Good copy is assembled, not invented
I don’t sit down and invent the words for a landing page. Most of the lines that end up doing the heavy lifting were said by a customer first, and my job was to notice them and put them somewhere a stranger would read them. That’s why research isn’t the boring bit before the writing starts. It’s where the writing comes from.
Which leaves two practical questions: where does that raw material live, and what do you ask once you’ve got someone’s attention? Most marketers jump straight to customer interviews and then stall for three weeks because nobody answers the booking link. Start with what’s already sitting in public instead, and you’ll reach the calls with better questions anyway.
Competitors leave their evidence in public
Facebook and LinkedIn both run ad libraries, free to search. Look up a competitor or a keyword and you’ll get their current and past ads laid out for you. You can’t see how any of it performed, but you can see how long it ran, and an ad that’s been live for two months is a safer bet than one that vanished after a week. Rough signal, still a signal.
Then follow the promising ads through to the landing pages behind them. Don’t copy a competitor blindly, because you have no idea what’s actually converting, but you can make an educated guess: if the ad looks like it’s earning its budget, the page it points at is probably pulling its weight too. Run that page through the Internet Archive and you get the edit history rather than a snapshot. The headline they’ve rewritten four times is the one they’re unsure about, and the section that survived every redesign is the one they know works.
Search engines tell you what people type when nobody’s watching
Type your customer’s problem into Google and scroll to “people also ask.” Every question in that box is a phrasing real people searched, and expanding one spawns three more. Ten minutes there gets you a list of what your market is confused about, worried about, and comparing, in their words rather than your category’s.
I do this before any page that has to rank, but also before interviews, because it flags the questions so common that asking them on a call would waste the call.
Forums are where buyers talk without a filter
Reddit and Quora do the job traditional forums used to do, at a scale worth mining. Dig into the subreddit for your niche, or the question threads on your category, and you’ll find objections, comparisons, workarounds, and the answer people give when someone asks “what should I use for this?” Most of it is written under a pseudonym, so nobody’s protecting a professional reputation and nobody’s hedging.
That’s also what makes forums the best vocabulary source you’ll find: the shorthand your buyers use, the nickname they’ve given the problem, the tool everyone names as the default. Search site:reddit.com "competitor name" alternative and read every thread that comes back. Those are people describing their switching moment out loud.
Reviews hand you the before and the after
Review sites are the most concentrated source of customer language on the internet, so read yours and read every competitor’s on G2, Capterra, Trustpilot, or whatever your market uses. Skip the five-star ones, because “great product, love it” tells you nothing. The three-star reviews are gold: someone liked it enough to keep paying and still had a reason to complain, and that reason is usually the exact hesitation your next prospect will have.
And don’t stop at software reviews. Find three or four books on Amazon about the problem your product solves and read the reviews of those. People reviewing a book about a problem describe the problem at length, explain what they’d tried before picking it up, and say whether it helped. Nobody writes those with a purchase decision in mind, so you get framing that exists nowhere else.
The comments under the gurus beat the gurus
Social platforms are packed with self-appointed experts, and most of what they post is recycled. Their audience is another matter. Find the loudest voices in your niche, ignore the posts, and read the comments underneath: that’s where people bring their half-formed questions, their “does this still work if I’m B2B,” their story about what went wrong last quarter. A focus group that assembled itself.
Your own inbox is the source most people never mine
Your own customers leave evidence everywhere, and it beats a stranger’s, because you can go back and ask follow-up questions. Support tickets, chat logs, sales call recordings, cancellation reasons, the questions in the last email before a deal goes cold: research nobody has to schedule.
The fastest version is to sit in on three sales calls and write down every question the prospect asks, then ask whoever runs onboarding what people get stuck on in week one. If the same question shows up in both places, it belongs on the page. What you’re after is the objection so routine your team stopped noticing they answer it daily.
Desk research gets you the vocabulary, conversations get you the reasons
Reading gets you a long way and then hits a ceiling: you can see what people say, but you can’t ask them why. Customer interviews are the most valuable tool in the kit for exactly that reason, so once the reading’s done, book the calls. Thirty minutes, recorded with permission, no slides, no demo. Here’s the list I work from, grouped by what you’re trying to learn.
Product development
- If you could have us make something just for you, what would it be?
- If you had a magic wand and could instantly solve your problem, how would your life change?
- What else should we offer you?
- What is one thing you wish we could solve?
- Is there anything you’d change about how our product works?
Product adoption
- How are you using our tool or service?
- Do you use any other tools or resources alongside it?
- What would persuade you to use our product more often?
- How would you react if you were told you have to stop using our product?
- Tell me about the last time you used our product. How did it go?
- What made you hesitate the most before starting to use it?
Marketing and messaging
- Do you ever talk about us to other people? What do you say?
- What’s your number one challenge related to the problem we solve, and why is it so challenging?
Customer journey
- What made you start using us? What convinced you?
- What would have made you more likely to start sooner?
You won’t get through all of them, and you shouldn’t try. Pick five, and let the good answers pull you off-script.
How to run the call so you get real answers
The questions matter less than what you do in the gaps between them, and the biggest improvement most people can make is to stop talking. Ask, then wait. Five seconds of silence feels unbearable when you’re hosting, and it’s usually the moment someone stops giving you the polite answer and starts giving you the real one.
Two habits do the rest. Always ask for the last time rather than the usual, because “tell me about the last time you used it” gets you a story with details while “how do you usually use it” gets you a summary they’ve already sanded smooth. And follow every interesting answer with “why was that hard?” until the answer stops changing. The third why is where the useful material sits.
Then resist defending the product. When someone says the onboarding confused them, the win is finding out what they expected instead.
Turning a transcript into copy
A recording is only worth something if you mine it, so read the transcript with a highlighter in hand and mark every sentence that sounds like a person rather than a company. Verbatim, typos and swearing intact. Then dump those lines into one document sorted by what they do: trigger moments, objections, before-states, outcomes, comparisons.
That document is your copy source. Need a headline, and you’re picking from things customers said instead of staring at a blank page. Need to handle an objection, and the words are already written in the phrasing that made you nod. It’s slower than making it up, and it’s the difference between copy that sounds like your market and copy that sounds like a marketing team.