There is quite a bit of unpredictability in sports; that’s why they are fun to watch. While being able to predict the results of every play in a game or a match would take all the excitement out of it, owners and managers could certainly benefit from predicting what it could take to lead their team to a winning season. Like a prophetic crystal ball, data analytics helps sports teams and businesses alike make predictions like these with confidence.
If there was a single recipe to follow, everyone would be (or at least should be) projecting predictions to guide their path. However even among different sports there are different degrees of unpredictability. Looking at variation amongst top performing teams year to year, the National Hockey League affords viewers the greatest degree of discrepancy each year. Does this make it the most exciting? That’s for you to decide.
Regardless, there is an entire industry built up around predicting sports (with varying degrees of success) worth billions of dollars. Probably one of the most popular modern stories was depicted in the 2011 sports drama film, Moneyball. Based on a true story, Moneyball gives the account of how assistant General Manager Billy Beane led the Oakland A’s to the playoffs in four consecutive years and the first team to win 20 consecutive games in the 100+ year history of American League baseball. Since then, other teams within Major League Baseball, across other sports, and in businesses alike are challenging the old way of doing things and relying more and more on data analytics to drive decisions based on forecasted predictions.
For Want of a Crystal Ball
This is one of the primary functions of data forecasting. By estimating the chances of future events occurring or not occurring, businesses can make the decisions they must make in order to not just survive but to thrive.
Inventory orders are the classic example of this. How much should you buy? Will demand stay consistent, or will it spike or drop off? If you order too much of course you will be stuck with excessive storage costs and perhaps wastage. Order too little and you won’t be able to deliver to demand, perhaps permanently damaging consumers’ trust in your brand.
And the rate at which customers come in is likewise unpredictable. Having no knowledge of this, no ability to forecast it at all leaves a business totally exposed to the elements. One minute you are inundated with customers, the next you are considering laying employees off and downsizing. Even if an organization can survive this, the damage is permanent.
A lack of forecasting can even rear its head in marketing and outreach efforts. How can a company ever justify spending money on marketing and sales if they have no ability to forecast the results within some range? Too major of a result and they are inundated, too minor a result and they are left wondering what went wrong. Either way the money was wasted.
“What’s the Forecast?”
Forecasting can prevent all of these scenarios, or at least lessen the likelihood of them occurring. While there are many different methods of forecasting – each bringing their own considerations – all serve the same core function of allowing a company to react to events before they happen.
Forecasting makes use of several different factors. Like any analytical project, we of course need data. The more we have, the more confident we can be in our prediction. If History is the best teacher, then historical data is key. If there is consistently a spike in demand for certain product categories during December, for example, this will factor into the analysis.
The type and extent of data available must be a consideration for those performing the forecasting. Some methods are more accurate than others, but require data that is not always extant. If an attempt is made to force fit data to a forecasting method or vice versa, the results are predictably bad.
Likewise, certain forecasting methods, while more reliable, are more extensive and therefore costlier. Balancing this cost with the potential cost of an error is key. We have to remember that the purpose of forecasting is not to achieve the absolute maximum certainty, but to economically minimize that uncertainty so as to consistently make informed decisions that lead to profit. Forecasting is yet another tool for our Risk Management strategy.
And forecasting cannot exist outside the rest of the analytical process. Data visualization, the process of creating legible charts and graphs out of a mass of data, is a first step towards determining what actually needs to be forecasted. Forecasting is of course a part of the decision making process, but can also be a product of it. When analysis leads to the decision to do X, forecasting can help ascertain the best way to get there and quantify the expected result.
Don’t leave business success to chance! In business, we need to know with the utmost confidence possible the potential outcomes of our decisions. Forecasting naturally leads into our next discussion – prescriptive analytics. With all the data at our fingertips, which course should we pursue? Next week we will discuss this process in depth and how it relates to your business.