Behavioural Analytics

Behavioural Analytics

To truly inspire insights through market research, we must first create a model that deciphers current and past behaviours to predict future outcomes

PMG Intelligence started as a full-service market research organization that developed a core competency in asking the right questions and finding ways to analyze and present results that answered our clients’ informational objectives.  In the 90s we grew our data collection and focus group facilities, entered the realm of online survey technology, and added multivariate data analysis techniques to our process.  Our analysis methods were really starting to evolve and the insights we were producing deepened.  However, there was still a gap in understanding the relative impact and meaning in certain areas in our findings.

Researchers have different philosophies and processes for how they design and interpret research data. From which scales to use (Likert, 7-point, and 10-point are a few popular choices) to how the data is analyzed (for example, top box vs. top 2 box), the possibilities are endless. Standard deviation thresholds vary as well as the sequence of advanced statistical processes. Factor analysis, PLS modelling, regression, clustering, ANOVA, t-tests, and correlations can be applied in various combinations to help answer simple to complex research questions.

Raising the Bar in Analysis

We sought to innovate new approaches to how we interpret data that go beyond conventional methodologies. To increase our data analysis capabilities and maximize the insights produced, we developed an analysis methodology built on benchmarks established over millions of interviews across various topic areas and research samples. This model helped us establish algorithms that we use to more effectively test hypotheses and uncover patterns in the data that we may not have otherwise observed.

“We wanted to identify the specific triggers that improve outcomes between company and client and establish new insights into consumer decision-making and behaviour change.”

Our Behavioural Analytics framework has changed how we think about data and was foundational in the development of our Behavioural Segmentation (P3), Predictive Modelling, and our Path to Engagement model.


Segmentation is an overused word in research and can simply mean a crosstab that compares data by geography, gender, or any other demographic variables. In our view, segmentation is the process of identifying unique, homogenous groups that are connected by specific behavioural and attitudinal markers.   Through a specific multivariate methodology, we identify how many different pathways exist in the data by examining response patterns.  Like identifying the number of highways that exist on a map based and how someone gets from A to B,  a company may have several different types of customers; our process discerns the number of customer types and then overlays models to determine how to most effectively influence them.  The outcomes are then tested against benchmark references and qualitative research. Today, we are using this framework to create customized segmentation models for our clients across many industry sectors and have also developed industry-specific segmentation applications in financial services, risk management, health science, and technology.

How Was the Model First Established?

The process was initiated by first developing a consumer Psychological Personality Profiling model (P3).  We essentially created our own Myers-Briggs analysis.  We determined there are unique segments of Canadian consumers and business professionals reflecting their personality, risk tolerance, behavioural, and attitudinal characteristics.  Where applicable, we apply this model to our research design and data analysis processes so that we can leverage more insights and present recommendations that answer the important questions such as: “How do we increase satisfaction?”, “How do we improve retention?”, or “What tools or messages have the greatest impact on our market and segments?”.

Take The Test

Predictive Modelling

We don’t work in “I think”; we work in probability and certainty. In our view, our responsibility as a consultancy is to help our clients make informed decisions – to take risks when the probability of success is high and to make the best decisions that are financially sound. With this, predictive modelling is our process of using historical data to help anticipate the future behaviours of consumers and businesses. Our analytical and segmentation models help interpret data and maximize insights and those insights are then applied against our data stores to develop recommendations. More specifically, we have tested segmentation profiling and evaluated the attitudinal and behavioural impact of various messaging and tactical marketing strategies through pre- and post-quantitative measurements with tens of thousands of consumer households. In application, we have helped our clients maximize the impact of their marketing and operational spend by supporting strategy development, messaging design, product development, and digital and usability evaluation.

Path to Engagement Framework

PMG’s Path to Engagement framework plays a critical role in how we decipher research data.  It is built on the premise that people, consumers of products and services, and business professionals work through a specific process when making decisions.  Whether it is to switch from one company to another, adopt a new technology, change lifestyle, or manage money, human decision-making travels through a routinized process.  We may spend a different amount of time at each step of the process, but all decisions travel through the Path to Engagement.

This framework helps identify where the consumer or customer resides in the process in terms of their relationship or connectivity to a brand, technology and product/service adoption, their predisposition to recommend, and, importantly, changing their way of doing things.  By leveraging our segmentation models,  P3, benchmarking, and qualitative data we provide key insights in how to move consumers through messaging, marketing, and digital media.

Advanced NPS Modeling

The Net Promoter Score is one of the most effective mechanisms available to understand the retention and promotion qualities of a company’s brand.

To enhance the level of insight in terms of customer experience and what specific variables have the greatest influence on increasing NPS and Gamma outcomes, PMG has developed the Customer Experience Composite Analysis (CECA).  CECA connects four (4) key variables, including propensity to refer, satisfaction, confidence, and switch/adoption rates/propensity to consider counter-offers.