Audience Segmentation Dimensions

Relationship Between Dimensions

WHAT they do       → Behavioural
WHO they are       → Demographic (B2C) / Firmographic (B2B)
HOW they think     → Psychographic
WHERE they are     → Geographic
WHEN/HOW READY     → Purchase Intent

No single dimension is sufficient. The strongest segmentation models layer at least three: typically one structural (Demographic or Firmographic), one motivational (Psychographic or Behavioural), and one contextual (Geographic or Purchase Intent).

1. Demographic

Who they are: objective, observable identity characteristics.

Pros: Data is cheap and widely available. Much of it is captured passively at signup or sourced from third-party providers. It is stable over time and indispensable for media planning on most ad platforms, which are built around demographic signals.

Cons: Tells you almost nothing about motivation. Two people with identical demographics can have opposite needs and values: a 35-year-old earning £80k might be a first-time homebuyer, a committed renter, or a serial mover. Demographics are increasingly poor at predicting preferences as lifestyles diverge from traditional life-stage patterns. Some attributes (ethnicity, religion) are legally restricted in several jurisdictions. Over-reliance on demographics is one of the main critiques Byron Sharp levels at conventional segmentation practice.

Key source: Philip Kotler & Gary Armstrong, Principles of Marketing (Pearson) — the standard academic marketing textbook, providing foundational coverage of demographic segmentation alongside all core marketing principles. Pearson

2. Psychographic

How they think and live: subjective, internal characteristics not visible from demographics alone.

Pros: Explains why people behave as they do rather than simply describing who they are. Enables more resonant messaging, brand positioning, and product decisions by connecting to underlying values rather than surface characteristics. Critical for differentiating brands beyond functional benefits.

Cons: Hard to measure directly. Psychographic data relies on self-report surveys, which are expensive, slow, and subject to social desirability bias (people describe who they aspire to be, not how they act). Difficult to operationalise in paid media, where most targeting is built on demographic or behavioural proxies. Byron Sharp argues that psychographic segments are often too small and too heterogeneous in actual purchase behaviour to justify the investment. Best used to inform tone and messaging rather than audience selection.

Key source: Arnold Mitchell, VALS Framework (SRI International, 1978) — the foundational psychographic segmentation system, formally inaugurated at SRI in 1978. Still maintained by Strategic Business Insights (SBI), though SBI's website availability has been intermittent as of 2025–26. Strategic Business Insights

3. Behavioural

What they do: observable actions, patterns, and history with a product or category.

Pros: Grounded in what people actually do, not what they say they do, making it among the most predictive inputs for personalisation, retention, and CRM. RFM has a decades-long track record in direct marketing. In digital products, behavioural signals are generated continuously and can be acted on in near real-time.

Cons: Requires data infrastructure to capture and act on, a significant barrier for early-stage products. The cold-start problem means new users have no behavioural history, making initial segmentation impossible. Past behaviour can mislead in fast-changing categories or after major life events (a new parent's consumption patterns shift dramatically). Privacy regulation (GDPR, CCPA) increasingly restricts the collection and use of granular behavioural data, particularly across third-party contexts.

Key sources:

• RFM model: an industry practice in direct mail marketing dating to the 1930s–40s, formalized by Arthur Middleton Hughes in Strategic Database Marketing (Probus Publishing, 1994). Jan Roelf Bult & Tom Wansbeek, Optimal Selection for Direct Mail (1995), provided the first rigorous academic statistical treatment of direct mail targeting and is frequently cited in the academic literature on the subject.
• For digital product context: Indi Young, Mental Models (Rosenfeld Media, 2008) — a UX methodology using mental model diagrams and task-based audience segmentation. Rosenfeld Media

4. Geographic

Where they are: location, spatial context, and environmental factors.

Pros: Among the easiest dimensions to collect. Location is captured passively at the IP or device level and is essential for any product with physical distribution, local pricing, or regulatory complexity. Language and time zone data directly inform localisation and support operations.

Cons: Increasingly limited as a standalone signal for digital products, where users are global and location correlates weakly with behaviour or preference. Remote work has further decoupled geography from lifestyle: a London postcode no longer implies a London income or cultural identity. Country-level segmentation can mask enormous intra-national variation (São Paulo vs. rural Brazil; London vs. the rest of the UK).

Key source: Claritas PRIZM — the dominant geodemographic segmentation system in the US market, combining geography with demographic clustering. Originally developed by Claritas (founded 1971); now operated as an independent company under The Carlyle Group ownership since 2017. Claritas

5. Purchase Intent

Where they are in the buying process: readiness and motivation to transact.

Pros: Directly actionable for funnel-stage targeting. Knowing whether someone is problem-aware, actively evaluating, or ready to convert determines the appropriate message, channel, and offer. Reduces wasted spend by matching content to readiness. In high-consideration B2B sales, intent data can meaningfully shorten sales cycles.

Cons: Notoriously hard to measure accurately. Behavioural intent signals (search queries, page visits, content downloads) are noisy proxies: research behaviour does not reliably indicate purchase intent, and most category research never converts. Self-reported intent is even less reliable, as survey respondents systematically overstate purchase likelihood. The linear funnel model underlying most intent frameworks (AIDA and its descendants) is a simplification; real decision journeys are non-linear, iterative, and vary enormously by category and individual.

Purchase intent is sometimes treated as a behavioural sub-category. It warrants its own dimension when funnel stage is a primary variable in the product or sales motion, which is common in AdTech, e-commerce, and long consideration-cycle B2B.

Key sources:

• E. St. Elmo Lewis, AIDA model (c. 1898–1900): the conventional attribution for the original purchase intent funnel. Note that Lewis's verified writings articulated three steps; the four-step AIDA acronym was coined by C.P. Russell in Printers' Ink in 1921.
• Jim Lecinski (Google), Winning the Zero Moment of Truth (2011): an updated purchase intent model for the digital era. Think with Google

6. Firmographic (B2B only)

What kind of organisation they are: the B2B analogue to Demographic; essential when buyer and end user diverge.

Pros: Large amounts of firmographic data are publicly available or cheaply sourced. LinkedIn, Crunchbase, Companies House, and similar registries cover industry, headcount, funding stage, and geography. Essential for ABM, where the unit of targeting is the account rather than the individual. Buying committee composition and budget authority vary predictably by company size, making firmographics a reliable proxy for sales motion complexity.

Cons: Describes the organisation, not the individual buyer. Two companies with identical firmographic profiles can have completely different internal cultures, priorities, and procurement processes. Firmographic look-alikes (companies that match an ICP on paper) frequently fail to convert for reasons firmographics cannot capture: wrong timing, internal politics, or a bad previous experience with a competitor. Tech stack data degrades quickly and is expensive to maintain. Funding stage is a particularly poor proxy for readiness, as a recently funded Series A is not automatically in buying mode.

Key source: "Firmographics" is a portmanteau of "firm" and "demographics" that entered common B2B marketing usage gradually from the 1970s onward, with no single documented originator. The concept relates to Shapiro and Bonoma's nested segmentation model (1984). It is now standard in ABM (account-based marketing) practice, a discipline pioneered by ITSMA in 2003. ITSMA

Further Reading