Copywriting Research Process: From Evidence to Message Strategy

Copywriting research becomes useful when it supports a decision.

Interviews, reviews, competitor pages and product documents may all contain useful material. Yet a folder full of sources does not show what the copy should emphasize. You need a process that connects evidence to a specific sales question.

This copywriting research process produces five working outputs:

1. Prioritized prospect concerns
2. Documented connections between those concerns and the product
3. Clear limits around the available evidence
4. Message and lead hypotheses
5. A register of unresolved claims

A lead is the opening approach of a sales message. A message hypothesis is a provisional direction supported by available evidence. It is not a proven winner.

The central principle is to begin with prospect priorities rather than the feature the company most wants to promote. [1]

Sales question
      ↓
Separate source groups
      ↓
Context-rich source records
      ↓
Voice-of-customer coding
      ↓
Priority and capability mapping
      ↓
Message and lead hypotheses
      ↓
Working research brief

Start with the decision the research must support. Record:

• The intended audience
• The offer
• The action the audience is being asked to take
• The uncertainty preventing confident drafting

Imagine a software product called Relay.

A broad prompt would be:

A decision-focused question would be:

The second question directs the research without assuming that automation is the audience’s main concern.

Separate facts from assumptions

This table stops an internal belief from quietly becoming a research conclusion.

Collect material in four groups:

• Prospect evidence: interviews, surveys, reviews, support conversations and sales calls
• Product documentation: capabilities, limitations, demonstrations and technical records
• Competitive messaging: positioning, offers and common category language
• Proof sources: primary documents, tests and other evidence used to support claims

These sources have different roles. A customer anecdote can document one person’s experience. A product specification can document a capability. Neither automatically establishes a broad outcome.

A proclamation-style lead opens with a bold central assertion. That assertion should emerge from supporting research, rather than being invented first and justified afterward. [2]

Record four evidence layers

Assign a stable ID to each useful item.

Customer language is evidence of what someone said or perceived. It is not automatically true, representative or suitable for an unqualified claim.

Keep product facts separate from prospect priorities as well. “Relay displays task ownership” is a documented capability. It does not establish that ownership is the audience’s most important concern.

Voice of customer, or VOC, is the language prospects and customers use to describe their situations, priorities and doubts.

Save the context around every excerpt:

• Who said it and where
• What situation prompted it
• Which product or alternative was under discussion
• Whether the person fits the intended audience
• Whether the statement describes a perception, behavior or checkable fact

A Relay source log might begin as follows:

Context prevents a vivid quotation from being stretched beyond what its source supports.

Coding makes records comparable while preserving the original language.

Evidentiary strength means what a source can legitimately support, along with what it cannot establish.

Voice-of-customer coding sheet

Completed hypothetical record

Do not invent scoring formulas, frequency thresholds or sample-size rules without a defensible measurement method.

Treat awareness as a conditional signal

Broadly, a familiar and ready audience may permit a more direct opening. A skeptical or less familiar audience may require more explanation or a gentler bridge. [3]

Use tentative coding notes:

• Familiar with the category but skeptical about switching
• Recognizes the problem, but knowledge of available approaches is unclear
• Readiness cannot be inferred because the source was not discussing a purchase

These notes are planning inputs. They are not definitive awareness-stage assignments or rules for choosing a lead.

Group related pains, desires and objections without erasing differences between sources.

Frequency shows recurrence within the material collected. It does not prove effectiveness or market-wide importance. Emotional intensity is also insufficient by itself.

A problem-solution direction should reflect a prospect concern and introduce relevant hope without dwelling needlessly on distress. [4]

Next, map the clusters to documented capabilities:

Remove themes that cannot connect credibly to the product. The opening, claims, benefits and offer should remain organized around one unifying idea. [5]

Each message hypothesis should retain its evidence trail.

A mechanism is the proposed explanation connecting a capability to a result.

For example:

This is a hypothesis, not a validated performance claim. Each arrow needs support. If one link is missing, record it as an open question rather than writing it as fact.

Practitioner frameworks describe recurring openings such as Offer, Promise, Problem-Solution, Big Secret, Proclamation and Story. Use these as overlapping labels for exploration, not rigid or mutually exclusive categories. [6] [7]

Possible leads for MH-01 include:

• Direct: “See current task owners and recorded workflow changes in one place.”
• Problem-solution: “When ownership disappears between handoffs, the next meeting becomes an investigation.”
• Story: “On Monday, Maya’s workflow looked complete. By Thursday, nobody could explain who changed the approval step.”

All three are original, hypothetical lines.

Indirect leads can use emotion or curiosity, but they may become slow or disconnected from the product. [8]

Record the bridge for each indirect opening:

Remove the lead if it cannot make that transition clearly.

Curiosity gaps around an overlooked component or process appeared in six examples within one stratified convenience sample. The repeated structure revealed the point before explaining it. This is only a sampled pattern, not evidence of effectiveness or a market-wide norm. [9]

A restrained hypothetical adaptation might be:

The copy must then explain the issue plainly. Curiosity does not permit a fabricated secret, hidden cause or unsupported claim.

The final brief should make drafting decisions and evidence gaps visible.

Maintain an unresolved-claims register:

This process produces organized evidence and traceable hypotheses. It does not prove that participants represent the market, establish that a message will convert or replace legal and specialist review.

For broader research context, see /copywriting/research-voc/. Use the separate research checklist for execution tasks and the research template for a reusable document.