Advanced Analytics

Cluster Analysis is a class of statistical techniques that can be applied to data that exhibit “natural” groupings. Cluster analysis sorts through the raw data and groups them into clusters. A cluster is a group of relatively homogeneous cases or observations. Objects in a cluster are similar to each other and dissimilar to objects in other clusters.
Multivariate Linear Regression analysis helps to understand how the value of the dependent variable changes when any one of the independent variables is varied, while the other independent variables are held fixed – for example, one could build a predictive model that helps understand how the typical weekly consumption of chocolate varies as a result of such factors as a consumer’s age, income, gender, habits and preferences.
Discriminant Analysis is a statistical method that is used by researchers to help them understand the relationship between a dependent variable and one or more independent variables – for example, one could use discriminant analysis to “quantify” the relationship between age, gender, income, etc. (independent variables) and whether a person would prefer dark, semi-sweet, or milk chocolate (categorical dependent variable).
Factor Analysis is a data reduction technique, where the goal is to summarize the information in a larger number of correlated variables into a smaller number of factors that are not correlated with each other – for example, in profiling segments of dark, semi-sweet and milk chocolate consumers one could use a multitude of consumer habits and preferences, where many might be correlated beyond variables such as age, gender and income.

Factor analysis can help summarize the information captured by this large number of correlated habits and preferences into a more manageable number of factors, which can then be used as inputs to cluster and discriminant analysis.

Structural Equation Modeling (SEM) is a powerful multivariate statistical technique which combines factor analysis and multivariate regression to study complex relationships among variables, where some of them can be hypothetical or unobservable.
Analysis of Variance (ANOVA) tests for significant differences between means, or between 2 groups. This technique is often used to test the hypothesis that the means among two or more groups are equal – for example, whether males and females consume different amount of chocolate on a weekly basis.
Financial Analysis allows companies to proactively seek out ways to change and enhance their business models so that they are constantly up-to-date with the current financial environment.
Financial analytics modeling requires the integration and assessment of a broad spectrum of data that may potentially affect a company.

Some of the data the analytics may take into account include; which customers provide the company with the most profit; how the company’s customer-base is spread out geographically, and which product brings in the most profit. Once the data has been collected, it can be displayed in charts or graphs so that complex and diverse information can be easily visualized.

Predictive Analytics and data mining both apply sophisticated mathematics to data in order to solve difficult business questions and help inform decision makers to make the right decision at the right time. At b3Intelligence, we do both. Data mining refers to an analytic toolset that automatically searches for useful patterns in large data sets. Tools that can extract relevant pieces of your data format it and then display it in a comprehendible method to take it to the next level.

Predictive analytics is an analyst-guided (not automatic) discipline that uses data patterns to make forward-looking predictions, or to make complex statements about customers by evaluating multiple data patterns. Data mining searches for clues and predictive analytics delivers answers that guide you to a “what next” action. Data mining is often the initial stage in developing a predictive model.

Asset Intelligence is used to gauge the company’s asset health. When assets are not working due to an operational failure or as the result of external or environmental factors, they impact revenue and service availability, and expose lives and property to undue risk. b3Intelligence has created a solution to this issue with their predictive analytics modeling

This solution allows our clients to predict the health of their assets using certain variables already collected though their inventory process. We combine notions of age, location, make, model and other information to accurately predict the heath of certain assets and can integrate with our clients work ordering system to give them a proactive approach to asset maintenance.

The cost of fixing an asset when it has already failed is significantly higher than the cost of proactively taking action to avoid that failure. b3Intelligence’s Asset Intelligence solution is designed to facilitate that proactivity. Using general linear modeling and predictive analytics we are able to help operations personnel diagnose problems, and enable rapid remedial action to issues that occur.

Market Segmentation uses cluster analysis in conjunction with discriminant analysis to reveal and score hidden segments for a telecom customer.

