Module 02 · Research and site architecture

Find customer language and read search intent

Use customer evidence and live results to understand the task behind a query before you decide what page to create.

Lesson 04Research and site architecture · Practical courseLast updated
11% through the course

What you will learn

Keyword research is not a contest to collect the largest list. It is a way to hear how customers describe a problem, what they know already, and what they need before they can move forward. This lesson shows how to turn those clues into a useful page decision.

By the end of this lesson: You can build an intent brief from first-party language, search results, and customer context without treating search volume as the only signal.

Why this matters

01

A phrase can mean different things in different contexts. ‘Payroll software’ may be a broad comparison, while ‘payroll software for restaurants’ may signal an industry-specific evaluation. The words matter, but the decision behind them matters more.

02

First-party evidence keeps research grounded. Sales calls, support tickets, reviews, on-site searches, product demos, and customer interviews reveal objections and vocabulary that a keyword tool cannot explain.

Keep the boundary clear: Search results are clues, not a command to copy the current winners. A result page can include mixed intents, location effects, personalization, and formats that do not match your customer or offer.

The intent evidence stack

Field note 04The intent evidence stackRead the evidence

Start with what customers say, then examine the query and current results. Infer the decision stage carefully. Finally, choose a response that gives more useful help than a superficial variation of an existing page.

Core concepts

01

Language is evidence of a problem

Collect exact phrases from customers, including the context around them. ‘Can I run payroll for staff in two provinces?’ tells you more than the isolated word ‘payroll.’ Preserve the question, source, audience, and date.

Use it when: Can you point to a real person or first-party record behind this phrase?

02

Intent is a working hypothesis

Classify the task as learning, comparing, solving, navigating, buying, or completing an action. Then state your confidence and the evidence. Intent is useful when it guides a testable page decision, not when it becomes a rigid label.

Use it when: What would you expect a helpful result to let the customer understand or do?

03

Format reveals expectations

A result set with product pages, comparisons, videos, maps, and official documentation may indicate several needs. Look for the formats that recur and the questions left unanswered, rather than cloning a title pattern.

Use it when: Which format appears to serve the dominant task, and what gap could your page fill?

04

Qualification changes the brief

A page should attract people you can genuinely help. Add constraints such as country, business size, product fit, price range, expertise level, or legal eligibility. Clear qualification can improve both trust and conversion quality.

Use it when: Would the right customer recognize themselves and the wrong customer understand the limit?

The practical method

  1. 01

    Collect raw language

    Pull questions and objections from calls, tickets, reviews, chat logs, on-site search, and interviews. Keep the source and surrounding context.

  2. 02

    Normalize without erasing meaning

    Group clear variants together, but retain meaningful distinctions such as country, audience, urgency, product type, and use case.

  3. 03

    Inspect live results

    Review representative queries in the relevant market. Note result types, recurring promises, source types, and unanswered questions.

  4. 04

    Write an intent hypothesis

    State the likely task, stage, audience, evidence, confidence, and the page format that may help. Make uncertainty visible.

  5. 05

    Check the business fit

    Confirm that your offer, expertise, operations, and proof can support the response. Do not create a page for demand you cannot serve well.

  6. 06

    Choose the next research or page action

    Decide whether to improve an existing page, create one focused response, gather more evidence, or deliberately do nothing.

Guided workshop

Collect customer language before you turn it into a content plan

This section turns the lesson into a bounded working session. It is designed to leave you with an intent evidence log that distinguishes what people say, what results currently offer, and what your page can uniquely help with.

Practice scenario

Practice scenario: A home-services company wants a page for ‘emergency heat pump repair.’ The first draft uses a keyword tool list and broad service language. Call recordings show that customers are actually worried about the time of day, whether a technician serves their area, what happens if a part is unavailable, and how to avoid damaging the system before help arrives.

The team gathers language from calls, chat, reviews, search suggestions, and the results people can currently choose from. It separates urgent repair intent from people comparing repair versus replacement. It also notes which questions need a service page, which need a short emergency guide, and which need a receptionist script.

The deliverable is not a giant keyword sheet. It is an evidence log that makes the next page brief more useful and keeps the team from copying competitors' vague wording.

Build it step by step

01

Collect first-party language safely

Use sales calls, support tickets, chat logs, reviews, and customer interviews. Remove personal details and preserve the exact wording of the problem, context, objection, and desired outcome.

Make it tangible: Save a redacted language collection. It helps the team decide which phrases describe real customer tasks. Check inventing customer language from internal jargon before moving forward.

02

Group statements by the decision behind them

Cluster phrases into tasks such as diagnose, compare, price, book, confirm location, or solve an urgent problem. Keep similar words apart when they imply a different next action.

Make it tangible: Save an intent group list. It helps the team decide which pages or experiences need different jobs. Check assuming all phrases containing one keyword mean the same thing before moving forward.

03

Read the current result set

Inspect the types of pages that appear, their format, freshness, local intent, commercial evidence, and unanswered questions. Record the observation date and market because results can change.

Make it tangible: Save a result-set observation note. It helps the team decide what format and evidence the searcher may expect. Check treating the result set as a permanent rule before moving forward.

