Choose which B2B customer segment to explore next

Start with two or three candidate customer contexts that could realistically use the same underlying offering.

Define each context through dimensions such as industry, company size, geography and customer role. Then keep as much as possible constant:

• the product or service
• the problem you believe it addresses
• the initial customer role where a reasonable equivalent exists
• the core questions you want to answer

This gives you a fairer comparison.

If one conversation is with a finance leader at a German enterprise and another with an operations manager at a Swedish startup, you are changing too many things at once. You may learn interesting things, but it becomes difficult to understand what deserves further investigation.

Do not ask simulated customers to vote on which segment you should pursue. Compare what you would need to learn for each segment instead.

Lens · What to explore

• Problem: Is the underlying problem present, and what happens when it remains unsolved?
• Workflow: Where does the problem appear in everyday work, and how is it handled today?
• Current workaround: What do people already do instead of using your proposed approach?
• Priorities: What else competes for attention, budget or organizational effort?
• Constraints: Which operational, technical, regulatory or organizational conditions could matter?
• Stakeholders: Who experiences the problem, influences change, approves it or must implement it?
• Evidence gap: Which important assumptions are still based on internal belief or simulated exploration?

The useful comparison is not “Segment A scores 78 and Segment B scores 63.”

It is “These are the assumptions we would need to resolve before making a serious investment in either segment.”

One candidate segment may raise a small number of familiar questions. Another may introduce a different workflow, more stakeholders and several assumptions you have never tested.

That does not automatically make the first segment better.

The second may have a more important problem. The first may prove difficult to reach or unwilling to change. Virtual Customer cannot settle those questions by producing more simulated agreement.

Instead, examine the learning burden behind each candidate:

What do we already know? What are we assuming? What would be expensive to get wrong? What can only a real customer, actual usage data or another external source resolve?

That gives you a basis for allocating the next research effort.

Illustrative example, not a customer result.

A B2B software company has developed a planning product for coordinating teams, resources and changing priorities. The team sees two possible expansion paths: field-service companies and logistics providers.

It is too early to decide that either industry is the better opportunity.

The team keeps the product concept, company-size range and core learning questions consistent. It explores equivalent operational roles through AI-simulated customer perspectives in each industry.

The conversations focus on how planning happens today, what forces plans to change, what information people depend on, which workarounds exist, what consequences poor planning creates and which other roles influence a new tool.

The two contexts may raise different questions. In field services, the team might need to investigate how local supervisors balance scheduled and urgent work. In logistics, a more consequential uncertainty might concern coordination across planning roles and systems.

Those are not findings about the industries. They are candidate research questions.

The team can now decide which real interviews or observations would reduce the most important uncertainty in each segment.

After exploring the candidate contexts, create a simple team-compiled map.

For each segment, record:

What appears reusable: assumptions already supported by evidence that may reasonably transfer.

What is uncertain: hypotheses exposed or challenged during the comparison.

What could change the decision: assumptions whose answer would materially increase or reduce the case for further investment.

What evidence is needed: customer interviews, observation, product data, technical research, market data or another appropriate source.

What to research next: the smallest credible next step that could resolve a consequential question.

This keeps research allocation separate from segment selection. You do not need to declare a winner before you know enough to make the comparison meaningful.

Virtual Customer can help teams examine customer contexts, expose different assumptions and prepare sharper research. It cannot tell you how large a segment is, how many companies will buy, what market share you can win or which segment will generate the most revenue.

Those questions require appropriate market, commercial and behavioral evidence.

A segment that produces compelling simulated conversations may still be difficult to reach, too small, poorly funded or already well served. A segment that initially appears less obvious may become attractive once real research reveals a severe problem and a practical route to adoption.

Use simulation to improve the questions. Use real evidence to make the investment decision.