Finding a Happy Medium in Your Forecasting
Forecasting has become the “third rail” of business – a practice considered both untouchable and controversial in the finance world due to the wide-ranging opinions on the ‘right’ approach. Some companies have a good handle forecasting and effectively achieve their financial goals, but many others struggle with meeting their targets and suffer financially from their incorrect predictions. One characteristic of the companies who get it right is their keen understanding of the best data needed to fuel their forecast models. A recent article in the Harvard Business Review, The Forecasting Sweet Spot between Micro and Macro, discusses the analytical challenges with forecasting and highlights how financial leaders can find a happy medium between data types to better forecast their future.
A KPMG survey of 500 finance executives showed that only one percent of companies hit their financial forecasting goals, and only 20% were within 5% of their goals. One challenge to consistently hitting forecasted targets is managing the multiple agendas that exist in the organization that surround the forecast. Sales leaders responsible for delivering a revenue forecast may want to lower expectations to improve their success rate. Support departments seeking more resources may look to increase expected expenses to get more discretionary funds. While competing priorities will exist in every company, finding the right balance of data and assumptions will drive the accuracy of any forecast.
In addition to harmonizing internal expectations, there are other data considerations crtical to the success of a forecast. With the explosion of big data, professionals are hopeful that the availability of information will improve the accuracy of forecasts. However, the truly successful organizations will have deep insight into the kinds of external data necessary to deliver effective forecasts.
Some companies focus on macro data – GDP, population, demographics – as part of their models. But using too much macro data makes it impossible to focus on the trends that are significant to one specific industry or company. Other models emphasize micro data but these data points are often too specific to capture bigger disruptions that can affect the forecast. This study highlights the importance of “middle data”.
Middle data can include information about life events and triggers at the individual level that can drastically influence the need for certain products or services. Such information includes getting married, moving to a new home, or finding a new job. In the article, middle data is described as data that is “closer to actual consumers than far-flung data like GDP, but it elevates the frame of reference, as most companies mistakenly believe consumers spend more time thinking about their categories and brands than they really do.” Middle data is the best prediction of consumer behavior because it puts prospective customer’s wants and needs in perspective for the company trying to sell to them. This information helps the seller forecast more accurately, which will ultimately lead to better business decisions.
An example in the apparel industry is the influence of religion on how people dress in a certain locale. In areas where Christianity is growing, the sales of Western style clothing generally increases. Apparel producers looking to build a refined sales forecast should use trend information like this to make decisions on what product to sell and where to manufacture them. Another example is an entertainment company using middle data like home types and local climates to help determine the demand for their products.
The UN human development index, a publicly-available composite statistic of education, economy and infrastructure, is another great predictor of human behavior. Food and beverage companies can use the index to determine whether the population of an area are “eating and drinking to live” or “living to eat and drink”. For the former group, companies should focus on selling functional items like bread and dairy products while the others have an appetite for more enjoyment-related items like snack foods and alcohol.
In forecasting, like many analytical practices, you only get what you put into it. Take the time to consider and include the types of relevant “middle data” that can improve your forecasting process.
It also helps to have an enterprise modeling tool to leverage both internal and external data to get the most accurate results. Platforms like ImpactECS by 3C Software make it possible to build operational budgets to guide decisions. Companies have the ability to adjust or flex any budget to forecast expected future results, generate unlimited scenarios adjusting any model variable, and maintain multiple versions of budgets and forecasts simultaneously.