If you have a Small or Midsize Business (SMB), you too need to evolve from a highly subjective HIPPO (Highest Paid Person’s Opinion) decision making style into a more data-driven one. Problem is that you cannot easily take time off from your busy days running your business to get a new degree in Data Science. Luckily tools like BigML are reducing the barriers to conduct advanced analytics using proven and well-researched Machine Learning algorithms such as decision tree models or Random Forest. All this is achievable by using tools like Excel (or Google Sheet) and without any coding skills.
You first need to pick a predictive use case. In other words, what are you trying to predict? Is it the sales for a given product on inventory needed next month? Is it the likelihood for a repeat customer to churn? Is it what other products go well with a given product? Their domain knowledge and creativity are keys to this step.
Once you decide on that, you can export data out of your POS (Point of Sale) systems or similar systems of record onto a spreadsheet. In the spreadsheet, you can reorganize data columns, create new calculated columns (such as average spend per customer etc.) as long as you think these additional variables may correlate with your target variable that you are trying to predict. This is where your domain expertise and judgement is also key.
When you have your final dataset with all the columns representing your variables and the rows representing each instance or observation, you can easily import it into BigML in seconds by dragging and dropping into your browser screen. BigML automatically gives you basic statistics of your dataset so you can quickly identify oddities or potential data quality problems like missing values. After that, with few clicks and no coding whatsoever you can build a very visual model to predict your objective field. One of the cool things about the model is that it will tell you what factors are most useful in predicting your objective field (in this case expected sales or customer satisfaction or customer life-time value, etc.) and which other variables can be practically ignored. We have many other features that can also help you. Please visit this other question for a general overview.