Website Feedback: How to Get Cheap User Feedback

What is the cheapest way to get feedback on a website?

The cheapest way to get website feedback is to ask a small number of people one clear question about one page: show the page, then ask what they remember or what they think the product does. A 5-second test, a one-question on-site survey, five short user interviews, or an unmoderated remote test that asks a small recruited panel to describe your page in one sentence can each be done for little or no money in a few days. The trick is not the tool, it is asking the right question in a way that does not lead people to the answer you hope for.

This guide covers six methods, when each one fits, how to write questions that give honest answers, how many responses you need, and how to turn a pile of comments into a decision.

Start with the question, not the tool

“Get feedback” is too vague to plan. Write down the one thing you want to learn, because it decides the method:

• Is the page clear? Do new visitors understand what you offer and who it is for? This is a first-impression question.
• Can people use it? Can someone sign up, find the pricing, or complete checkout without getting stuck? This is a usability question.
• Why do visitors leave or hesitate? This is a motivation question, and your existing visitors are the best people to ask.
• Which version works better? This is a comparison question, which needs more answers, or a real A/B test.

One question per round keeps the feedback focused and keeps the cost down.

Six cheap ways to get website feedback

1. Five-second tests

Show someone your page for five seconds, hide it, and ask: “What is this website for?” and “What do you remember?” If people cannot say what you sell after five seconds, your headline and first screen need work, whatever else is on the page. You can run this with dedicated testing tools such as Lyssna’s five second test, or simply with a screenshot and a timer on a video call.

2. Short on-site surveys

A one or two question survey on the page itself reaches people with real intent. Good moments to ask are on the pricing page (“Is anything stopping you from signing up today?”), after sign-up (“What almost stopped you from signing up?”), or on exit. Keep it to one open question and an optional second one. Every extra question cuts the number of people who answer. For longer surveys sent to a wider audience, How to get survey respondents covers sources and quality checks.

3. Session recordings and heatmaps

Recordings show what people do rather than what they say: where they scroll, which buttons they click, and where they stop. They are excellent for spotting a button nobody sees or a form field that makes people quit. They cannot tell you why. Use them to find the problem, then ask about it with a survey or an interview. Recording tools collect personal data, so mention them in your privacy policy and check whether you need visitor consent in the countries you serve.

4. User interviews and usability tests

Sit with one person, in person or on a video call, give them a task such as “find out how much the basic plan costs”, and ask them to think out loud. Do not help them. It is the slowest method per person, but it finds problems no survey will: the moment someone hesitates, rereads a sentence, or clicks the wrong thing. Recruit from your own users, your email list, or people who fit your target customer, and offer a small thank-you. How to find beta testers covers more places to find willing people.

5. Communities

Many forums, subreddits and Discord servers have feedback threads or days where members review each other’s sites. Peers often spot technical and design issues fast. Two cautions: read each community’s rules on self-promotion before posting, and remember that fellow founders and designers are not your customers. Their opinion on your pricing page is not the same as a buyer’s.

6. Unmoderated remote tests with a recruited panel

When you need many fresh pairs of eyes quickly, use an unmoderated remote testing tool and recruit participants through its research panel. You build the test once, for example a 5-second test of your first screen followed by one written question, and people who have never seen your site take it on their own device, without you on a call. Because you get dozens of independent answers within days, you can count how many people got it right, which interviews cannot give you.

Three widely used tools, as described on their own sites (checked on October 22, 2026):

• Lyssna offers five second tests, surveys and other quick studies. Its free plan includes 15 self-recruited responses, and responses from its participant panel are priced separately, pay per use (Lyssna pricing).
• Maze lists prototype testing, surveys and live website testing, and says recruitment from its participant panel is available on all plans, with limits (Maze pricing).
• UserTesting lists unmoderated tests, including surveys, interaction tests and think-out-loud sessions, plus a global participant panel. Prices are on request (UserTesting plans).

Plans, limits and panel prices change, so compare the current offer on each site before you buy. Panel participants are paid by the platform for their time, not drawn to your product by need, so use this method for clarity, first-impression and usability questions, not for questions like “Would you buy this?” Use the panel’s targeting options, such as country, to get closer to your real audience.

Two details make this work. First, count every honest answer, including wrong or critical ones: a confused answer is exactly the feedback you are paying for, and leaving it out would hide your problem. Set aside only blank or off-topic answers when you tag them. Second, ask an open question that cannot be answered by copying the headline. In a 5-second test the page is hidden before the question appears, which helps.

How to write questions that give honest answers

The way you ask decides what you hear. Most bad feedback comes from questions that suggest the answer.

• Ask open questions. “What does this product do?” tests clarity. “Is it clear that we help freelancers send invoices?” has already told them.
• Ask about behavior, not predictions. “What would you do next on this page?” is more useful than “Would you sign up?”, because people are poor at predicting their own future choices.
• Ask one thing at a time. “Is the page clear and attractive?” gets one answer to two questions.
• Avoid loaded words. “How easy was our simple sign-up?” pushes people toward “easy”.
• Do not reveal what you hope for. Do not say you built it, or that you are testing a new headline you love.
• Leave room for the unexpected. End with “Anything else that confused you?” as an optional extra.

How many responses do you need?

It depends on the kind of question.

• Finding problems: about five people per round. Jakob Nielsen’s widely cited 2000 article for Nielsen Norman Group, Why you only need to test with 5 users, argued that testing with five users finds about 85% of the usability problems in a design, and recommended several small rounds rather than one large study. Fix what the first five reveal, then test again with five new people.
• Measuring a share: closer to 40. When the question is “what share of people understand our headline?”, you are counting, not exploring. Nielsen Norman Group suggests about 40 participants for most quantitative studies, which gives a margin of error of around 15% at 95% confidence. Fewer answers are fine if you accept a wider margin.
• Comparing two versions: more again. A small difference between two headlines needs many answers per version to be real. For live traffic, a proper A/B test is better; A/B testing on a small budget explains why low traffic needs patience.

How to analyze website feedback

Raw comments feel convincing one by one and are easy to misread in bulk. A simple process keeps you honest:

1. Write down what you expected before you read anything, such as “most people will say we help freelancers send invoices.”
2. Read every answer once without judging, then create three to six themes, for example “understood correctly”, “thought it was accounting software”, “no idea”.
3. Tag each answer with one theme in a spreadsheet and count them.
4. Pull out exact words. The phrases people use to describe your product are often better headline copy than your own.
5. Decide one change that addresses the biggest theme, make it, and run the same question again.

Common feedback mistakes

• Asking friends and family. They know you, want to be kind, and are rarely your audience.
• Changing five things after one round, so you never learn which change helped.
• Treating one loud comment as a trend. Count themes before you act.
• Asking paid panel testers whether they would buy. Their answer tells you about the incentive, not your market.
• Collecting more personal data than you need. A clarity question needs no email address.
• Never retesting, so you do not know whether the fix worked.

Before you send paid traffic to a page, run it through our landing page checklist. Feedback works best on a page that already loads fast and works on phones.

Frequently asked questions