THE COMPANY-BUILDING FIELD NOTEBOOKRESEARCH EDITION / SEPTEMBER 2026
Startup
Research.
Search
EARN CUSTOMERS / PRACTICAL GUIDE

Choose the first customer segment by constraint, not charisma

Segment prospects by trigger, workflow, buyer, and budget path, then make the weighting judgment visible instead of declaring an intuitive ICP.

  1. Trigger
  2. Budget owner
  3. Access to buyers
Conceptual relationship map, not measured data or a guaranteed sequence.

An early segment is not 'SMBs' or 'healthcare.' Those labels describe a census bucket, not a repeatable sale. The first useful segment shares a trigger, an operating workflow, a buyer and budget path, and constraints that let one product and sales motion serve several customers.

The validation chapter starts with a specific person, problem, workaround, and willingness to pay; read the wider sequence at validation chapter.

Build segments from buying context

Start with observed variation in interviews and pilots. Group accounts by:

  • Trigger: what changed now—new rule, headcount threshold, failed audit, backlog, new contract.
  • Workflow: how the work is done, including handoffs and systems.
  • Buyer: who owns the outcome and who can release money.
  • Budget path: existing line item, discretionary spend, project approval, or new budget.
  • Constraints: integration, security, regulation, geography, service load.

GOV.UK guidance frames a user need with the person, goal, trigger, and constraint, and warns against defining the need as a solution (Learning about users and their needs). It is service-design guidance. The editorial adaptation is to add payer and adoption mechanics.

Reject broad labels that hide variance

NAICS can count establishments by production activity, but it is not a buying-behavior model (Census NAICS). Two firms in one code may have different triggers, software stacks, and authority structures. Conversely, two industries may share a workflow but face different rules. Use industry as a filter only when it predicts something operational.

Use a weighted selection, openly

Pick five or six criteria and weights before scoring. A reasonable starting set is pain frequency 25%, urgency 20%, budget access 20%, reachability 15%, workflow similarity 10%, and delivery burden 10%. Those weights are editorial judgment, not a universal formula. If runway is short, reachability may deserve more weight; in a safety-critical market, feasibility and approval burden may dominate.

Score only from evidence. Mark unknown rather than converting optimism into a three out of five. Then run the result as a hypothesis, not a declaration.

Worked hypothetical selection

Hypothetical: A compliance-workflow founder compares three groups.

Segment Pain 25 Urgency 20 Budget 20 Reach 15 Similarity 10 Low burden 10 Weighted result
Regional lenders after an exam finding 5 5 4 3 4 2 4.10
Fintech startups preparing first audit 4 4 3 5 3 3 3.75
Large banks modernizing controls 5 3 5 1 2 1 3.30

The founder selects regional lenders for the next six weeks. That is a judgment: the trigger and budget path outweigh service burden. The table does not prove the segment will buy. It makes the tradeoff inspectable and sets up a test.

Segment decision artifact

Write a one-page segment contract:

  • Included trigger, workflow, organization shape, buyer, and geography
  • Explicit exclusions
  • Evidence behind each score and unknowns
  • One sentence explaining why this segment now
  • A six-week acquisition and delivery test
  • Advancement evidence: for example, three paid trials with the same core scope
  • Reconsideration evidence: no budget path, incompatible workflows, or custom work above a set cap
  • Date for rescore

Limits

A weighted matrix can dignify weak inputs. Scores are ordinal judgments, and small differences are not meaningful precision. The first segment is also not a permanent brand boundary. Choose it to compress learning and establish repeatability, then revisit when evidence shows a neighboring group can use the same core product and motion.

Sources & scope

Sources checked 19 September 2026. Worked scenarios are illustrative; recommendations are editorial analysis. These checks do not re-verify the entire original notebook.

  1. Learning about users and their needs — UK Government Service Manual

    The guidance frames user needs around goals, triggers, and constraints and distinguishes needs from proposed solutions.

    Source publication date: 2016-04-04 · Retrieved 2026-09-19

  2. North American Industry Classification System — U.S. Census Bureau

    NAICS classifies establishments by production activity for statistical use; it does not define buying behavior.

    Source publication date: Not established · Retrieved 2026-09-19

Developed from the original notebook

Keep the question moving.

Next in this path: Build a founder sales pipeline that records rejection

All practical guides →