When most people hear the terms big data, data analysis and machine learning they associate these terms with big corporations and large-scale manufacturers. Few people dig deeper and question the application of these technologies in small to medium businesses. Those that do uncover some interesting ways that small business can leverage these technologies for growth and competitive advantage.
Aim for simplicity in Data Science. Real creativity won’t make things
more complex. Instead, it will simplify them.
– Damian Duffy Mingle
Cost is often the first obstacle small business owners are faced with, transforming into a data driven business can at times be a high-cost investment. These costs are associated with skills acquisition, data scientists and data engineers don’t come cheap. Next is technology, cloud infrastructure, data processing and mining tools also come at a cost. A run through these numbers would make a small business owner reconsider their journey to being a data driven business. But this is changing thanks to tech companies and their agile business models, SaaS, IaaS, PaaS solutions from companies like Microsoft Azure, AWS and Google have brought about competitive and affordable software as a service solutions that small business owners can leverage off.
Use Cases for Data Science
There are various possible use cases for data science and analysis in small business with each one being unique. Many machine learning algorithms have been developed over the years and have been used by many businesses to find insights in their data and make predictions. Much popular is the linear regression model, which is a supervised machine learning algorithm that predicts a given outcome dependent variable (y) based on independent variable (x). Say you are an Internet Service Provider (ISP) and you’re looking to answer the question of “How long will a given customer use your service”.? This could be a question posed to solve a marketing challenge or for your financial forecasts. The variable to be predicted here is “Service usage”, through the analysis of historic customer data on customer usage patterns. A model could be developed to predict an individual’s “service usage”. Armed with these insights as a business owner, you could now develop strategies to ensure that those with low service usage scores are offered incentives to stay with your service much longer instead of losing customers.
If that’s not enough consider Thabo who operates a busy supermarket in the city, Thabo could make use of market basket analysis a technique used to find associations between data points. In this case these would be the various transactions recorded by the Point of Sale (Pos) machines in the supermarket. A model could be developed to answer questions such as “What items were bought together the most by customers”.?. This analysis could reveal that people like to purchase rice, fish and mayonnaise together, the shop owner could use this newfound information to offer product discounts, product bundling and other marketing tactics that could drive sales through the roof for the supermarket.
These examples we’ve mentioned are only the tip of the iceberg, various other models exist that could be applied on data generated by small businesses. Becoming a data driven business in a fast changing and highly competitive business market is a must. Business owners need to seriously consider the transition form being old school intuition based decision makers to becoming modern day data driven decision makers.
