in the company. The work is louder than the leadership, so the leadership disappears. My 12-week program is about getting that founder back, the one who steers instead of carries. 10+ hours a week, returned to the job only you can do. Headline: Get back to actually leading CTA: Book a free 20-minute Founder Clarity Call
Angle 10: Before you hire your way out Audience: SaaS founders actively hiring, interests in recruiting, startup hiring Awareness level: Problem-unaware Primary text: Most overwhelmed founders try to hire their way out of being the bottleneck. Then they discover the new hire needs them for every decision too, and now they’re managing more people and still drowning. The problem was never headcount. It’s that everything routes through you by default. Fix that first and your next hire actually lightens the load. That’s the first month of my program. Headline: Hiring won’t fix the bottleneck CTA: Book a free 20-minute Founder Clarity Call
Test these 3 first: 1. Angle 5 (The 90-day proof) — leads with your strongest real proof, and product-aware buyers convert fastest. 2. Angle 1 (The 70-hour trap) — sharpest pain plus the easiest audience to target precisely. 3. Angle 7 (The Sunday-night founder) — most emotionally specific scenario, great for cheap engagement and learning.
Testing note: Run these 3 angles to ONE audience (early-stage SaaS founders) first. Once a winning angle emerges, take it and test it against 3 audiences. Never change angle and audience in the same test, or you won’t know what moved the needle.
That’s a full testing plan, not a single ad. Pick the three, write your offer in your own voice where it sounds stiff, and you can launch this week.
Why this works
Four LLM principles are doing the heavy lifting. Learn them and every prompt you write improves:
- Role priming. “You are a senior direct-response strategist who thinks in angles, not adjectives” tells the model which slice of its knowledge to use. “Write me some ads” pulls the bland average of the internet; a sharp role pulls direct-response thinking. The role even encodes a philosophy (“the job of a first ad is to find what responds”), which shapes every output that follows.
- Define your terms so the model can’t cheat. The biggest failure here is ten reworded headlines pretending to be ten angles. The prompt closes that loophole twice: it defines an angle (“a different reason someone buys”), hands the model a menu of ten distinct psychological levers, and adds a self-check rule (“if two angles could swap headlines, they’re the same angle, redo it”). When you give the model a clear definition and a way to check itself, you get real variety instead of synonyms.
- Constraints are quality control. The word counts, the 40-character headline cap, the awareness-level tag, and the banned-buzzword list each remove a common failure mode. The “use only the proof I give you, never invent statistics” rule is the most important one for coaches: it keeps your ads honest and compliant. Telling the model what NOT to do is as powerful as telling it what to do.
- Conditional clarifying questions. “Ask up to 3 questions only if a critical input is missing” gives the model a way to fill gaps by asking instead of guessing, which is the single biggest fix for generic AI output, without making it interrogate you when your inputs are already solid.
Do this now
- Install the skill as a Custom GPT or Claude Project (or just paste it into a fresh chat).
- Fill in your seven inputs. Be brutally specific about your IDEAL_CLIENT and use only real PROOF.
- Run it, then read the “Test these 3 first” section, that’s your launch list.
- Lightly rewrite the winning angles in your own voice and ship one test this week.
Pro tips
- Give it real language. Paste two or three sentences a real client said about their problem into the IDEAL_CLIENT field. The angles will mirror that language and out-convert anything written in marketing-speak.
- One variable per test. The model already warns you, but it bears repeating: test 3 angles to one audience, or one angle to 3 audiences. Never both. Otherwise a win teaches you nothing.
- Re-run for a different awareness level. Ask it to generate 10 more angles but only for problem-unaware cold traffic. Cold and warm audiences need completely different doors in.
- Turn winners into a hook bank. Save the headlines and primary-text openers that perform. Within a few campaigns you’ll have a swipe file of openers proven on your exact audience.
0 comments
No comments yet.