Run a small manual outreach experiment

A one-week, low-volume outreach experiment

Use this recipe when you need to learn whether a defined group recognizes a problem or will take a next step. It is not a bulk-marketing campaign. The goal is a decision about audience, message, or workflow—not a vanity count.

Day 1: set the question

Write one question, such as “Can we recruit five independent clinic administrators who recently reconcile appointment cancellations for workflow interviews?” Choose an audience rule that another person could apply. Define what evidence would change your next action. Set a time limit and stop rule.

Day 2: find 10–20 appropriate people

Use a legitimate source such as personal introductions, a professional directory, or a community that permits research invitations. Verify each person fits the criteria. Do not scrape prohibited sources, use private group data without permission, or contact people through a channel they have reserved for personal use. Keep a minimal private tracker.

Day 3: write and review the message

State your identity, why this person is relevant, what you are asking, how long it takes, and how to decline. If it is sales outreach, say so. If it is research, do not lead with a product pitch. Remove unsupported claims, fake personalization, urgency, and multiple asks. Check current outreach rules where sender and recipient are located.

Days 4–5: send individually and listen

Send a small number of messages that you have actually reviewed. Record sent, delivered if available, reply, opt-out, qualified conversation, and outcome. Do not infer that an unopened email was rejected or that a reply means purchase intent. Honor opt-outs immediately. Keep only the minimum suppression record needed to avoid contacting the person again; delete other notes when they are no longer useful.

In a conversation, ask for a recent example. Learn the existing workflow and workaround before presenting a concept. If you do show an offer, ask the person to explain what they believe is included and what would happen next.

Day 6: classify the evidence

Separate counts by the original denominator and fit. For example:

18 invited; 6 replied; 4 fit the audience; 3 conversations held.
All 3 described monthly reconciliation; 2 had a spreadsheet workaround.
One asked about the product; none made a purchase commitment.

This may support another workflow test. It does not establish market size or repeatable acquisition. Note counterexamples and what the sample cannot tell you.

Day 7: decide one next step

Continue with the same segment if people describe a recurring costly task and agree to an informative next step. Change the segment if fit is poor. Rewrite the message if people misunderstand the request. Stop if the channel is unwanted or community rules prohibit it. Record the decision and run only one meaningful variation next week.

Use the experiment brief to precommit the threshold and first 100 users to build a broader learning loop.