Segmentation

segmentation

customer-understanding

Segmentation divides an audience into distinct groups based on shared behaviors, attitudes, or needs — enabling targeted strategy, messaging, and product development

Author

TVE Analytics Team

Published

March 27, 2026

Overview

Segmentation divides a broad audience into meaningful, distinct groups so that each group can be understood and targeted differently. The goal is to move beyond demographics and uncover groups that genuinely differ in what they need, how they behave, or what motivates them.

TVE uses a range of analytical approaches depending on the project objectives — from simple clustering on survey responses to full strategic segmentations with typing tools for ongoing classification. The analytics team selects and configures the right model based on the data, the business question, and how the segments will be used.

What the analytics team handles:

Selecting the appropriate clustering method (K-means, K-prototypes, latent class analysis, archetypes, decision trees, and more)
Testing multiple solutions with different input combinations
Evaluating and recommending the strongest solution based on statistical and strategic criteria
Building typing tools for ongoing segment classification (when required)

What consultants need to provide is covered below — the inputs, objectives, and desired outputs that shape the analysis.

Types of Segmentation

Not every segmentation requires the same level of complexity. The approach depends on what the segments will be used for:

Type	Description	Example
Simple clustering	Groups respondents based on a small set of survey questions to understand behavioral patterns. No typing tool needed.	Clustering gamers based on how they approach playing a game (story-focused vs. exploration-focused vs. completionists)
Strategic segmentation	A comprehensive segmentation designed for long-term use across the business. Includes multiple tested solutions, full profiling, and a typing tool for classifying new respondents.	Segmenting crafters based on motivations, attitudes, and engagement to inform brand strategy, product development, and communications
Needs-based grouping	Maps existing demographic or firmographic groups against their needs and behaviors. May not require statistical modeling — instead consolidates known groups based on observed need patterns.	Grouping fleet decision-makers by their EV adoption status and charging needs to identify which segments have the strongest product-market fit

What Makes a Strong Segmentation

A strong segmentation is not just statistically sound — it must be usable by the business. Key criteria:

Viable segment sizes — not so small they can’t be found, not so large they lack clear identity
Clearly differentiated — homogeneous within each segment, distinct from all other segments
Not driven by demographics alone — to avoid segments that simply reflect life stage or geography
Stable over time — built on inputs that don’t shift rapidly, so segments remain useful for strategic planning
Easily explainable — segments should feel intuitive and be easy to communicate internally
Actionable — clear hooks within each segment that the business can act on

Business Questions This Answers

Segmentation is ideal when clients ask: “Are all our customers the same, or are there meaningfully different groups?”

It helps answer strategic questions such as:

Audience understanding: Who are our customers, and how do they differ in needs, behaviors, or motivations?
Targeting: Which customer groups should we prioritize, and how should we reach them differently?
Product development: Do different segments need different products, features, or service levels?
Messaging: Should we tailor our communications to different audience groups? What resonates with each?
Portfolio strategy: Which segments are we winning with today, and where are the growth opportunities?
Market sizing: How large is each segment, and what is the revenue potential of each?

It is particularly valuable when:

The business suspects “one size fits all” is not working
There is a rich dataset (survey or behavioral) with enough variables to reveal genuine differences
Stakeholders need a shared language for talking about customer groups
The segmentation will inform strategy across multiple functions (marketing, product, sales)

When NOT to Use

Segmentation may not be a good fit when:

The audience is genuinely homogeneous: If a category has very little variation in needs or behaviors, forced segmentation will produce artificial groups. Run an initial exploration first.
The sample is too small: Robust segmentation typically needs 300+ respondents at minimum, and more if you expect many segments. Small samples produce unstable solutions.
You only need a ranking or prioritization: If the question is “what matters most?” rather than “who are the different groups?”, use MaxDiff or Key Drivers Analysis instead.
The data is purely demographic: Segmenting by age/gender/income alone rarely produces actionable groups. Effective segmentation usually requires attitudinal, behavioral, or needs-based inputs.
There is no plan to act on the segments: Segmentation is a strategic investment. If there is no clear use case for the segments (targeting, product development, communications), the effort may not justify the cost.