04

Find the evidence gap

Ask what a careful customer cannot learn from the current options. The gap may be a missing comparison, local constraint, cost explanation, preparation checklist, or clear next step.

Make it tangible: Save a short gap statement. It helps the team decide what your source can make clearer without repeating existing pages. Check creating a page that only rephrases the same answer before moving forward.

05

Choose the smallest useful format

Match the format to the task. An urgent question may need a concise service page and safety guidance. A comparison may need a table, sources, and an honest boundary.

Make it tangible: Save a format decision with rationale. It helps the team decide whether to create, update, merge, or support a page. Check forcing every question into a long blog article before moving forward.

06

Validate language with a human owner

Ask the team that hears customers to check the brief. Confirm that the wording is accurate, respectful, and not overstating what the business can deliver.

Make it tangible: Save a reviewed intent brief. It helps the team decide whether the page will sound useful in a real conversation. Check publishing a phrase that customers would never use before moving forward.

Working template

Use these fields in a document, task, or spreadsheet. Keep the evidence close to the decision.

  1. Exact customer wording: Capture the phrase or question, removing names and personal details. A support or sales owner should confirm the wording is authentic.
  2. Situation: Describe what happened before the customer searched and what makes the decision urgent or difficult. A writer should be able to picture the customer without making assumptions.
  3. Intent group: Choose the underlying task: learn, compare, solve, verify, book, or another clear action. A strategist should be able to distinguish it from nearby groups.
  4. Result-set observation: Record the date, market, common formats, and gaps noticed in available results. A reviewer should know this is an observation, not a universal ranking rule.
  5. Best response: Name the page, guide, tool, script, or product evidence that can help the customer next. The page owner should be able to state why this response fits the task.
  6. Proof and boundary: List what supports the answer and where the business must be careful not to promise too much. An expert or compliance owner should be able to review the claim.

Quality review before you ship

Use these checks while the evidence, owners, and customer context are still easy to correct.

  1. Compare every proposed phrase with its original customer source. Preserve the meaning, context, and uncertainty rather than polishing a real concern into a generic keyword or a promise the business cannot keep.
  2. Test the page outline with someone outside the project. They should be able to point to the question, decision stage, proof needed, and next action without being told the team's internal topic label.
  3. When two customer groups use the same words for different needs, split the evidence before writing. A shared phrase is not proof that one page, offer, or call to action serves both groups.

Decision rules for the real world

A phrase has high volume but weak fit

Do: Treat it as a research lead, then verify the customer task before assigning a page.

Avoid: Do not let volume override relevance and capability.

Two phrases share words but imply different urgency

Do: Give them separate intent records and potentially different page paths.

Avoid: Do not collapse them into one generic article.

A competitor page covers the topic well

Do: Look for a clearer task, better evidence, local nuance, or a better next action.

Avoid: Do not copy its structure or claims just to match it.

Internal teams disagree on wording

Do: Bring a small sample of customer evidence to the review and test the language with the customer-facing team.

Avoid: Do not settle the question by seniority alone.

Coach notes

  • A phrase becomes useful when the team knows who says it, in what situation, and what help they need next.
  • Keep evidence notes close to the page brief. Detached keyword lists lose the customer context that makes them valuable.
  • Use respectful language in every sample. Real customer frustrations are evidence, not copy to exploit.

Worked example: a home-energy advisor

The team sees searches for ‘heat pump cost’ and assumes one generic price article is enough. Customer calls reveal a different question: homeowners want to know whether their older electrical panel, climate, and local rebates make a heat pump realistic.

The intent brief separates broad cost education from a decision guide for homes with older electrical systems. The guide uses local rebate sources, names the conditions that change cost, includes a checklist for an installer visit, and links to a transparent assessment request.

What changed: The business avoids a thin city-page rollout. It creates one deeper guide that helps a real decision and exposes the information needed for later local variations.

Make it stronger

Watch for hidden mixed intent

When results switch between explanations, products, maps, and news, split the research by audience or task. One page does not need to satisfy every possible meaning.

Use internal site search carefully

On-site searches are strong first-party clues, but spelling errors, navigation failures, and staff traffic can distort them. Sample the actual sessions before acting.

Use volume as a directional input

Volume can help you size demand, but small-volume questions may matter greatly for high-consideration purchases, local services, and retention.

Lesson artifact

Customer-language evidence log

Create an intent brief for one customer question your business hears often.

Raw evidence: Add three real customer phrases with their source and the context in which they appeared.

Hypothesis: Describe the task, audience, stage, likely format, confidence level, and what evidence supports your view.

Result review: Record the recurring result types, common promises, missing details, and any local or device differences you observe.

Page decision: Choose improve, create, research further, or do nothing. Explain how the response will help the customer better.

Done looks like this: Your proposed page is based on a real customer task and can be tested against both search behaviour and customer feedback.

Before you move on

  • I used real customer language, not only tool suggestions.
  • My intent label is a hypothesis with evidence and confidence.
  • I examined the result formats rather than copying a competitor page.
  • The proposed response fits what the business can honestly deliver.
  • I can explain why this page should exist as a distinct customer task.

Put the lesson into practice.

Create a free Spacebrain account and use the SEO suite with your own data providers.

Start for free →