To test a homepage message, show it to people who resemble the intended customer and ask them to explain what the company offers, who it is for, what problem it solves, why it may be different and what they would do next. Observe their interpretation before asking for opinions. You are testing the message, not whether participants can produce better copy on demand.
Decide what you need to learn
“Does the homepage work?” is too broad for a useful session. Turn uncertainty into specific research questions.
For example:
- Can intended customers identify the offer from the opening section?
- Do they recognise themselves as the intended audience?
- Which outcome do they think the product enables?
- What do they assume the product or service includes?
- Do they understand the next step and feel ready to take it?
- Which claim needs more evidence?
Choose two or three priority questions for one round. If you test navigation, visual design, pricing, brand preference and every page at once, it becomes hard to explain what caused a response.
Write down the team’s current assumption for each question. This makes the session a genuine test rather than a search for comments that support a preferred draft.
Recruit for the decision, not convenience
Participants should resemble the people who need to understand the message in the real buying situation. Colleagues can help you rehearse a session, but they know too much to test first impressions.
Define practical recruitment criteria:
- role in using, influencing or buying the offer;
- relevant business stage or situation;
- experience with the problem or current alternative;
- level of familiarity with your company;
- any access needs that affect how they use the page.
Include a mix of existing customers and suitable non-customers when possible. Existing customers can explain whether the message reflects the value they know. Non-customers reveal what the page communicates without prior context.
Recruit a small, relevant group for an initial round, then repeat after changes. Qualitative testing identifies patterns and misunderstandings; it does not produce a statistically representative conversion forecast. Keep personal information secure, explain how notes or recordings will be used and obtain informed consent.
Prepare the page and a neutral guide
Test a live page or realistic prototype at the size and device relevant to participants. Remove obvious placeholder content. If the page is unfinished, tell participants which parts work and which do not without explaining the intended message.
Build a short discussion guide:
- Welcome, purpose, consent and reassurance that you are testing the page.
- A few context questions about how the participant handles the relevant problem today.
- A first-impression task.
- A realistic exploration task.
- Follow-up questions about meaning, relevance, trust and next steps.
- A closing question for anything you did not cover.
Avoid leading tasks such as “Find our automated reporting feature”. Try: “Imagine your monthly reporting process has started taking too long. Look at this page and decide whether this company may be relevant.” Good tasks provide a believable goal without giving away the answer.
Illustrative example
Suppose the homepage opens with:
One workspace for ambitious teams to unlock better performance.
Show the opening for a short period, hide it, then ask: “What do you think this company does?”
One participant may describe project software. Another may assume management training. A third may not know. Do not correct them. Ask, “What on the page gave you that impression?” The ambiguity is evidence about the message.
Now show a clearer alternative:
Plan weekly stock orders for every café from one sales forecast—without rebuilding spreadsheets for each location.
Ask the same questions. If intended users can explain the offer and context accurately, explore what remains unclear: where the forecast data comes from, who sets orders or what systems it connects to. A clearer message often reveals the next information gap; that is useful, not failure.
Run the session without teaching the answer
Ask participants to think aloud, then give them space. Silence can feel uncomfortable, but explaining the page removes the very uncertainty you need to observe.
Useful prompts include:
- “What do you think this means?”
- “Who does this seem to be for?”
- “What would you expect to happen next?”
- “What makes you say that?”
- “Which words helped you reach that conclusion?”
- “What, if anything, would you need to believe this?”
Ask about behaviour and past experience before hypothetical preference. “Tell me about the last time this problem happened” provides stronger context than “Would you buy this?” A participant cannot reliably predict a future purchase from a short research session, but they can explain what they understand and what they have done before.
Have one person facilitate and another take notes when possible. Record exact phrases selectively, but do not treat a colourful quote as a finding until you understand its context and see whether the issue repeats.
Analyse patterns against the intended meaning
After each session, capture observations under the research questions. Separate:
- what the participant did or said;
- your interpretation;
- a possible design or copy response.
This separation prevents the first proposed rewrite from becoming the finding.
Create a simple message map with rows for audience, offer, problem, outcome, difference, proof and action. In one column, write what the team intended. In another, summarise what participants understood. Mark gaps that appeared across sessions and note whether they block comprehension, reduce relevance or weaken confidence.
Prioritise problems before solutions. If nobody understands the category, changing a button label is unlikely to matter. If people understand the offer but hesitate at an unsupported claim, add credible proof rather than rewriting the entire page.
Revise one layer at a time
Start with the highest-impact misunderstanding. Change the strategic message if the audience or value is wrong. Change the copy if the strategy is sound but the language hides it. Change hierarchy or layout if the right words are present but easy to miss.
Run another round with revised material. Keep a short decision log showing what changed, why and what evidence informed it. Website analytics can then show how people behave at scale, while interviews and usability sessions help explain why. Neither source should stand alone.
Your useful next step
Choose one homepage section and write the five meanings a visitor should take from it: offer, audience, problem, outcome and next step. Recruit three suitable people for a first learning round, use the same neutral questions with each, and record their words before discussing possible rewrites with the team.