In short: buyer personas are a legitimate tool born in UX research (Cooper 1999) but misused in strategic marketing. They work for product design, conversational AI and long-form copy. They fail as a basis for advertising segmentation and media planning, because they overlook light buyers (70-80% of revenue) and create illusions of homogeneity within segments. Evidence-based alternatives: Category Entry Points (Ehrenberg-Bass), Jobs To Be Done (Christensen).
Origin of personas: UX research, not marketing
The modern buyer persona was formalised by Alan Cooper in "The Inmates Are Running the Asylum" (1999), then refined by John Pruitt and Tamara Adlin in "The Persona Lifecycle" (2006). The original goal was specific: help software designers empathise with real users when they only met a few during projects. The persona was a narrative construct based on direct ethnographic interviews, to be used in interface design.
The mechanism is: instead of "the average user", I build "Marina, 38, project manager at a manufacturing company with little time, frustrated by Excel multitasking". This helps designers make consistent interface choices. It works well for UX because: (1) decisions are many and granular (microinteractions); (2) the number of actual users is limited; (3) needs change slowly.
When personas actually work
(1) UX design and product design. To choose how to structure an interface, which microcopy to use, which features to prioritise, personas remain useful. Cooper's original use case is still valid.
(2) Conversational AI and chatbot design. To design the conversation arc of an AI agent, modelling a prototypical user is effective. It is the same activity as interface design, applied to a conversational channel.
(3) Long-form B2B copywriting. To write whitepapers, case studies, B2B email sequences with sales cycles of months, a decision-maker persona helps calibrate tone and technical depth. It works because in B2B the number of buyers is genuinely small and homogeneous.
(4) Customer support and onboarding. Modelling 3-5 archetypes of new users helps design onboarding flows and proactive FAQs.
In all these cases, the persona is a product-centred design tool, not a market map. It is used to make experience choices, not to allocate budget.
When personas fail
(1) Segmentation for mass advertising. The most common mistake is turning 3-5 personas into 3-5 advertising segments, allocating budget to each. This violates the Pareto law of buying behaviour documented by Ehrenberg-Bass: in most categories, 80% of revenue comes from occasional light buyers who do not match any archetypal persona. "Marina the 38-year-old project manager" is a caricature, not a high-penetration market segment.
(2) Media planning. Buying audiences that mirror the personas (e.g. "Meta targeting: woman 35-44, manager, northern Italy, business interests") concentrates spend on a narrow and homogeneous public. Sharp and Romaniuk show that brands grow by reaching a broad category of buyers (mass-marketing reach), not by focusing on narrow segments.
(3) Pricing strategy. Personas often suggest pricing tiers (e.g. "Marina wants entry-level €19/month", "Carlo the executive wants premium €99/month"). Real consumer data (NielsenIQ POS, Circana) show that willingness to pay varies more within each "persona" than between personas. Pricing tiers based on personas are a simplification that masks the real variance.
(4) Content marketing strategy. Creating 1 persona = 1 content cluster leads to niche content that does not scale. Real Search Console queries rarely follow persona boundaries: a "Marina" persona may search for a topic the marketer had assigned to "Carlo".
Light buyers and the representation problem
Sharp ("How Brands Grow Part 2", 2015) devotes an entire chapter to the fact that personas systematically exclude light buyers. The light buyer purchases once or twice a year, with low loyalty, on specific occasions (CEPs). They have no distinctive demographic profile, no identifiable "passions", do not associate with a cause: they buy because they find themselves in a context of need and the brand is mentally available.
Light buyers are heterogeneous, statistically average, demographically normal. They do not produce a memorable persona. But they generate the majority of revenue. Persona-based segmentation discards exactly them, opting for the colourful "heavy buyers" (Marina the demanding project manager) who however are 5-10% of the market.
Alternative 1: Category Entry Points (CEP)
CEPs are contextual triggers that activate thought of the category. For coffee: "work break", "breakfast", "guest at home", "personal reward". For B2B marketing agencies: "new product launch", "declining revenue", "post-merger rebranding", "new CMO".
The CEP framework (Romaniuk 2018) proposes: identify the 5-10 most frequent CEPs in the category through interviews and surveys, map the brand's existing associations with each CEP, build communication that reinforces the most strategic associations. It is measurable (CEP coverage score), scalable, and includes light buyers because it is not based on demographic profiles but on consumption occasions.
