Standard Frameworks for Finding Insights in Data

The following frameworks often prove to help in uncovering insights from data:

• The Delta Principle
• Contrasting attitude with behavior
• Conversion through stages of development
• Driver analysis and purchase hierarchies
• Punctuation versus small and gradual processes

The Delta Principle

One of the ugly truths of survey research is that all survey results contain some error, and the level of error is often both unknown and unknowable. A fix for this is to focus on comparing relativities rather than the absolute magnitude of results. This is done by focusing on:

• Comparisons by sub-group.
• Comparisons against benchmarks.
• Comparisons over time.
• Comparisons to very similar questions.

For more information, see The Delta Principle of Data Analysis.

Contrasting attitude with behavior

Many insights in real-world insights come by contrasting attitude with behavior. Do people do what they say? Why not?

For example, see Contrasting Attitude with Behavior (Basic Brand Vulnerability Matrix)

Conversion through stages of development

The comparison of attitude to behavior can be thought of as a special case of a more general approach to assessing how people convert through different stages of development. For example:

• Think →  Feel → Do
• Aware → Know → Like → Prefer → Try → Loyal

For an example, see Comparing Multiple Conversion Metrics.

Segmentation

It is commonly the case that consumers in a market vary in their needs and wants. Segmenting consumers into similar groups in terms of their needs and wants can be a useful way of describing markets. Common ways of doing this include:

• Age
• Gender
• Family life stage
• Price sensitivity
• Occasion
• Needs/Demand-creating conditions
• Satisfaction/loyalty

See also Overview of Market Segmentation

Driver analysis and purchase hierarchies

Another framework is to attempt to quantify the relative importance of different factors in driving behavior. In some areas, this is known as identifying purchase hierarchies.

For more information, see Introduction to Driver Analysis.

Punctuation versus small and gradual processes

When understanding the physical world, some of the most powerful frameworks involve showing that major changes reflect the cumulative impact of small and gradual processes over time. The most famous examples of this are the concepts of erosion and plate movement in geology, Darwin's theory of evolution, the various economic theories of history, and the ideas of the viral transmission of diseases and information (e.g., word of mouth effects). The central idea is that the magnitude of differences reflects differences in time. For example, the mountains in North America are much higher than those in Australia, because the ones in Australia are much older and thus more eroded, and Facebook is big because it has been consistently growing over a long time.

The alternative to the world being shaped by graduate processes is that it is shaped by major events or people. Examples of this include:

• The theory of punctuated equilibrium in evolution.
• The bang of the big bang theory.
• The effect of product tampering in consumer markets, whereby even after the product tampering is addressed, the products that were tampered with never recover (to use some jargon, this is rate-independent hysteresis).