Choose one buyer for one pass. A broad audience produces broad messages and muddled feedback.

Write the customer record before the message

1. Name one buyer

Write the role that owns the workflow and the situation that makes the problem relevant. “Small businesses” is too broad. “The person who reconciles supplier invoices before month-end” is specific enough to investigate, but it still needs evidence from your market.

2. Record one sourced problem observation

Use an exact interview note, support ticket, public complaint, sales email or observable workaround. Save the source location and date beside the observation. Separate what the person said or did from your interpretation.

3. Turn it into one message hypothesis

Write a message that a recipient can confirm, correct or ignore without sitting through a product pitch.

When [situation] happens, are you currently [observed workaround]? I am looking into a narrower way to [job], and I would like to understand how you handle it now.

Keep only the parts supported by the source. Do not add estimated time saved, customer counts or claims about what similar companies do.

4. Choose one channel test

Use the channel where this buyer already has context for the problem: a direct introduction, a relevant community, a reply to an existing conversation, search-led content or carefully researched outreach. Define the test by learning value and permission, not by a borrowed daily quota.

Keep a learning log

Date and sourceBuyer and observationMessage and channelWhat happenedWhat changes next
Link to the note or artifactRole, situation and observed problemExact message version and routeReply, correction, action or no responseKeep, revise or reject one assumption

Do not combine several buyer types, messages and channels in the same pass. If nothing happens, the log should still tell you which assumption to revisit.