What the Consultant Needs to Define

Before analysis begins, the analytics team needs clear answers to four questions. These shape the entire approach:

1. Objective — What will the segments be used for?

Objective	What It Means	Typing Tool Needed?
Audience understanding	“We want to know who’s in our audience and how they differ” — used for internal insight and reporting	Usually no
Strategic segmentation	“We want a segmentation we can use across the business for targeting, product dev, and comms” — long-term use	Yes
Typing tool for tracking	“We want to classify new respondents into these segments in future surveys”	Yes
Behavioral grouping	“We want to understand different usage or behavior patterns within a known audience”	Usually no

2. Inputs — What variables should drive the segmentation?

Inputs are the variables used to “pull consumers apart” — they define what makes the segments different. Common input types:

Attitudinal: Attitudes, beliefs, values (e.g., “I prefer quality over quantity”)
Behavioral: Frequency of use, purchase patterns, engagement level
Motivational: Reasons for using a product or category
Needs-based: Stated needs, unmet needs, pain points
Mixed: Combinations of the above (often produces the strongest solutions)

Tip

The analytics team will often test multiple combinations of inputs and present shortlisted solutions. In the WATG project, three solutions were tested — attitudes + motivations, attitudes + frequency, and motivations + spend — before recommending the strongest one.

3. Profiling variables — What do you want to see differences across?

Profiling variables are not used to create the segments — they are used to describe them after clustering. These are the variables where you hope to see meaningful differences between segments:

Demographics (age, gender, income, life stage)
Brand perceptions and preferences
Media consumption and channel usage
Category behaviors (frequency, spend, repertoire)
Lifestyle attitudes and hobbies
Product needs and pain points

4. Number of segments — How many groups are useful?

There is a trade-off between granularity and usability:

3–4 segments: Easier to communicate and act on, but may miss important nuances
5–6 segments: A common sweet spot for strategic segmentations
7+ segments: More granular but harder for the business to operationalize

The analytics team will test a range and recommend the number that balances statistical quality with business utility.

Time Allocation (For a strategic segmentation)

Stage	Hours
Kick-off & planning	5
Analysis - overall	35
Total	40 hours

Note: Strategic segmentations with typing tools typically require more hours. Simple clustering may require fewer. Hours are indicative and scoped per project.

Key Milestones (Analytics Perspective)

Analytics Briefing: Initial briefing to understand segmentation objectives, input candidates, profiling requirements, and desired number of segments
Data Review & Preparation: Data cleaning, variable recoding, and exploratory analysis to assess suitability of candidate inputs
Solution Testing: Multiple clustering solutions tested with different input combinations and segment counts
Alignment Session: Shortlisted solutions presented to the project team. Key decision: which solution is most useful for the business?
Full Profiling: Recommended solution fully profiled across all profiling variables
Final Outputs: Segment narratives, profiling tables, and visualization delivered
Typing Tool (if required): Classification model exported for use in future surveys
Debrief Attended by Analytics (if necessary): Analytics team available for findings presentation and Q&A session

Questionnaire Considerations

Segmentation does not have its own question format like MaxDiff or Van Westendorp. Instead, it draws on variables already collected in the survey. However, the questionnaire design directly affects segmentation quality:

For segmentation inputs:

✅ Include attitudinal batteries (agree/disagree scales) covering the attitudes and motivations most relevant to the category
✅ Include behavioral questions — frequency, spend, usage patterns
✅ Use consistent scales across input batteries (e.g., all 5-point or all 7-point) to avoid mixing signal strength
✅ Ensure enough input variables — typically 10–30 candidate inputs to test different solutions
✅ Screen out straight-liners — respondents who give the same answer across all scale questions produce noise

For profiling variables:

✅ Include rich demographics — age, gender, income, household composition, life stage
✅ Include brand and category questions — awareness, consideration, usage, perceptions
✅ Include media and channel questions — if targeting implications are important
✅ Include needs and pain points — to understand what each segment is looking for

Sample size:

Minimum 300 respondents for simple clustering
500–1,000+ preferred for strategic segmentation with typing tool
Each segment needs enough respondents for stable profiling — if you expect 5 segments, plan for at least 100 per segment

Segment Overview

The primary output is a clear summary of each segment — who they are, how they differ, and what matters to them. This typically includes:

