Customer Analytics

Data Mining automates the detection of relevant (correlations,trends and) patterns in a database. Data Mining uses well established statsitical and machine learning techniques to build models that predict (rather than retrospect) customer behavior.

What data mining will discover nobody can anticipate. There in lies the artistic beauty of the new age science that it is. It is important for Marketing managers to understand and appreciate these customer insights to turn them in to actions.

Data Mining/Machine Learning in the Hadoop world has three categories by learning mode -

1. Supervised Learning – Involves utilizing training data set to identify significant variables and correlations across these variables. Once trained, the model is used to predict outcomes on new data set.

2. Unsupervised Learning – There is no training involved. The model looks for patterns and correlations above certain significance factor and then these rules are used analyze and act as input for patterns.

3. Reinforcement learning – There is no training here as well but this is a feedback loop based learning approach. These models are adaptive and optimize themselves over time and has the potential to solve very difficult time dependent problems.

Another interesting way of looking data mining is the context of applications it solves -

1. Classification Problem – Classify a transaction as high potential for fraud.

2. Clustering/Segmentation – Classifying a population based on common behavior exhibited or other unknown factors. Example those who like movie A also tend to like movie B.

3. Associations – Identify associations between different entities or actions, those who bought A also bought B.

4. Sequencing – Pertains to time based patterns. Those buying A now will usually buy B 6 months later.

Machine Learning, Statistics and AI methodologies are used in data mining.

According to Forrester research findings on the state of customer analytics in 2012, firms do not Measure profitability and engagement, only trial conversion.

Customer Churn  - Some customers will leave no matter what. Some will stay no matter what. The focus needs to be on those that require touch, require attention and that is where TrendMatrix comes in. We will identify the target customers that need your intervention. The key is determining who these customers are and what kind of intervention works best to retain them. Timing is another important factor, it is not just how to intervene but when to intervene that is as important if not more, to attain higher retention.

CRM is a process that manages the interaction between a company and its customers. In that regard we can classify ourselves as a CRM company. Our solution will deliver the customer insight to help marketers, sales professionals and business executives to identify the profitable customer segments, understand what features are most valuable to them, inform how they find your service and help you deliver the right features and attract the right customers.

Identifying the profitable customers involves not just understanding who is signing up for your most high end plan but also understanding who is sicking with you the longest. These results must be fed back in to the campaign management software to derive the maximum ROI for your your marketing spend.

Data mining and campaign management must be tightly integrated.

When a customer signs up for a trial, as time goes by, you should be able to tell with more and more accuracy the probability that the customer will sign up for your service or just churn in trial! A conversion score is the output of your data mining software. As more and and more customer data is fed to the system the prediction will become stronger and deliver you greater results. This builds a case for why you should start today instead of waiting to acquire more customers to help you increase trial conversion in the long run.

Customer analytics and executing on the derived insights will provide you a sustainable competitive advantage.

Data warehousing is the combination of technologies that integrate operational databases into a single cohesive environment. The warehouse is aimed at providing input to decision makers by providing aggregate data across operational databases. A data warehouse provides an holistic perspective of the data across departmental/operational boundaries in the company and this is fundamental to realizing strategic benefits from this data. In that sense TrendMatrix is your data warehouse. While you store a copy of all your operational data, you surrender a copy to TrendMatrix which is digest the data, aggregate, summarize and provide you a clean informational view of business ready for drill-down along various dimensions.

It is no longer sufficient to rely on one tool to understand your customers, you need an integrated platform for customer analytics, a data warehouse that provide you individual account level detail to identify opportunities to do better an attain your profitability and customer experience goals.