We used transactional, service profiles and demographic (anonymous) data from the client’s database to uncover segments, which varied in terms of these variables via cluster analysis. An algorithm was developed to “score” the respondents by using discriminant analysis.

The client has successfully used the algorithm to not only track how their segment mix is changing over time but also modifying their service offerings to appeal to these segments for maximal impact.

Chi-square Analysis is used to assess two types of comparison: tests of goodness of fit and tests of independence. A test of goodness of fit establishes whether or not an observed frequency distribution differs from a theoretical distribution. A test of independence establishes whether paired observations on two variables are independent of each other – for example, whether males and females differ in the frequency with which they consume chocolate.
Logistic Regression is a type of predictive model that can be used when the target variable is a categorical variable with two categories – for example, if chocolate consumption were categorized into two categories, namely High/Low – then the predictive model would take the form of a logistic regression.
Perceptual Mapping is a graphics technique used by asset marketers that attempts to visually display the perceptions of customers. Typically the position of a product, product line, brand, or company is displayed relative to their competition – for example, one could graphically display the perceptions of different chocolate brands on a variety of product characteristics such as taste, price, color, packaging.
Big Data Management and Analytics is the collection and management of information from one or more sources and the distribution of that information to one or more audiences. We have built the infrastructure and expertise to help companies organize and deliver the information they need to effectively make better business decisions that are strategically aligned with the company goals. b3Intelligence adopts this model through the following process:

  • Identification of information needs
    • Acquisition and creation of information
    • Analysis and interpretation of information
    • Organization and storage of information
    • Information access and dissemination
    • Information use

The benefits from managing information strategically are centered on key standard objectives such as how to reduce costs; reduce uncertainty or risks; add value to existing products or services; or create additional value through new information-based products or services. b3Intelligence looks at all these factors when taking on your information management needs.

Brand Related Studies ensure you get to assess the brand value of your product and that of your competitors. The studies help clients understand how to create brand value and equity for their products and position them distinctly to the target audience.
b3Intelligence also studies medical practitioners from close quarters as part of pharmaceutical research. This helps in identifying their prescription practices, the way they subscribe medicines, and how the drugs get a market.
Market Share Analysis provides a deep understanding of the market and competitive landscape and can generate the development direction for defining the unique product attributes and positioning. b3Intelligence’s analysis helps provide the insights to create the strategies that can increase product and company success.
Affinity Analysis is a statistical method for learning relationships amongst variables in databases. The procedure helps in deciphering the multiple connections between variables and their eventual impact. This method is a cost effective way to identify opportunities for positioning and selling products in the market.
For example, b3Intelligence’s Affinity Analysis can provide medicine and drug makers a road map on how to position their products, find new sales channels and reduce R&D costs.
Market Basket Method is used for taking the cover off purchasing habits of consumers. This method is used to help make buying habits more transparent and ensure an understanding of what consumers value and want to buy.
This method is key to execute for companies in the pharmaceutical industry, as consumers are very cautious about the drugs they purchase and concerns about product combinations, chemical concentration, and other preferences.
b3Intelligence delivers the insights that the pharmaceuticals can leverage to develop the product mix that consumers favor.

We use multivariate weighting for several experimental design type projects, where the groups need to be balanced on multiple variables.

Rather than using cell weighting, which would be practically impossible when dealing with multiple variables in their entire granularity, we used propensity-weighting schemes to balance the groups.

These multivariate weighting techniques generally employ logistic regression models to generate the weighting schema.

Market Structure Insights and Product Positioning uses perceptual mapping to help clients understand the competitive market structure in addition to ascertaining whether adding a particular feature to their product would bring them closer to the “ideal” point described by the market.

This analysis showed our client their position relative to the competitors as well as the ideal point in terms of the market perceptions. This analysis also helped address the issue whether they should invest in the additional feature, which they thought would bring their product closer to the market ideal – the answer in this case was no.

The insights revealed that price adjustments rather than investing in an additional feature was the optimal option.