CAN PREDICTIONS WITH R HELP A SMALL START-UP COMPANY INCREASE ITS POTENTIAL SALES?

• Main Authors: Mircea Radu GEORGESCU, Ionuț-Daniel ANASTASIEI
• Format: Article
• Language: English
• Published: Technopress 2020-06-01
• Series: Journal of Public Administration, Finance and Law
• Subjects: r predictions; logistic regression; digital marketing
• Online Access: http://www.jopafl.com/uploads/issue17/CAN_PREDICTIONS_WITH_R_HELP_A_SMALL_STARTUP_COMPANY_INCREASE_ITS_POTENTIAL_SALES.pdf
• ISSN: 2285-2204; 2285-3499

The ERP solutions which include the predictions modules are very expensive and very hard to comprehend whenever a new company starts its activity. According to a survey (Ly, 2019), small to medium-sized businesses can expect to pay somewhere between $75,000 and $750,000 for implementation and this expenditure grows even larger for large businesses. Fortunately, there are some free tools available which can be used to implement a small part of an ETL professional process. Of course, we may have to admit that you also need some technical skills in order to learn how R, but also some statistical ones. The R language is very easy to use when it comes to implement regressions on the actual data of companies, and it comes at zero costs. Also, there is almost none ETL (extract-transform-load) technics needed because the client portfolio of small businesses is not large enough to be worth investing into. The statistical formula used for predictions was logistic regression and it intends to create a model to predict the probability of buying a product based on the yearly income of a costumer. To make these concepts easier to explain in this article, we have considered a toy problem where you only have one customer characteristic (the customer’s yearly income) and a data scientist from a small company wants to predict if the customer will buy. This matter can be extended in future studies which can conduct the predictions of multiple independent variables, binomial or multinomial. Mainly, this article also admits that the use of digital marketing to reach the potential customers is very important, but more important is to predict the behaviour of a potential client whether it will buy or not our solution so that the company may set its own expectations.