Your move
Choose one reachable buyer
Use one narrow list so the replies test one buyer, one problem, and one offer.
Choose one buyer type and name real people or businesses before you write a pitch. For this working session, ten names is the PickBits starter default—not a market law.
The situation
Start where the choice gets difficult.
You open a blank sheet and write ten names. A former customer comes first. Then a person who asked a sharp question last month. Then three local businesses whose public pages show the same messy handoff you know how to fix. The final rows take longer because you are no longer recalling people; you are searching for anyone who might fit. You send a careful note to the ten. Two people answer. One is curious. One says the timing is wrong. The other eight are quiet. Your first reaction is not to study those two answers. It is to make the list longer. Forty names feels safer than ten because a bigger sheet looks like a fuller pipeline. You start opening directories and copying businesses you have never watched, people whose role you do not know, and contact routes you have not checked. By row twenty, the original group is mixed with buyers who have different problems and different reasons to care. You are working harder, but the useful part of the list is getting harder to see. The question is no longer whether you can find more names. You can. The question is which names contain a real path to a useful conversation.
What changes the decision
Read the evidence before you add more work.
The two replies are not proof that this buyer group will work, and silence is not proof that it will not. A ten-name starter batch is too small for market claims. It is large enough to reveal where your evidence is strongest. Look at how the names entered the sheet. The replies are more likely to come from rows where you can state the shared problem, point to a visible signal, and explain why your route to that person is appropriate. The weakest rows usually need a story to connect them to the offer. Adding more of those rows does not strengthen the test. It hides the contrast between strong and weak reach paths.
Your job is to preserve that contrast. Mark the names you can reach for a truthful reason. Keep the names that share one problem and one buying role. Set aside names that merely resemble your imagined customer. Then ask what the two replies teach you: which signal was visible, which relationship made the note natural, and which part of the problem the person recognized. Do not turn two replies into a forecast. Use them to decide where the next small batch should come from. A short list with reasons gives you a test you can read. A long list assembled to avoid choosing gives you activity you cannot interpret. When a few well-supported rows produce nearly all the useful movement while the long weak end consumes most of the search, this is called the Pareto tail.
Do this now
Three moves, in order.
- Start with people you already know, past conversations, referrals, local businesses, or a directory you already trust.
- For each name, record the visible situation that suggests the problem exists. Use evidence; do not invent urgency.
- Keep one group whose problem, buying role, and reachable channel are similar enough to test one offer.
Working session
Build it from the evidence in front of you.
- Write ten real names before opening a new directory
Use people, businesses, past conversations, referrals, events, and sources you already know. Ten is a PickBits starter default, not a required market size. If you only have six evidence-backed names, work with six instead of inventing four.
- Attach one visible signal to every row
Record what you actually saw or heard and where it came from. A public service page, a question asked directly, or a permitted referral can support a row. A guess about revenue, urgency, or authority cannot.
- Circle the shared buying situation
Read the rows without looking at your service. Keep the group that shares a recognizable problem, a similar role in deciding, and a route you can use honestly. Move the rest to a parking list instead of forcing a fit.
- Choose the source of the next small batch
After replies arrive, compare strong rows with quiet rows. Pick one evidence source to revisit. Do not simply choose forty more names. Choose where another small, readable group of names should come from.
Inspect the finished shape
A ten-buyer list you can inspect
BUYER GROUP
Shared problem: [plain-language problem]
Likely buying role: [role shared by this group]
Starter batch size: 10 names, or the smaller evidence-backed set you have
KEEP | [real person or business] | Signal: [what you observed] | Source: [where/date] | Reach path: [permitted route]
KEEP | [real person or business] | Signal: [what you observed] | Source: [where/date] | Reach path: [permitted route]
CHECK | [real person or business] | Missing: [problem evidence / role / reach path]
PARK | [real person or business] | Reason: [different problem, role, or no honest route]
AFTER THE BATCH
Replies came from rows: [row numbers]
Shared evidence in those rows: [observation, not a claim]
Source for the next small batch: [one place]
What I will not do: add names only to make the list look fullBuild the artifact
Ten-buyer list
Copy these fields into a note, document, or sheet you already use.
Give AI the bounded job
Prompt from your artifact—not from vibes.
Organize my notes into a buyer list. Do not add facts or contacts I did not provide. Group only the buyers who share the same observable problem. Put missing evidence in a column called NEEDS CHECK.
Before you use it
The work leaves your hands only when:
- Every row names a real buyer and an evidence source.
- The contact route is appropriate and permitted.
- You can explain the shared problem without naming your solution.
Failure patterns
Watch for the move that looks useful but is not.
Directory drift
- The tell
- The first names have a reason beside them. Later names only have an industry label or a copied web address.
- The fix
- Stop the search. Move unsupported names to CHECK and return only when you can add a visible signal and an appropriate reach path.
The forty-name reflex
- The tell
- Two replies from ten make you want four times as many names before you have read what made those two rows different.
- The fix
- Compare the evidence on replied and quiet rows first. Let that comparison choose the source of the next small batch.
Solution-shaped grouping
- The tell
- Everyone on the list could use what you sell, but they do not share the same problem, buying role, or reason to act.
- The fix
- Describe the group without naming your service. If the sentence falls apart, split the list until one clear buying situation remains.
Use the boundary
When this rule does not fit.
Do not narrow to one buyer group when you already have clear, current demand from several distinct groups and enough capacity to serve them separately. Do not use a short-list exercise to ignore a buyer who arrives through a trusted referral merely because they sit outside your current pattern. And do not keep contacting a small group after its members have declined or asked you to stop. The point is a readable test, not permission to pursue the same people indefinitely. Respect each answer and the rules of the channel you use.