Why the first sentence is often the wrong place to start
A team finds a name, scans a profile, and adds one personal sentence to a message whose audience and offer were never examined. The sentence may be accurate; the reason for contacting that person may still be weak. Research should establish the connection between an offer, a company condition, a responsible role, and a useful next step before the first line is written.
This is an operating method, not a claim that more research always produces more replies. Its value is that the team can explain why each company and person is on the list.
The offer profile: what you can actually defend
Describe the service in the language a buyer would use to recognise the problem. Separate what the company demonstrably does from what it hopes to do. Keep proof points, exclusions, and claims that require qualification next to the offer.
For a security advisory firm, “we help prepare evidence for enterprise vendor reviews” is a usable offer. “We guarantee your next enterprise deal closes” is not. If this profile is wrong, every downstream message can be wrong with greater fluency.
The market map: what the buyer can do instead
A competitor list alone is too narrow. The buyer may build internally, hire a generalist, buy software, accept the current process, or postpone the project. Document those alternatives and what would make a specialist useful.
For the security advisory example, the real alternative may be asking an engineering lead to answer procurement questionnaires between product deadlines. That changes how the introduction should be framed: as focused support for a recurring workflow, not a grand replacement for the team.
The audience hypothesis: a testable definition
Specify the company conditions that might make the offer relevant, the function that owns the work, the geography and size that can realistically engage, and who should be excluded. A title by itself is not a buyer definition. Apollo’s persona guidance makes the same practical point: roles are useful when combined with other prospecting filters.
Write the hypothesis so it can be rejected. “Every SaaS company” cannot be tested meaningfully. “B2B software vendors beginning to sell into regulated enterprise accounts, with a security or engineering owner and a public vendor-review process” can.
Research: Apollo persona guidance
The prospect reason: an internal quality check
For each company, record a sourced observation, the possible consequence, and the role that could own it. For example: “The company has published an enterprise security review process; answering larger buyers may create more evidence work for its security team.” The first clause is an observation. The second is an inference. Keep the distinction visible.
Do not paste the entire internal reason into an email. A private-looking diagnosis based only on public data can sound presumptuous. The prospect reason exists to decide whether an introduction is warranted and what it may honestly say.
The message strategy: one useful next step
Choose the point of view before generating a draft. What single observation is relevant? Which part of the offer addresses the likely work? What is a low-friction question the buyer could answer?
In the example, an opening could acknowledge the public enterprise review process and ask whether evidence preparation is already handled or worth a short conversation. It should not claim that the company has failed a review, is behind schedule, or has an approved budget. It should also respect applicable commercial-email rules, including accurate sender information and opt-out handling. The FTC makes clear that its CAN-SPAM rules have no business-to-business exception.
Research: FTC CAN-SPAM guide
A practical research threshold
Not every company needs an exhaustive dossier. Before a draft is made, require a current source for the company observation, an offer claim you can defend, a plausible responsible role, and a sentence that keeps facts separate from inference. If any one is missing, pause. This is a quality threshold, not a promise of reply rate.
The threshold can be applied in a few minutes to a narrow set of accounts or automated across a larger market. What matters is that the system preserves the chain. A faster workflow that silently drops the source or turns a possibility into a fact is not better research.
Put the artifacts in one chain
The five documents should not become five disconnected PDFs. The prospect reason should inherit the offer and audience definition; the draft should inherit the prospect reason; a changed offer should make old drafts suspect. That is the practical purpose of a connected workflow.
The fastest useful research is not exhaustive. It is enough to answer four questions without bluffing: What do we solve? Why this company? Why this person? Why this message now? If one answer depends on a guess, either research more or leave the company out.
What this method does not establish
A sourced company condition does not establish buying intent. An identified role does not establish budget. A verified address does not establish permission, relevance, or deliverability. Keeping these limits visible protects the usefulness of the research: it tells you where a respectful question is warranted and where further evidence is still needed.
Review the chain when the offer changes, a source becomes stale, or a recipient replies with a correction. The best research artifact is not the longest document; it is the one a teammate can inspect, update, and use without inventing the missing facts.
Sources and further reading
Apollo: create and use a persona Apollo: create and use a signal FTC: CAN-SPAM compliance guide for businessPut the method into practice with nteni.
Research your own offer