NEED HELP ?

Too much data and not enough time and resources to leverage it intelligently? Need someone to assist you in decision-making by mining the data you have on your Customers, Products, Employees or Operations? Have an urgent need for sophisticated data analytics? Or just plain data management? Get a 2-pager on us. And contact us .

August 28, 2014

Credit Risk Scoring using non-traditional data

Unbanked and underbanked Customers are regarded by creditors/investors as high-risk borrowers due to insufficient information about their assets and liabilities. Given lack of this information, traditional risk-evaluation methodologies do not work.  However, psychometric risk profiling has been growing in popularity for determining the credit worthiness of such individuals. This exercise utilizes a blend of psychology & statistics to come up with questions that are used for reliability & validity of attitudes and behaviors of people. We helped our client, a leading financial services company, measure risk and business potential for such customers in emerging countries through psychometric tests, which asked questions about their attitude and beliefs, financial acumen, problem solving skills etc. to generate a risk-score through proven statistical techniques.

Credit risk modeling is a very common statistical technique used in the Banking and Financial Sector which involves analyzing historical data of borrowers to identify certain characteristics that predict the likelihood of the borrower defaulting on his/her loan in the future. The data could include, for example - past transactions, credit history, default in months, years in business, etc. At Marketelligent, we leveraged this conventional modeling technique to come up with a predictive credit score for individuals based on the psychometric test conducted. The data included socio-demographic, situational, psychological and behavioral variables. The operational procedures for building the predictive model involved:
  • Analysis of characteristics of each variable and relations between them
  • Identify possible inconsistencies and missing cases
  • Individual comparison of variables with good and bad credits
  • Redefinition of some variables and creation of other new derived variables
Various approaches like Decision Trees, Neural Networks (Gradient Boosting), Genetic Algorithm and Logistic Regression were used with the inclusion of psychological variables and scales, in order to come up with a robust model that would help us provide a probability score on each individual based on their survey responses. The analysis was done by calculating the Area under Curve (AUC) of all the countries and emphasized on to improve the AUC’s of the 5 worst performing countries.  The predictive scorecard developed had strong ability to identify risky customers. Model also worked well across countries.  Credit risk score generated helped the issuer in lending to previously un-bankable but credit-worthy customers, thereby directly improving the lives of struggling customers across many developing economies.

More details can be found here.

July 16, 2014

OOMF – Order of Magnitude Forecast

In today’s fast paced business ecosystem, manufacturers are increasingly looking to expand globally and utilize vast opportunities available in emerging and rapidly growing new markets. In this scenario, it becomes extremely important to identify the right markets and the right entry strategy. One of the key factors in developing a robust entry strategy is in understanding the market structure. Market structure analysis helps companies in understanding how various products are positioned, their interactions and in identifying gaps in the marketplace.

Our client, a leading pharmaceutical and CPG manufacturer wanted to establish presence in a new market. We did the market structure analysis to help them understand the competitive landscape and their current position in that environment, and also to identify opportunity areas to enter the new market. This was done by doing an in-depth analysis of historical sales by products, segments, formats, ingredients, price, packaging, distribution, communication, etc. This was overlaid with macro-economic factors, trade regulations, demographic factors and other tactical parameters in forecasting how the structure will change over the years.

The gaps identified were mapped with the company’s products and portfolio strategy to develop various product portfolios to be evaluated for the launch. A simulator was developed to help the manufacturer get an Order of Magnitude estimate of sales for various scenarios of the launch, the sequence of launch, marketing investments required and various other parameters impacting the launch. The client was able to quickly evaluate the opportunity and identify the right portfolio to launch in the market. The market structure analysis and the simulator developed are detailed below.

Market Structure Analysis:
We conducted a Market Structure Analysis to assess the size of prize, right to win and an order of magnitude forecast for a new entrant which helped the manufacturer understand the key value drivers impacting a new market entrant. This understanding helped to simulate various scenarios and develop the best market entry strategy.
Market Structure
Using information from the strategic and tactical launch parameters, the leadership was able to design a launch plan for the product in terms of identifying target markets, target consumers, suggesting the order of entry strategy, establishing the right product positioning and the right marketing mix & distribution. Along with this, the leadership got an idea on the overall product performance in the new market with respect to customers, financials and technology.

Simulating ‘New Market’ entry:
After the holistic analysis of the ‘new market’, we developed a simulator to weigh in different launch strategy options to enter the ‘new market’ in a way that marketing spends were optimized and best returns on investment were derived for the manufacturer entering the ‘new market’. The simulator helped to understand the impact of the newly launched product over the years and estimated the size of prize for the product launched in the ‘new market’.With this simulator, the manufacturer was able to select from a basket of new products; the right product to be launched in the market and the right time to launch the product.

For multiple product launches, this tool helped time the market right and sequenced the launches in an optimal manner in order to maximize returns. The manufacturer was able to visualize the market impact of the product launch by predicting the value share of the new product in the market on an yearly basis for the next ‘x’ years and thus foresee the success/failure of a brand launch in the long term.

Based on the above analysis, we leveraged the simulator to evaluate scenarios and created an optimal portfolio to help build the launch strategy, sequence and time the new product launches in the market in order to uphold the long term financial goals and vision of the manufacturer.


OOMF - Order of Magnitude Forecast Tool
This robust analytical approach helped the company in significantly reducing the risk and fastened the entire launch process. 

More details can be found here.

June 18, 2014

Who are my Social Influencers ?

According to a survey, influencer marketing campaigns drive 16X more engagement than paid or owned media.

Influencer marketing incorporates engaging with specific key individuals that have influence over potential buyers and who can have a significant impact on their purchasing decisions. Reaching potential buyers during their key decision-making processes is one of the great ways to enhance company’s sales with minimal investment. 

These influencers can help generate genuine brand awareness and more importantly persuade others to take action. Influencers with deep social media presence help spread the message to a wider audience.

Our client - a leading publisher of scientific journals - was trying to deal with a rapidly evolving industry. Researchers felt the publisher’s pricing policy was too aggressive and this was hampering dissemination of quality research to a wider audience. The publisher wanted to engage with researchers and thought leaders having influence in the publishing domain to gain insights regarding best practices of the industry, so that they could fine-tune their business model accordingly.

Marketelligent helped the publisher identify key influencers using human analysis and recognition tools. All influencers were ranked based on their digital media footprint and reach among consumers of publishing industry.         


Based on the ranking, top ten influencers were profiled with specific details like Klout Score & # of Twitter followers. Influencer profiling also included connection diagrams which showed linkages of each influencer with organizations or individuals who are authorities in the publishing industry.

The Result:  The academic publishing company engaged with top influencers of the industry to gain great insights regarding business model going forward.  Key enhancements were made:
  • Open access was given to a number of articles in tune with other publishing companies
  • Subscription rates were overhauled to make articles easily available for researchers and scholars
  • Positive reviews in terms of articles and blogs written by these influencers helped the company enhance its brand value
More details can be found here.