Segment name and size — a descriptive label and the proportion of the audience in each segment
Key defining characteristics — the 3–5 attributes that most distinguish this segment from others
Behavioral patterns — how this segment behaves differently (usage, frequency, spend)
Needs and motivations — what drives this segment and what they’re looking for
Strategic implications — how the business should engage with this segment differently

Segmentation Assessment + AI Interpretation

TVE provides an interactive segmentation dashboard that gives the project team direct access to explore the data, evaluate solutions, and build understanding of each segment. The dashboard includes:

Data exploration:

Variable explorer — browse and visualize all variables in the survey, with filtering and sorting
Correlation matrix — examine relationships between variables to understand how inputs and profiling variables relate to each other
Cross-tabulation tables — traditional profiling tables showing how each segment differs across all profiling variables, with statistically significant differences highlighted

Segmentation solutions:

All tested solutions — every segmentation solution built during analysis is available for review, not just the recommended one
Summary variable distributions — see how the clustering inputs and other key variables are distributed within each segment, making it clear what drives each solution and how segments are different from each other
Segment migration table — understand how respondents move between segments when the number of clusters changes (e.g., moving from a 5-segment to a 4-segment solution)

AI interpretation:

Automated segment narratives — AI-generated descriptions of each segment based on the profiling data, giving the project team an immediate starting point for segment storytelling
Key differentiators — AI-highlighted variables where segments differ most, helping consultants quickly identify the most actionable distinctions

Additional Output Types

Beyond the dashboard, segmentation analysis may also include:

Segment size distribution: Proportion of the audience in each segment, with breakdowns by market or demographic
Variable importance: Which input variables contribute most to distinguishing the segments
Typing tool (if required): A classification model that assigns new respondents to segments based on a short set of “golden questions”

Previous Project Examples

Project 1: Sony PlayStation — Ghost of Yotei Post-Launch (Simple Clustering)

Clustered players based on play style — story-focused, exploration-focused, completionists
No typing tool needed — used for post-launch reporting only
Project Folder

Project 2: Wool and the Gang — Crafting a Segmentation (Strategic Segmentation)

Tested three solutions with different input combinations (attitudes, motivations, frequency, spend)
Recommended solution combined behavioral and attitudinal inputs
Full profiling across brand interest, platform usage, lifestyle, and buying attitudes
Typing tool developed for ongoing classification
Project Folder

Project 3: Edenred — EV Layer Segmentation (Needs-Based Grouping)

Mapped fleet decision-makers by EV adoption status and charging needs
Demographic and firmographic groups consolidated based on observed need patterns
No statistical modeling required — groups defined by existing business logic
Project Folder

R Package Documentation

tveSegmentation package documentation (internal)
– TVE’s proprietary R package for segmentation analysis, including K-means, K-prototypes, latent class, and archetypes clustering, plus profiling and typing tool export

Ready to use segmentation in your project? Contact the analytics team to discuss your requirements and next steps.

Email: Analytics@dtadvisorygroup.com

What to prepare for our discussion:

The business objective — what will the segments be used for? (insight, strategy, targeting, tracking)
Candidate input variables — what attitudes, behaviors, or needs should drive the segmentation?
Desired profiling variables — what do you want to see differences across?
Preferred number of segments (or a range to test)
Whether a typing tool is needed for classifying future respondents
Target audience definition and expected sample size
Decision timeline and budget parameters

