We develop the best solutions by balancing statistical validity and real-world usability. The benefit of our process is that we take nothing for granted. We use as many analytical methods as necessary and will work with you to determine key profile variables, evaluate solutions, and select the best one together.

CMB uses a variety of techniques to produce multiple potential segmentation schemes and test a variety of potential basis variables. We commonly leverage the following:

• K-means, hierarchical, and ensemble clustering
• Factor Analysis to reduce the number of attributes for segmenting.
• Latent Class Modeling to segment through robust models that include valuable diagnostic statistics.
• Detailed Profile Sheets to flesh out rich quantitative segment personas.
• Predictive Modeling (TreeNet) to understand key drivers: technology, habits, content, and more, and to develop classification algorithms (e.g., typing tools).
• Habit loops to reveal consumers’ relevant psychological drivers: we’ll segment occasions based on social psychology and consumer behavior and loyalty. This approach delivers insights into ways to trigger use–or reduce barriers and make the routine easier.

We help our clients bring new products to market, tweak existing products, and design and re-design product lines. We also specialize in product and line optimization–maximizing reach and revenue. That optimization can involve pricing, product features, and information about product performance vs. the competitive set. We commonly employ Discrete Choice and Conjoint modeling as well as MaxDiff and TURF analysis.

Driver modeling helps us uncover the “why” behind the metrics. We use Driver modeling techniques like Gradient Boosted Modeling, Shapley Value Regression, Multivariate Adaptive Regression Splines, and CHAID/C&RT to provide insight into increasing NPS scores, identifying which customer issues are most important to solve, and which attributes are most important to consumers.

We combine neural networks and AI technology with human intelligence (HI) to analyze large amounts of data at scale, including open-ended responses. This integrated approach leverages AI to overcome biases and expectations inherent to human interpretation in order to identify new and unanticipated themes, while HI enables fine-tuning of themes based on knowledge of what will be most useful to stakeholders.

If you’re focused solely on what your consumers think of you, you’re missing a critical piece of the puzzle. Our revolutionary consumer-centric BrandFx℠ solution delivers a complete picture of your brand’s performance, based on the four key psychological benefits that drive consideration, trial, loyalty, and advocacy:

• EMOTIONAL BENEFITS: How I want to feel (e.g., positive feelings, enhanced joy, reduced frustration)
• IDENTITY BENEFITS: Who I want to be (e.g., strong self-esteem, pride, a positive self-image, relating to a tribe)
• SOCIAL BENEFITS: Where I want to belong (e.g., sense of community, conversation, social connection)
• FUNCTIONAL BENEFITS: What I want to do (e.g., ability to accomplish tasks or goals, saving time or money)