Marketing
Beginner
18 min
Customer research and personas — built from evidence, not from imagination
The problem
Your persona document says your customer is a 28–40 year old urban professional who values quality and convenience. So does every other brand's persona document in your category. It tells you nothing about what to say, what to fix, or which objection is killing conversion.
The workflow
1Gather the text
2Strip identifiers
3Cluster the themes
4Extract their vocabulary
5Rank objections
6Mine competitor reviews
7Write the persona
★Test against a decision
What you need
Business: D2C stainless steel water bottles, ₹899–₹1,499, Amazon + own site. Source: 240 of our reviews plus 180 competitor one-to-three star reviews, identifiers removed.
The tool
An assistant's file upload plus code mode
The prompt
Subscribers. The full prompt is part of the SMEMinds AI Playbook. See plans.
What comes back
1. PROBLEMS (ours, ranked) Leaking in a bag — 61 of 240 mentions.
Verify before you use it
Human review required. Check every line below before this leaves your screen.
- Confirm the counts by spot-checking. Ask for the code and verify one theme's count manually against the source file.
- Check that no demographic detail crept in. If the persona mentions age or income and the reviews do not, it was invented.
- The 'leaking versus insulation' finding is a claim about your marketing — verify it against your actual product page before acting on it.
- Verify the competitor review sample is representative rather than the most recent 180, which may be skewed by one bad batch.