Key difference from personas: a CEP is not "who buys" but "when one buys". It is a mapping of demand, not of the consumer.
Alternative 2: Jobs To Be Done (JTBD)
Christensen et al. ("Competing Against Luck", 2016) propose JTBD: consumers do not buy products, they "hire" products to do a job. Classic example: McDonald's morning milk-shake is not a dessert but "something that keeps me busy during the commute".
JTBD is useful to: ideate new products, reposition existing products, identify non-obvious competitors (the milk-shake competes with the granola bar, not with other desserts). Unlike personas, JTBD is centred on the problem to be solved, not on the type of person solving it.
The operational framework: identify the "job" (functional + social + emotional), map progress measures, identify obstacles, design the solution. It is particularly effective for B2B SaaS, durable consumer products, professional services.
CEP setup for an Italian SME: concrete workflow
- Qualitative interviews (15-25 buyers). Open questions: "How did the last purchase in this category happen? What happened beforehand?" Transcribe and code recurring themes.
- Quantitative survey (300-500 respondents). For each candidate CEP, ask "How often does this situation happen to you?" and "Which brand comes to mind first?" on a Likert scale.
- CEP-Brand map. For each CEP, calculate share of mind of your brand vs competitors. CEPs with low association are growth opportunities.
- Communication planning. Build messages and creative that anchor the brand to the 3-5 most strategic CEPs.
- Measurement. Re-test every 12-18 months to track CEP coverage evolution.
Decision table: which framework to use
| Decision | Personas | CEP | JTBD |
|---|---|---|---|
| Product UI/UX design | ✅ | — | ✅ |
| Advertising media planning | ❌ | ✅ | — |
| Brand positioning | — | ✅ | ✅ |
| New product development | — | — | ✅ |
| Long-form B2B copywriting | ✅ | — | ✅ |
| Pricing strategy | ❌ | — | ✅ |
| Content marketing/SEO | — | ✅ | ✅ |
FAQ
Should I remove personas from my marketing playbook?
No. Keep them for the use cases where they work (UX, product design, long-form B2B copy). Drop them as a strategic framework for segmentation, media planning and brand positioning. The rule: personas as a design tool, not as a market map.
How do I explain to a client that personas should not be used for advertising?
Show data. Open the "How Brands Grow" report or an Ehrenberg-Bass case: brands that win do so by reaching a broad audience, not narrow targets. Show that 70-80% of revenue comes from light buyers not represented in the personas. Propose CEPs as a measurable alternative.
Are data-driven personas different from classic ones?
Only marginally. Even when built with cluster analysis on CRM data, they remain archetypes that simplify. The problem is not the quality of the input data but the nature of the construct: reducing millions of buyers to 3-5 archetypes erases the real variance, especially light buyers.
What do I say if the client insists on personas?
Reframe them as a design tool (useful) and separate the strategic decision (segmentation, budget, media). Work with two tools: personas for experience design, CEPs for budget and media allocation. They work in parallel.
Are JTBD and CEP the same thing?
No. JTBD identifies the "job" the product must do for the consumer (e.g. "keep me awake during the commute"). CEP identifies the occasion that activates thought of the category (e.g. "driving to work"). JTBD is better suited for product development, CEP for brand communication. They are often used together.
Are there Italian case studies of companies that abandoned personas?
Publicly rare, because companies seldom expose their frameworks. Well-known international cases: Mars Petcare publicly migrated to a CEP framework with documented market share gains (Ehrenberg-Bass case). Coca-Cola and Unilever use CEP+DBA instead of personas for brand planning. In Italy, some Barilla and Ferrero divisions have integrated CEPs into planning (references in Effie Awards Italia proceedings).
Sources and references
- Cooper, A. — "The Inmates Are Running the Asylum" (1999, Sams Publishing)
- Pruitt, J. & Adlin, T. — "The Persona Lifecycle" (2006, Morgan Kaufmann)
- Sharp, B. — "How Brands Grow Part 2" (2015, Oxford University Press)
- Romaniuk, J. — "Building Distinctive Brand Assets" (2018) and "Better Brand Health" (2023)
- Christensen, C. et al. — "Competing Against Luck: The Story of Innovation and Customer Choice" (2016, Harper Business)
- Ehrenberg-Bass Institute — Category Entry Points research papers (2017-2023)
- NielsenIQ & Circana — POS panel reports on light/heavy buyers
- IPA Effectiveness Awards — case studies documenting the shift from personas to mass marketing
