Intelligence 01
Know every person who can say no.
Strong B2B customer research turns an ideal customer profile (ICP) from a firmographic description into a buying system made of real people. It captures what each person hopes to gain, fears losing, trusts, questions, and needs to believe before the group can reach yes.
From Label To Empathy
Industry, company size, title, and budget help locate a market. They do not explain how someone experiences a problem or which message will resonate. That requires a deeper picture of the person's world.
Fears, hopes, pressures, identity, interests, trusted voices, and exact language all matter. Not every detail belongs in the copy. The details build the empathy required to make the copy accurate.
Role, company stage, industry, team size, geography.
Responsibilities, pressure, constraints, current alternatives, and where time disappears.
Hopes, fears, reputation, identity, ambitions, interests, and what they cannot afford to get wrong.
Buying triggers, proof standards, objections, exact language, preferred sources, and reasons to stay with the status quo.
The B2B Reality
A single buyer persona can hide the real decision. The person who wants the product may not control the budget. The person who controls the budget may not use it. Security, legal, procurement, or another operator may be able to stop the purchase without ever appearing in the campaign brief.
These roles are illustrative. The actual map may include an executive sponsor, technical evaluator, procurement lead, department owner, or several end-user groups.
Candid Feedback
No research format is always best. The right method is the one that reduces social pressure, exposes actual behavior, and gives people enough safety to disagree with the company.
Probe past the polite first answer and ask for concrete moments, choices, and consequences.
Remove the social cost of saying the product, message, or idea is weak.
Compare what people say with what they actually click, ignore, repeat, or abandon.
Show the work before explaining it. Their unaided reaction is the useful evidence.
Reconstruct the sequence, stakeholders, doubts, alternatives, and moment the decision changed.
Use it to transcribe, retrieve exact language, compare reactions, and surface patterns. Any fear, hope, motive, or veto condition that has not been observed remains a hypothesis until a real person supports it.
Current Tools
The stack depends on the missing layer. Some teams need better access to hard-to-reach people. Others need a research memory, a repeatable customer panel, or a way to test messaging with a precise B2B audience.
Default approach: start with direct access to customers and the company's existing call tools. Add software only where recruiting, organization, or repeated testing is slowing the feedback loop.
Examples Worth Studying
The useful pattern is not the tool each company selected. It is the decision to put an internal belief in front of real people and allow the response to redirect the product or message.
Rahul Vohra described segmenting feedback around the users who would be most disappointed without Superhuman, then building a detailed picture of the high-expectation customer the company could serve best.
Principle: learn most deeply from the people receiving the strongest value, then use their language to sharpen the focus.
Read the sourceIntercom reported conducting 45 customer calls and closely studying 15 of them to identify four jobs customers hired the product to perform. Those jobs helped reshape its products and go-to-market story.
Principle: motivations can connect buyers that firmographic labels divide.
Read the sourceWynter reports that Appcues tested key pages with product managers and marketers in its target audience, then used their direct reactions to revise the messaging and design.
Principle: real audience reactions expose confusion that an internal review can easily miss.
Read the sourceThe Working Loop
An ICP profile is a living hypothesis. It becomes more accurate when messages, ideas, offers, and products repeatedly meet the people they are intended to move.
List every likely participant who can influence or veto it.
Record what is known, what is assumed, and what still needs proof.
Use a message, concept, offer, page, or product experience.
Choose the format most likely to produce honesty.
Preserve exact language, revise the profile, and change the work.
Tool sources: Dovetail product information, User Interviews pricing and audience details, Wynter Research Playbook, and Great Question product information.
Example sources: Rahul Vohra's Superhuman account, Intercom's Jobs-to-be-Done account, and Wynter's Appcues case study.
Sources and tool information reviewed Q3 2026. Tool fit should be rechecked before procurement.
Intelligence 01
Keep putting the work in front of real people. The quality of the message will follow the quality of the empathy behind it.