Market and Customer Research · Segmentation

The first step in marketing is not writing a message but choosing whose situation to target.

Questions Research Must Answer

In what situation, and toward what progress, is the customer trying to move?
What do they use to solve that problem today? (Including competitive alternatives and the status quo)
Who makes the purchase decision, in what order, and with what information?
Which group is willing to spend money and time on the problem?

The Jobs to Be Done Lens

The Christensen Institute describes JTBD as the progress a person is trying to make in a particular circumstance. The idea is that customers do not buy products; they "hire" them to make progress.

Bad example:

Our customers are developers aged 25–34.

Good example:

A developer running a new service alone who, when an incident happens, must find the cause within 10 minutes and explain it to the team.

Demographics are useful for reach (targeting) but do not explain why people buy. JTBD looks at functional, social, and emotional dimensions together.

Choosing a Research Method

Method	What you learn	Limitations
Customer interviews	Context, reasons, buying process	Small samples; words and actions can differ
Switch and churn interviews	Why they bought / why they left	Recall bias
Surveys	Scale, distribution, priorities	Question design shapes the result
Product analytics	Actual behavior	Cannot tell you why
Search queries and community observation	Customer language, alternatives	Unclear representativeness
Sales and support records	Recurring objections and requests	Bias toward the loudest customers
Competitor analysis	Market standards and gaps	Tends to show only the surface

Principle: Form hypotheses qualitatively; confirm size quantitatively.

Segmentation

A segment is "a group of customers similar enough to be handled with the same message and the same channel".

Basis	Example	Use
Firmographic (B2B)	Industry, size, region	Sales coverage, pricing
Demographic (B2C)	Age, region	Media targeting
Behavioral	Usage frequency, feature use, purchase history	Lifecycle, PQL
Needs / JTBD	Desired progress, situation	Positioning, message
Value	Willingness to pay, LTV	Budget allocation

Conditions for a good segment:

Identifiable (it can be separated in data)
Reachable (a channel exists)
Large enough (it can sustain a business)
Responds differently (there is a reason to treat it differently)

ICP and Persona

ICP (Ideal Customer Profile)
= The conditions of the most valuable, most likely-to-succeed "customer organization or group"

Persona
= The role and motivation of a "person" within it who is involved in buying or using

In B2B, a single purchase often involves users, decision makers, budget approvers, and security or legal reviewers. Personas are only useful when split by these roles.

Example: ICP for the Incident-Monitoring Tool

Written for this track's example product, the ICP looks like this. The conditions are pre-research hypotheses (assumptions) to be corrected with interviews and paid-conversion data.

Item	Condition (assumption)
Industry	SaaS companies, startups, and agencies that run their own web or app services
Size	1–5 developers, no dedicated SRE or operations team
Tech stack	Deploys to a public cloud, Git-based CI/CD, logs already collected in one place
Trigger events	First paying customer, refunds or churn caused by an incident, rising deploy frequency, being on call alone
Exclusions	Organizations that already have a dedicated operations team and a large observability platform contract, or that allow only on-premises installs (an early product cannot serve them)

Trigger events tell you when to reach out; exclusions tell you whom not to spend budget on. Without both, an ICP becomes a wish list.

The Buying Journey Is Not a Straight Line

Think with Google's "Messy Middle" research (2020) described the middle of the buying process as a loop in which exploration and evaluation repeat.

flowchart LR
    T[Trigger] --> E1[Exploration]
    E1 --> E2[Evaluation]
    E2 --> E1
    E2 --> P[Purchase]
    P --> X[Experience]
    X --> T

B2B is similar. In a Gartner survey of 646 B2B buyers conducted in August–September 2025 and published on 2026-03-09, 67% said they prefer a rep-free buying experience. In other words, customers have largely decided before they meet sales.

Research Outputs

Problem / JTBD statement
One or two priority segments and the reasons for choosing them
ICP definition and exclusion criteria
Role-based personas (where needed)
A list of words customers actually use
A list of competitive alternatives (including the status quo)

Common Mistakes

Defining everyone as a customer
Building segments from demographics alone
Asking in interviews, "Would you use it if it had this feature?"
Doing research once and stopping
Using research only to confirm a conclusion already reached

References

Jobs to Be Done Theory — Christensen Institute (accessed 2026-09-28)
The 'messy middle' and purchase behaviour — Think with Google (2020, accessed 2026-09-28)
Gartner Sales Survey Finds 67% of B2B Buyers Prefer a Rep-Free Experience — Gartner (2026-03-09, accessed 2026-09-28)
The B2B Buying Journey — Gartner (accessed 2026-09-28)