--- title: "Segmentation" author: "TVE Analytics Team" date: last-modified categories: [segmentation, customer-understanding] description: "Segmentation divides an audience into distinct groups based on shared behaviors, attitudes, or needs — enabling targeted strategy, messaging, and product development" --- :::: panel-tabset ## What It Is ------------------------------------------------------------------------ ### Overview Segmentation divides a broad audience into meaningful, distinct groups so that each group can be understood and targeted differently. The goal is to move beyond demographics and uncover groups that genuinely differ in what they need, how they behave, or what motivates them. TVE uses a range of analytical approaches depending on the project objectives — from simple clustering on survey responses to full strategic segmentations with typing tools for ongoing classification. The analytics team selects and configures the right model based on the data, the business question, and how the segments will be used. **What the analytics team handles:** - Selecting the appropriate clustering method (K-means, K-prototypes, latent class analysis, archetypes, decision trees, and more) - Testing multiple solutions with different input combinations - Evaluating and recommending the strongest solution based on statistical and strategic criteria - Building typing tools for ongoing segment classification (when required) **What consultants need to provide** is covered below — the inputs, objectives, and desired outputs that shape the analysis. ------------------------------------------------------------------------ ### Types of Segmentation Not every segmentation requires the same level of complexity. The approach depends on what the segments will be used for: | Type | Description | Example | |------------------|-------------------------------|-----------------------| | **Simple clustering** | Groups respondents based on a small set of survey questions to understand behavioral patterns. No typing tool needed. | Clustering gamers based on how they approach playing a game (story-focused vs. exploration-focused vs. completionists) | | **Strategic segmentation** | A comprehensive segmentation designed for long-term use across the business. Includes multiple tested solutions, full profiling, and a typing tool for classifying new respondents. | Segmenting crafters based on motivations, attitudes, and engagement to inform brand strategy, product development, and communications | | **Needs-based grouping** | Maps existing demographic or firmographic groups against their needs and behaviors. May not require statistical modeling — instead consolidates known groups based on observed need patterns. | Grouping fleet decision-makers by their EV adoption status and charging needs to identify which segments have the strongest product-market fit | ------------------------------------------------------------------------ ### What Makes a Strong Segmentation A strong segmentation is not just statistically sound — it must be usable by the business. Key criteria: - **Viable segment sizes** — not so small they can't be found, not so large they lack clear identity - **Clearly differentiated** — homogeneous within each segment, distinct from all other segments - **Not driven by demographics alone** — to avoid segments that simply reflect life stage or geography - **Stable over time** — built on inputs that don't shift rapidly, so segments remain useful for strategic planning - **Easily explainable** — segments should feel intuitive and be easy to communicate internally - **Actionable** — clear hooks within each segment that the business can act on ------------------------------------------------------------------------ ## When to Use ### Business Questions This Answers Segmentation is ideal when clients ask: **"Are all our customers the same, or are there meaningfully different groups?"** It helps answer strategic questions such as: - **Audience understanding:** Who are our customers, and how do they differ in needs, behaviors, or motivations? - **Targeting:** Which customer groups should we prioritize, and how should we reach them differently? - **Product development:** Do different segments need different products, features, or service levels? - **Messaging:** Should we tailor our communications to different audience groups? What resonates with each? - **Portfolio strategy:** Which segments are we winning with today, and where are the growth opportunities? - **Market sizing:** How large is each segment, and what is the revenue potential of each? It is particularly valuable when: - The business suspects "one size fits all" is not working - There is a rich dataset (survey or behavioral) with enough variables to reveal genuine differences - Stakeholders need a shared language for talking about customer groups - The segmentation will inform strategy across multiple functions (marketing, product, sales) ------------------------------------------------------------------------ ### When NOT to Use Segmentation may not be a good fit when: - **The audience is genuinely homogeneous:** If a category has very little variation in needs or behaviors, forced segmentation will produce artificial groups. Run an initial exploration first. - **The sample is too small:** Robust segmentation typically needs **300+ respondents** at minimum, and more if you expect many segments. Small samples produce unstable solutions. - **You only need a ranking or prioritization:** If the question is "what matters most?" rather than "who are the different groups?", use **MaxDiff** or **Key Drivers Analysis** instead. - **The data is purely demographic:** Segmenting by age/gender/income alone rarely produces actionable groups. Effective segmentation usually requires attitudinal, behavioral, or needs-based inputs. - **There is no plan to act on the segments:** Segmentation is a strategic investment. If there is no clear use case for the segments (targeting, product development, communications), the effort may not justify the cost. ------------------------------------------------------------------------ ## Project Management Info ### What the Consultant Needs to Define Before analysis begins, the analytics team needs clear answers to four questions. These shape the entire approach: **1. Objective — What will the segments be used for?