BusinessOptics’ Template Models

Businesses today can gain a competitive advantage by informing their decisions using data analytics. Unfortunately most domain experts are not well equipped to do data analytics themselves due to a  lack of the skills required (programming, advanced statistical techniques, machine learning, etc). Consequently, most businesses hire “data scientists” at a high cost in order to extract value from their data.  

The BusinessOptics platform aims to improve that skills gap by hiding many of the technically difficult aspects of analytics from the user. Its visual programming paradigm allows business users to quickly build complex models without having prior programing experience or formal statistics or computer science training. Furthermore, BusinessOptics has developed templates that can be used to generalise these models in order to enable them to be easily reused. This means that businesses can easily reuse previously created models to quickly derive the vast majority of the value from your data without having to waste effort rebuilding the core of the model (i.e. immediately extract 80% of the value with zero effort). These analytics models can further be customized to suit your business needs and extract the additional business value on offer.

To illustrate how easy and quickly it is to deploy a model using the templates, let’s consider BusinessOptics’ Sales and Customer Analytics solution. This solution is designed to help you maximize the success of your sales campaigns in environments such as call centres using business metrics such as the Customer Lifetime Value (CLV). We want the sales analytics models to advise us on important decisions such as which customers to contact, at what time of the day, which sales agents should speak to which customers, what products should be sold to them and at what price. Making intelligent decisions here can have enormous impacts on profitability, we’ve seen improvements up to 50%. The diagram below illustrates the different models that make up the CLV solution.

Customer Lifetime Value
Different Models that can be implemented to optimize Customer Lifetime Value (CLV)

The customer lifetime value model is made of three submodels:

  • The response model
  • The take-up model
  • The persistency model

The response model predicts whether or not an individual with certain characteristics will respond positively when contacted about a product. The next model (the take-up rate) predicts the likelihood that individuals who responded positively to the offer actually take-up a contract and make their first payment. The persistency model looks at the payment behavior of the individual during the customer life cycle. Considering all of these together with product pricing and costing we can determine the expected customer lifetime value.

BusinessOptics’ Sales and Customer Analytics model can be deployed in just a few steps using the existing template to immediately assist in decision making and help improve profitability.

To utilize the template models, we want to:

  • upload and validate data. This helps us see what features of the data we can learn from, for example customers’ incomes, educations, etc.
  • Set model options. This helps take the generalised template models and make them specific to you.
  • Set model options. This helps take the generalised template models and make them specific to you.
  • Investigate the model outputs. We can look at insights to see trends in the data such as how customer lifetime value is affected by combinations of customers’ income and education.
  • Use the models in your business to drive decision making. The outputs of models can either be exported or integrated into external systems (such as 3rd party dialling systems) in order to make use of the model output.

These steps are illustrated in the figure below.

Data Process
Data Ingestion Process

Select Data

The data that will be used for the analysis is uploaded in this step. Our data engine will take care of cleaning the data using actions such as removing undesired characters and replacing blank cells with default values.

Accept Data

After uploading the data, the summary of the data will be displayed for easy high level sanity checks

Accept Data
Data Review

Accept Data Fields

Thereafter field manipulations can be performed such as dropping fields that are not required, renaming fields for consistency and changing variable types (text, number, date, etc)

Configure Model Options

This is one of the most important step in the use of template models. The parameters for that specific type of analytics model can be easily specified and give you control over specific functionality within such a model.  Examples of such parameters are parameters relating to:

  • The way to split the data into a partition for training the model and another partition for model evaluation
  • Specific methodology options for sub components. For example, parameters specifying options for fitting separate models relating each of the attributes to the target variable.
  • How data sparsity is handled in the models. That is, checking whether or not there are enough fields in the data for the model to learn on.
  • Optional features derived insights from the related data

Once all the desired options have been specified, the corresponding analytics model will be generated at the next step.

Investigate Analytics Model

At this point, one can

  1. Investigate the model that has been generated.
  2. Train (i.e fit the model to the data) all the machine learning models that are included at once.

This is achieved by selecting one of the two options in the figure below.

Investigate Workspace
Investigate Analytics Model

The analytics model with associated business logic can be accessed by selecting the “Investigate Workspace” option. One can then make additional changes at this stage to further tailor the model to one’s needs.  An example of the logic behind such models is given in the picture below. The window on the left shows a mind map of the different modules making up the model and how they connect to one another. By selecting any of those modules, the parameters associated with it will be displayed in the window on the right. These parameters can be changed according to one’s needs.

Completed Idea
Mind Map and Parameters

BusinessOptics’ templates make use of advanced machine learning algorithms and optimization solutions to derive actionable insights from your data. You have the option to choose a specific machine learning algorithm or to let the platform optimize and choose the best algorithm for you.

Once the model completes training, one has access to interactive dashboards that display the insights derived from the data. The results of the analysis can further be explored using parameters included in the dashboards.

Make Use Of Models In Production

The results of the analysis can easily be shared in your organization and integrated in your operation using the BusinessOptics extensive API.


Thus using just a few steps as described above, BusinessOptics’ templates can help you quickly deploy analytics solutions that model your specific business and extract value from your data. These models can help you improve sales conversions, increase customer lifetime value, decrease sales costs and increase customer retention.

Intrigued? Schedule a demo today!