Industry/Market Focus:

Advertising Agencies

Content Type:

Glossary

Kelly Triads Definition

A qualitative market research technique used to uncover the subconscious criteria and native language consumers use to differentiate competing brands, products, or advertisements. The technique helps researchers and ad agencies map consumer perceptions without imposing pre-set survey options.

Originating from George Kelly’s Personal Construct Theory (1955), this method reveals how consumers naturally categorize products. By avoiding pre-scripted questions, Kelly's Triads capture authentic consumer language and uncover the real drivers behind purchase decisions.

How the Kelly triad method works

• Selection: The researcher selects a set of competitive elements (e.g., six competing brands of toothpaste).
• Triadic Comparison: Three items are dealt or displayed simultaneously.
• The "Odd One Out" Prompt: The respondent is asked to pick the item that differs from the other two and explain the distinction—for example: "Which two are most alike, and how is the third one different?"
• Construct Capture: The moderator notes the specific language, descriptors, and contrast poles the consumer uses (e.g., "These two feel like everyday essentials, but this one feels like an indulgence").
• Repetition: The process repeats with different combinations of three until the respondent exhausts new ways of categorizing the elements.

What are some applications of Kelly triads

• Ad Copywriting: Captures authentic consumer vocabulary to craft resonant marketing messaging.
• Repertory Grids: Serves as the foundation for Repertory Grid Analysis, where products are scored against the newly discovered attributes.
• Perceptual Mapping: Identifies the true, underlying dimensions driving brand positioning and consumer choices.
• Survey Design: Informs the attribute lists and rating scales used in large-scale quantitative studies.

What are best practices when using Kelly triads

To get clean, actionable results from the Kelly’s Triads (triadic sorting) method, follow these core best practices during design, moderation, and synthesis:

1. Element Selection & Preparation

• Ensure Familiarity: Choose items (brands, packaging, ads, or concepts) that participants know well or can easily evaluate on the spot.
• Keep the Set Manageable: Aim for 6 to 10 total elements. Fewer than 6 doesn't yield enough unique combinations; more than 10 leads to respondent fatigue.
• Maintain Category Homogeneity: All elements must belong to the same competitive frame (e.g., compare premium coffee brands with other premium coffee brands, not coffee brands with luxury cars).
• Use Physical or Visual Cards: Use tangible cards or high-resolution digital boards so respondents can physically move and group the three items.

2. Facilitation & Moderation

• Frame the Question Clearly: Stick to the core prompt: "In what important way are two of these similar to each other, but different from the third?"   
• Enforce Both Ends of the Construct: Don't just ask why two items are alike—always capture the opposite pole (e.g., if two are "approachable," ask what the third one is: "intimidating," "exclusive," or "corporate"?).
• Capture Native Vocabulary verbatim: Write down the exact words, slang, and phrases respondents use. Avoid "cleaning up" or translating their language into internal marketing jargon.
• Probe Deeper (Laddering): Once a basic functional difference is named (e.g., "This one has cleaner packaging"), ladder up to emotional drivers: "What does cleaner packaging imply about the brand?" or "How does that make you feel when buying it?"
• Watch for "Boring" Discriminators: If a respondent gives an obvious physical distinction (e.g., "These two are red and that one is blue"), acknowledge it and ask: "Aside from color, what is a functional or perceptual difference?"

3. Execution & Synthesis

• Mix Up the Triads: Rotate element combinations so every item is paired with different competitors across multiple rounds.
• Stop at Saturation: End the sorting exercise when the participant starts repeating constructs or struggles to find new ways to differentiate the items.
• Feed into Quantitative Analysis: Use the extracted "constructs" (bipolar scale labels) directly as the attribute list for subsequent Repertory Grids, perceptual mapping, or quantitative rating surveys.