Most marketing problems aren’t execution problems. They’re understanding problems. People rush to write copy, rush to run ads, rush to “test” — all before they actually understand who they’re selling to. The result is fast iteration on the wrong premise, which is just guessing faster.
This is a simple exercise that fixes that. You answer a small set of honest questions, run them through AI, and use the output to expose the tension that’s actually blocking conversions.
Step 1: Answer the Eleven Questions Honestly
The exercise starts with eleven questions about your product, your customer, and your competitors. Don’t draft polished answers — answer the way you’d answer a friend who asked you over coffee. The honesty is the whole point.
Step 2: Run the Answers Through ChatGPT
Paste your answers into ChatGPT and ask it to do three specific things:
ChatGPT is the right tool for this specific job because it’s built for synthesis across messy, contradictory inputs — which is exactly what your eleven answers will be once you answer them honestly. You’ll get back a profile that names tensions you didn’t see while writing the answers yourself.
Step 3: Pressure-Test the Profile Against Real Conversations
Here’s where the exercise either becomes powerful or stays theoretical.
The output ChatGPT gives you is a hypothesis, not a truth. The way you confirm or break the hypothesis is by looking at what your customer actually says — not what you think they say. The fastest way to do this is to monitor the conversations they’re already having in public.
Devi is built for exactly this. It watches Reddit, Facebook groups, LinkedIn, and Twitter for the keywords and problem statements you set, and surfaces real conversations where your audience is articulating their frustrations in their own words. You feed those real phrases back into your profile and the messaging starts writing itself — because you’re no longer guessing what they think, you’re literally reading it.
Step 4: Look For Tension, Not Confirmation
This is where most people get this exercise wrong.
They run the questions, get a clean summary, nod at the parts that match what they already believed, and move on. The analysis sounds smart but nothing changes downstream — same copy, same offers, same conversion rates. The exercise becomes a ritual instead of a tool.
That gap — between what they say they want and what they keep doing instead — is where conversion actually lives. Read the ChatGPT output looking for friction. Look for contradictions. Look for cost. Look for the thing the customer would never admit out loud but is obviously true once you see it. That’s where real messaging comes from, and it’s the part most marketers skip because it requires sitting with discomfort instead of jumping to tactics.
Step 5: Translate Understanding Into Output
Once the profile is honest and pressure-tested, the downstream work gets dramatically easier. The same understanding feeds copy, offers, ad creative, content, and product decisions — because you’re no longer making each of those individually, you’re expressing the same insight across surfaces.
The bottleneck for most marketers isn’t producing more output. It’s deciding what the output should say. With a real customer profile in hand, the saying-what gets answered, and tools like LoudAF — which generates carousels, threads, articles, and posts in your voice across formats — turn one clear insight into weeks of distribution. Without the profile, LoudAF just produces more polished noise. With the profile, it produces messaging that actually converts.
The Rule Most People Miss
Here’s the part that matters more than any of the steps above.
If your understanding of the customer is wrong, every campaign you run, every ad you test, every email you send is broadcasting the wrong message louder. You can iterate forever without getting closer, because iteration only sharpens what you already believe. It can’t tell you that what you believe is the problem.
That’s why this exercise is the one most marketers skip and the one that pays the most when it’s done honestly. Slow down at the output. Look for the gap between what your customer says and what they actually do. Use Devi to read their real words instead of guessing at them. Then let everything else — the copy, the offers, the content, the funnels — flow downstream from a profile you actually trust.
Until you do that, you’re just guessing faster.
The Full Customer Profiling Stack
Three tools, used in sequence:
Run this once for any product or offer you’re working on, and the rest of your marketing gets noticeably easier the next day.
Frequently Asked Questions
How often should I redo this exercise? At launch, when conversions plateau, and any time the underlying market shifts. For most products, twice a year is enough. For fast-moving niches, quarterly.
What if my answers feel obvious or generic? That usually means you’re answering the way you wish your customer thought, not the way they actually think. Use Devi to pull real customer language from public conversations and rewrite your answers using the words they actually use.
Can I use Claude or another AI instead of ChatGPT? Yes. Any frontier model handles this kind of synthesis well. The exercise is about the questions and your honesty, not the specific AI.
Why is this better than a survey? Surveys ask customers what they think, which is filtered through what they want to admit. This exercise plus public-conversation monitoring captures what they do and say when they’re not being asked — which is consistently more accurate than what they self-report.
Final Word
When understanding is clear, everything downstream gets easier — copy, offers, content, decisions. When it isn’t, no amount of execution closes the gap. The exercise above is small. The honesty it requires isn’t. Do the small thing honestly, and the rest of your marketing becomes a translation problem instead of a guessing game.
See the full growth stack on our tools page →