How to find out what your audience wants before you spend on ads
The expensive way to learn what your audience wants is to guess, build the store around the guess, and buy traffic until the numbers tell you that you were wrong. That lesson costs whatever you spent on ads.
The cheap way takes a few hours and no money. It's tedious rather than difficult, which is why most people skip it.
Read complaints, not praise
Positive reviews tell you what a product does. Complaints tell you what people wanted and didn't get, and that gap is where your opening is.
Go to the marketplaces where products like yours are already sold and read the two and three star reviews. Not the one stars, which are usually about a damaged parcel, and not the five stars, which are usually thin. The middle is where people explain what nearly worked.
You're looking for repeated phrases. If eleven people say the strap digs in, that's a product decision. If eleven people say it arrived faster than expected, that's a line for your product page.
Do the same in forums, subreddits and Facebook groups where your audience gathers. Search inside them for "recommend", "worth it", "anyone tried". Those threads are people telling each other what they're trying to buy and why they haven't.
Write down their words, not yours
Keep a plain document with two columns as you read. On one side, the exact phrasing people use. On the other, what they say gets in the way.
The exact phrasing matters more than it sounds. People search and click using their own language, not the language of a marketing team. A page written in the customer's words converts better than a polished one written in yours, and it takes less effort to produce.
By the end you should have twenty or thirty real sentences. Those become your headlines, your product descriptions, your ad copy and your FAQ. You're not inventing messaging at that point, you're selecting it.
Check that demand is real, not just loud
Enthusiasm in a niche community doesn't always mean people spend. A few quick checks separate the two.
Look at whether anyone is running paid ads for these products. Sustained advertising is the strongest signal you'll get for free, because nobody keeps paying for ads that lose money.
Look at how many established sellers exist. An empty market is usually empty for a reason. A crowded one is a nuisance to enter but proves people buy.
Look at search demand for the terms your audience uses. Steady is better than spiky. A term that peaks once a year is a seasonal business, which is fine if you plan for it and painful if you don't.
Talk to ten people
This is the step almost nobody does, and it's worth more than everything above.
Find ten people in the audience and ask them what they last bought in this category, what they nearly bought instead, and what almost stopped them. Not "would you buy this", because people are polite about hypotheticals. Ask about a purchase they already made, because memory is more honest than prediction.
Ten conversations will contradict at least one thing you were confident about. That contradiction is the whole point.
Turn it into decisions
Research is only useful if it changes what you do. When you're finished you should be able to fill in these five, in the customer's language rather than your own.
- Who this is for, specifically enough that you could describe them to a stranger
- The problem they're trying to solve
- What they've tried already and why it disappointed them
- The one objection most likely to stop them buying
- Where they gather, so you know where to reach them
Point four earns its place on the product page rather than in the FAQ, because objections you leave unanswered turn into abandoned carts.
Keep doing it after you launch
Your first customers are better research than anything you read beforehand. Every support email is someone telling you what wasn't clear. Every refund request is a mismatch between what they expected and what arrived.
Set aside twenty minutes a week to reread them and note the repeats. Two people asking the same question is coincidence; five is a page that needs rewriting.