Choose your first outbound segment

A first outbound segment should be small enough to recognize and useful enough to teach you something. “Companies that need more sales” will not help a researcher decide which account belongs on the list. “Independent accounting firms adding an advisory service, with a partner responsible for business development” gives the team something to investigate. The difference is a practical boundary, not a clever label.

Choosing that boundary is a judgment call. A scoring sheet can make the judgment visible, but it cannot manufacture customer evidence. Use the worksheet below to compare three plausible groups, surface the weakest assumptions, and select a pilot with a clear learning question. The point is to finish with a segment you can describe, find, and serve with your present resources.

Begin with a job someone already does

Write the job in the customer's language before describing your solution. A consultancy might need to keep next month's project calendar full. A software team might need to find implementation partners in a new region. Both involve business development, but the buyer, timing, and acceptable next step differ. Putting them in one campaign would make an ambiguous response harder to interpret.

For each candidate segment, finish this sentence: “When this situation occurs, this person needs to complete this job, and currently handles it this way.” Leave unknown details marked unknown. A blank is a research task. A confident guess often becomes an accidental requirement, then survives in the messaging long after everyone has forgotten where it came from.

Customer discovery offers a useful discipline here. NSF's I-Corps guidance describes customer discovery as a central part of its entrepreneurial training work. That does not validate your chosen market, but it supports the habit of talking with prospective customers while evaluating an idea. Your worksheet should capture things to investigate in those conversations, not act as a substitute for them.

Describe three candidates with the same fields

Give each segment a short name, then record company type, approximate team size, geography, the likely owner of the job, and an observable reason the job might matter now. Add the current alternative, including manual work or doing nothing. Finally, list one exclusion. Exclusions prevent the team from stretching the definition whenever an attractive company appears.

Suppose an illustrative research service is comparing specialist software consultancies, regional recruiting firms, and small industrial distributors. The consultancies might need introductions to businesses adopting a platform they support. Recruiting firms might need to identify hiring teams in one occupation. Distributors might need to open conversations with new purchasing accounts. These are hypotheses, not findings about those industries.

Use equivalent detail for all three. Comparing a thoroughly researched consultancy segment with “manufacturing” creates a false contest. If the distributor candidate is still broad, narrow it to a product category, buyer role, and selling region before scoring. You should be able to explain why a company qualifies without referring to how much you would like it as a customer.

Score evidence, access, fit, and effort

Use four categories with a simple zero-to-three scale. Evidence asks how directly you have observed the job and its consequences. Access asks whether you can identify appropriate people and reach them through suitable channels. Fit asks whether your current offer can help with the job. Effort asks whether research, delivery, and follow-up fit the team's available time. For effort, a higher score means more manageable work.

Define the scale before assigning numbers. Zero means there is no usable support or a serious obstacle. One means the case rests mostly on assumptions. Two means there is some direct support with unresolved gaps. Three means the claim has recent, relevant evidence and a workable next step. Use the numbers to start a conversation, not to create an appearance of precision.

Record one sentence of evidence beside every score. “Access: two, because the association directory identifies firms, but responsibility for partnerships varies” is more useful than a colorful cell. If two people disagree, preserve both reasons until you check the underlying fact. Averaging their scores would hide the question you need to answer.

Work through a hypothetical comparison

Imagine the consultancy candidate scores two for evidence, three for access, two for fit, and two for manageable effort. The recruiting candidate scores one, two, two, and three. The distributor candidate scores two, one, one, and one. These invented scores would make consultancies the provisional leader, but the total alone should not settle the choice.

Look at the weakest important field. If the consultancy offer requires a platform certification the team lacks, that could be a disqualifier despite the strongest total. If the recruiting segment includes several willing discovery contacts, it might offer a cheaper way to clarify the problem. Add a separate “must be true” line so a fundamental constraint cannot disappear inside arithmetic.

Then run a small sensitivity check. Lower the leading segment's most uncertain score by one point. Does your preference change? If it does, investigate that assumption before committing a month of work. If it does not, you have a reasonable basis for a limited pilot, provided you keep its uncertainty visible.

Translate the segment into inclusion rules

Turn the winning description into a short list-building instruction. For the consultancy example, require a public service page naming the relevant platform, evidence that the firm serves your intended market, and a plausible commercial contact. Exclude firms outside your service geography or those whose public offer does not match the job. Record the source supporting each inclusion.

Have a second person classify five accounts independently. Discuss disagreements before expanding the list. One person may interpret “specialist” as a dedicated practice while another accepts a single mention on a page. That difference will affect research time and message relevance. Fixing the rule now is cheaper than discovering later that half the pilot addressed a different audience.

Keep a short borderline list as well. Borderline accounts should help refine the rule, not quietly enter the campaign. If most prospects are borderline, the segment may depend on information you cannot reliably observe. In that case, choose a different screening signal or move the unresolved question into a discovery conversation.

Set a learning question and a stopping rule

A useful pilot question could be: “Do practice leaders treat partner introductions as an active problem, and what makes an introduction worth considering?” It is narrower than “Will outbound work?” It tells the team what to listen for even when a person declines a meeting. It also prevents a polite reply from being counted as proof of demand.

Decide how much research and conversation time you can spend before reviewing the choice. Pick a review date and write the possible decisions: continue, revise the boundary, change the offer, or stop. Do not treat silence as an explanation by itself. Poor contact selection, an unclear request, timing, and weak demand are different possibilities that may require different follow-up research.

Finish by putting all three segment sheets beside each other. Circle the strongest direct evidence, underline the most consequential assumption, and assign one person to resolve it. Your immediate next step is to score three customer segments and choose one bounded learning exercise. A useful first segment gives the team a shared decision rule and a reason to revisit it.