Project Proposal Predictive Analytics

Predictive Analytics
Project Proposal
Introduction
Predictive analytics has got to do with the extraction of information from the various data sets that are available at present with an aim of identifying specific patterns and predicting future trends as well as outcomes. It is important to note that predictive models do not necessarily determine the future. Instead, they are predictions or conjectures of future events. The relevance of predictive models cannot be overstated in business. This is more so the case given that they help in the analysis of historical trends as well as present data to not only highlight opportunities for a business enterprise, but also discern potential risks. In essence, predictive analytics could be seen as a consequence of big data. Today, organizations deem data as a source of competitive advantage. In a business setting, data could be inclusive of customer details, industry and market statistics, risk models, supplier information, employee info etc. Business entities could use data to not only inform strategy, but also assess the market for opportunities. Data sets could be complex and large – giving rise to the term big data.
Problem Statement
To remain relevant in the increasingly competitive business world, businesses ought to make use of various strategies and approaches. One such approach is the use of coupons. In essence, discount coupons are of great relevance in seeking to further enhance sales by guaranteeing financial discounts to persons who purchase various products or services. It is, however, important to note that a business entity cannot offer coupons to all customers. For instance, the customers who walk through our doors to make various purchases are in the millions. It would be a poor business decision to advance a 10% discount to all customers. Towards this end, there is need to conduct customer assessment so as to determine not only those who are most likely to make purchases, but also those to whom advancing a coupon is likely to optimize revenues. Predictive analytics could, therefore, be of significant utility on this front.
Research Question
Would advancing discount coupons to those most likely to buy our product increase redemption rates?
Some of the companies that have made use of predictive analytics to predict product choices for personalized recommendations include, but they are not limited to, Amazon, Netflix, Tesco, Target, and Pandora (Siegel, 2013). For instance, with regard to Tesco (UK), the company issues millions of coupons to customers at its various cash registers. Towards this end, “predictive modeling increased redemption rates by 3.6 times, compared to previous methods” (Siegel, 2013).
Brief
Currently, our company offers discount coupons in a largely erratic manner. Coupons are issued randomly – with the 100th customer at any of our stores receiving a coupon. Most of these customers may never redeem the coupons. Predictive analytics will in this case turn out to be useful in seeking to ensure that coupons are only issued to those customers who are most likely to make a purchase or place an order for various products. In that regard, therefore, the utilization of predictive analysis is likely to come in handy in seeking to optimize revenues. With a sound methodology of offering discount coupons, the company could get a more profitable way of disposing unwanted inventory, i.e. products that have stayed in the shelves for a long period of time. Further, the coupons will help in the further enhancement of customer royalty. This is more so the case given that customers are likely to view them as rewards.
In this endeavor, I would recommend that we pilot the undertaking at our New York stores. Successful outcomes should in this case be used as a basis for rolling out the idea to our other locations countrywide. The attributes that could be used as predictor variables include, but they are not limited to; age, gender, marital status, occupation, place of residence, value of purchases made within a period of three months, three products that the customer has routinely bought within a period of three months, average price of individual products the customer purchases, the location of stores most purchases are made from, and shopping preferences (online or physical).
References
Siegel, E. (2013). Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die. Hoboken, NJ: John Wiley & Sons.