** | Objective | What It Means | Typing Tool Needed? | |-------------------|-----------------------|-------------------------------| | **Audience understanding** | "We want to know who's in our audience and how they differ" — used for internal insight and reporting | Usually no | | **Strategic segmentation** | "We want a segmentation we can use across the business for targeting, product dev, and comms" — long-term use | Yes | | **Typing tool for tracking** | "We want to classify new respondents into these segments in future surveys" | Yes | | **Behavioral grouping** | "We want to understand different usage or behavior patterns within a known audience" | Usually no | **2. Inputs — What variables should drive the segmentation?** Inputs are the variables used to "pull consumers apart" — they define what makes the segments different. Common input types: - **Attitudinal:** Attitudes, beliefs, values (e.g., "I prefer quality over quantity") - **Behavioral:** Frequency of use, purchase patterns, engagement level - **Motivational:** Reasons for using a product or category - **Needs-based:** Stated needs, unmet needs, pain points - **Mixed:** Combinations of the above (often produces the strongest solutions) ::: callout-tip The analytics team will often test multiple combinations of inputs and present shortlisted solutions. In the WATG project, three solutions were tested — attitudes + motivations, attitudes + frequency, and motivations + spend — before recommending the strongest one. ::: **3. Profiling variables — What do you want to see differences across?** Profiling variables are **not** used to create the segments — they are used to describe them after clustering. These are the variables where you hope to see meaningful differences between segments: - Demographics (age, gender, income, life stage) - Brand perceptions and preferences - Media consumption and channel usage - Category behaviors (frequency, spend, repertoire) - Lifestyle attitudes and hobbies - Product needs and pain points **4. Number of segments — How many groups are useful?** There is a trade-off between granularity and usability: - **3–4 segments:** Easier to communicate and act on, but may miss important nuances - **5–6 segments:** A common sweet spot for strategic segmentations - **7+ segments:** More granular but harder for the business to operationalize The analytics team will test a range and recommend the number that balances statistical quality with business utility. ------------------------------------------------------------------------ ### Time Allocation (For a strategic segmentation) | Stage | Hours | |---------------------|--------------| | Kick-off & planning | 5 | | Analysis - overall | 35 | | **Total** | **40 hours** | **Note:** Strategic segmentations with typing tools typically require more hours. Simple clustering may require fewer. Hours are indicative and scoped per project. ------------------------------------------------------------------------ ### Key Milestones (Analytics Perspective) 1. **Analytics Briefing:** Initial briefing to understand segmentation objectives, input candidates, profiling requirements, and desired number of segments 2. **Data Review & Preparation:** Data cleaning, variable recoding, and exploratory analysis to assess suitability of candidate inputs 3. **Solution Testing:** Multiple clustering solutions tested with different input combinations and segment counts 4. **Alignment Session:** Shortlisted solutions presented to the project team. Key decision: which solution is most useful for the business? 5. **Full Profiling:** Recommended solution fully profiled across all profiling variables 6. **Final Outputs:** Segment narratives, profiling tables, and visualization delivered 7. **Typing Tool (if required):** Classification model exported for use in future surveys 8. **Debrief Attended by Analytics (if necessary):** Analytics team available for findings presentation and Q&A session ------------------------------------------------------------------------ ## Data Collection Guide ### Questionnaire Considerations Segmentation does not have its own question format like MaxDiff or Van Westendorp. Instead, it draws on variables already collected in the survey. However, the questionnaire design directly affects segmentation quality: **For segmentation inputs:** - ✅ Include **attitudinal batteries** (agree/disagree scales) covering the attitudes and motivations most relevant to the category - ✅ Include **behavioral questions** — frequency, spend, usage patterns - ✅ Use **consistent scales** across input batteries (e.g., all 5-point or all 7-point) to avoid mixing signal strength - ✅ Ensure **enough input variables** — typically 10–30 candidate inputs to test different solutions - ✅ Screen out **straight-liners** — respondents who give the same answer across all scale questions produce noise **For profiling variables:** - ✅ Include rich **demographics** — age, gender, income, household composition, life stage - ✅ Include **brand and category** questions — awareness, consideration, usage, perceptions - ✅ Include **media and channel** questions — if targeting implications are important - ✅ Include **needs and pain points** — to understand what each segment is looking for **Sample size:** - Minimum **300 respondents** for simple clustering - **500–1,000+** preferred for strategic segmentation with typing tool - Each segment needs enough respondents for stable profiling — if you expect 5 segments, plan for at least 100 per segment ------------------------------------------------------------------------ ## Output Examples ### Segment Overview The primary output is a clear summary of each segment — who they are, how they differ, and what matters to them. This typically includes: - **Segment name and size** — a descriptive label and the proportion of the audience in each segment - **Key defining characteristics** — the 3–5 attributes that most distinguish this segment from others - **Behavioral patterns** — how this segment behaves differently (usage, frequency, spend) - **Needs and motivations** — what drives this segment and what they're looking for - **Strategic implications** — how the business should engage with this segment differently ------------------------------------------------------------------------ ### Segmentation Assessment + AI Interpretation TVE provides an interactive segmentation dashboard that gives the project team direct access to explore the data, evaluate solutions, and build understanding of each segment. The dashboard includes: **Data exploration:** - **Variable explorer** — browse and visualize all variables in the survey, with filtering and sorting - **Correlation matrix** — examine relationships between variables to understand how inputs and profiling variables relate to each other - **Cross-tabulation tables** — traditional profiling tables showing how each segment differs across all profiling variables, with statistically significant differences highlighted **Segmentation solutions:** - **All tested solutions** — every segmentation solution built during analysis is available for review, not just the recommended one - **Summary variable distributions** — see how the clustering inputs and other key variables are distributed within each segment, making it clear what drives each solution and how segments are different from each other - **Segment migration table** — understand how respondents move between segments when the number of clusters changes (e.g., moving from a 5-segment to a 4-segment solution) **AI interpretation:** - **Automated segment narratives** — AI-generated descriptions of each segment based on the profiling data, giving the project team an immediate starting point for segment storytelling - **Key differentiators** — AI-highlighted variables where segments differ most, helping consultants quickly identify the most actionable distinctions ------------------------------------------------------------------------ ### Additional Output Types Beyond the dashboard, segmentation analysis may also include: - **Segment size distribution:** Proportion of the audience in each segment, with breakdowns by market or demographic - **Variable importance:** Which input variables contribute most to distinguishing the segments - **Typing tool (if required):** A classification model that assigns new respondents to segments based on a short set of "golden questions" ------------------------------------------------------------------------ ## Resources ### Previous Project Examples **Project 1: Sony PlayStation — Ghost of Yotei Post-Launch (Simple Clustering)** - Clustered players based on play style — story-focused, exploration-focused, completionists - No typing tool needed — used for post-launch reporting only - [Project Folder]('Z:\Shared\TVE%20Data\1.%20Client%20Projects\Sony\PlayStation\120.%2015029%20-%20Ghost%20of%20Yotei%20Post%20Launch') **Project 2: Wool and the Gang — Crafting a Segmentation (Strategic Segmentation)** - Tested three solutions with different input combinations (attitudes, motivations, frequency, spend) - Recommended solution combined behavioral and attitudinal inputs - Full profiling across brand interest, platform usage, lifestyle, and buying attitudes - Typing tool developed for ongoing classification - [Project Folder]('Z:\Shared\TVE%20Data\1.%20Client%20Projects\Wool%20and%20the%20Gang\4245%20-%20Crafting%20a%20Segmentation%20-%20December%202024') **Project 3: Edenred — EV Layer Segmentation (Needs-Based Grouping)** - Mapped fleet decision-makers by EV adoption status and charging needs - Demographic and firmographic groups consolidated based on observed need patterns - No statistical modeling required — groups defined by existing business logic - [Project Folder]('Z:\Shared\TVE%20Data\1.%20Client%20Projects\Edenred\4645%20-%20EV%20Layer%20-%20September%202025') ------------------------------------------------------------------------ ### R Package Documentation - **tveSegmentation** package documentation (internal)\ -- TVE's proprietary R package for segmentation analysis, including K-means, K-prototypes, latent class, and archetypes clustering, plus profiling and typing tool export ------------------------------------------------------------------------ ## Get Started Ready to use segmentation in your project? [Contact the analytics team](mailto:Analytics@dtadvisorygroup.com) to discuss your requirements and next steps. **Email:** Analytics\@dtadvisorygroup.com **What to prepare for our discussion:** - The business objective — what will the segments be used for? (insight, strategy, targeting, tracking) - Candidate input variables — what attitudes, behaviors, or needs should drive the segmentation? - Desired profiling variables — what do you want to see differences across? - Preferred number of segments (or a range to test) - Whether a typing tool is needed for classifying future respondents - Target audience definition and expected sample size - Decision timeline and budget parameters ::::